No. 19 SMU @ Florida State
Mon Sep 7, 7:30 PM EDT
Model: Florida State by -0.1 (43% win prob) · Market: SMU by 3.0
Moneyline pick: SMU (57% win prob) LOW Confidence†
Spread pick: Florida State +3.0 (-111) LOW Confidence†
The market favors SMU by 3.0, but the model actually favors Florida State, essentially a pick'em at -0.1 -- a 2.9-point swing that makes Florida State the pick, though this is a low-confidence, close-to-even edge.
- Florida State rates ahead once schedule strength is factored in, a 10.8-point opponent-adjusted margin edge (+17.0 vs +6.2), despite pre-game ELO‡ giving it only a 43% win probability.
- SMU has the better recent scoring margin (16.5 vs 2.0, last four games) and has covered the spread in 80% of its last five games, both real factors working against this pick.
- SMU has also faced tougher recent competition than Florida State (+2.7 vs -9.4) -- with the numbers this mixed, this is a genuinely close call, not a strong lean.
Per-model breakdown (home win probability)
Logistic Regression40%
Random Forest45%
XGBoost43%
LightGBM48%
Extra Trees43%
Final: Florida State 24 - 27 SMU — moneyline pick was correct, spread pick push the spread
No. 24 Louisville @ No. 9 Ole Miss
Sun Sep 6, 7:30 PM EDT
Model: Ole Miss by +10.0 (78% win prob) · Market: Ole Miss by 6.8
Moneyline pick: Ole Miss (78% win prob) HIGH Confidence†
Spread pick: Ole Miss -6.8 (-112) MEDIUM Confidence†
Ole Miss is favored by 6.8 in the market, and the model likes Ole Miss slightly more, by 10.0 -- a 3.2-point gap in Ole Miss's favor, making Ole Miss the pick at medium confidence (58% cover probability).
- Ole Miss has the edge in recent scoring margin (12.8 vs 3.2, last four games) and a 5.5-point opponent-adjusted margin advantage (+11.7 vs +6.2), against tougher recent competition (+8.5 average opponent rating vs +4.6).
- Louisville has covered the spread in just 20% of its last five games.
- Pre-game ELO‡ gives Ole Miss a 61% win probability -- a real but not overwhelming edge, matching the modest confidence tier here.
Per-model breakdown (home win probability)
Logistic Regression77%
Random Forest75%
XGBoost70%
LightGBM72%
Extra Trees73%
Final: Ole Miss 41 - 38 Louisville — moneyline pick was correct, spread pick did not cover the spread
Wisconsin @ No. 4 Notre Dame
Sun Sep 6, 7:30 PM EDT
Model: Notre Dame by +22.8 (91% win prob) · Market: Notre Dame by 20.2
Moneyline pick: Notre Dame (91% win prob) HIGH Confidence†
Spread pick: Notre Dame -20.2 (-110) LOW Confidence†
Notre Dame is favored by 20.2 in the market, and the model likes Notre Dame slightly more, by 22.8 -- a 2.6-point gap in Notre Dame's favor, though this stays a low-confidence pick given how close the numbers are (56% cover probability).
- Notre Dame has a large recent scoring margin edge (38.2 vs -3.5, last four games) and carries a 231-point ELO‡ advantage (1717 vs 1487).
- Notre Dame also has the edge in recent yards per play (7.9 vs 4.2).
- Wisconsin has faced tougher recent competition than Notre Dame (+10.4 vs +3.0), part of why this edge stays modest despite Notre Dame's clear form advantage.
Per-model breakdown (home win probability)
Logistic Regression96%
Random Forest92%
XGBoost94%
LightGBM98%
Extra Trees92%
Final: Notre Dame 41 - 13 Wisconsin — moneyline pick was correct, spread pick covered the spread
Washington State @ No. 17 Washington
Sun Sep 6, 4:00 PM EDT
Model: Washington by +10.5 (80% win prob) · Market: Washington by 23.2
Moneyline pick: Washington (80% win prob) HIGH Confidence†
Spread pick: Washington State +23.2 (-108) HIGH Confidence†
Washington is favored by 23.2 in the market, but the model's margin has it by 10.5 -- a 12.8-point gap that makes Washington State the pick, at high confidence (79% cover probability).
- Washington still rates as the better team: a 70% ELO‡ win probability, the edge in recent scoring margin (21.5 vs 14.5, last four games), and tougher recent competition (+5.6 vs -0.4).
- Washington State has covered the spread in 80% of its last five games, the strongest recent ATS trend in this matchup.
- Washington's edges here are real, but the market's 23.2-point spread runs well ahead of what those numbers support.
Per-model breakdown (home win probability)
Logistic Regression77%
Random Forest78%
XGBoost81%
LightGBM76%
Extra Trees72%
Final: Washington 24 - 10 Washington State — moneyline pick was correct, spread pick covered the spread
UCLA @ California
Sat Sep 5, 10:30 PM EDT
Model: California by +9.1 (79% win prob) · Market: UCLA by 2.0
Moneyline pick: California (79% win prob) HIGH Confidence†
Spread pick: California +2.0 (-110) HIGH Confidence†
The market actually favors UCLA by 2.0, but the model favors California instead, by 9.1 -- an 11.1-point swing that makes California the pick, at high confidence (76% cover probability).
- California has the better recent scoring margin (-4.8 vs -24.5, last four games), and pre-game ELO‡ gives it a 64% win probability.
- UCLA has faced tougher recent competition than California (average opponent rating +12.6 vs +3.1) and has covered the spread in just 20% of its last five games.
- Over their last five meetings, UCLA has outscored California by an average of 4.6 points -- a real head-to-head trend worth weighing against the model's lean toward California here.
Per-model breakdown (home win probability)
Logistic Regression71%
Random Forest79%
XGBoost73%
LightGBM78%
Extra Trees76%
Final: California 24 - 45 UCLA — moneyline pick was incorrect, spread pick did not cover the spread
Western Kentucky @ Nevada
Sat Sep 5, 10:30 PM EDT
Model: Nevada by -2.4 (44% win prob) · Market: Western Kentucky by 1.5
Moneyline pick: Western Kentucky (56% win prob) LOW Confidence†
Spread pick: Western Kentucky -1.5 (-110) LOW Confidence†
The market favors Western Kentucky by 1.5, but the model favors Nevada instead, by 2.4 -- a swing of less than a point, so this stays a low-confidence, essentially coin-flip pick toward Western Kentucky on the spread.
- Western Kentucky has the better recent scoring margin (5.2 vs Nevada's -2.8, last four games) and has covered the spread in 80% of its last five games.
- Western Kentucky also rates ahead once schedule strength is factored in, a 2.6-point opponent-adjusted margin edge.
- With the model and market numbers this close, this is a genuinely marginal call, not a real edge in either direction.
Per-model breakdown (home win probability)
Logistic Regression44%
Random Forest43%
XGBoost42%
LightGBM46%
Extra Trees46%
Final: Nevada 49 - 14 Western Kentucky — moneyline pick was incorrect, spread pick did not cover the spread
Morgan State @ Arizona State
Sat Sep 5, 10:00 PM EDT
Model: Arizona State by +39.9 (92% win prob) · Market: Arizona State by 43.5
Moneyline pick: Arizona State (92% win prob) HIGH Confidence†
Spread pick: Morgan State +43.5 (-110*) MEDIUM Confidence†
Morgan State has only 19 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 3.6-point gap between the model's margin and the market's line.
Arizona State is favored by 43.5 in the market, and the model's margin sits close behind at 39.9 -- a 3.6-point gap that makes Morgan State the technical pick, at medium confidence (59% cover probability).
- Arizona State still rates as clearly the better team: a 74% ELO‡ win probability, a 26.2-point opponent-adjusted margin advantage, and the edge in yards per play (6.4 vs 3.2).
- Morgan State has covered the spread in 60% of its last five games, though it has only 19 tracked games in the database as a recent FBS entrant -- a thin sample behind that trend.
- The gap between the model and market numbers here is small, matching the modest confidence tier on this pick.
Per-model breakdown (home win probability)
Logistic Regression96%
Random Forest97%
XGBoost98%
LightGBM99%
Extra Trees95%
Final: Arizona State 70 - 7 Morgan State — moneyline pick was correct, spread pick did not cover the spread
UNLV @ Hawai'i
Sat Sep 5, 10:00 PM EDT
Model: Hawai'i by -3.2 (47% win prob) · Market: UNLV by 3.0
Moneyline pick: UNLV (53% win prob) LOW Confidence†
Spread pick: UNLV -3.0 (-115) LOW Confidence†
The market favors UNLV by 3.0, and the model's own margin lands almost exactly there, with Hawai'i favored by just 3.2 -- essentially no edge, so this pick carries only a 50% cover probability at low confidence.
- UNLV has the better recent scoring margin (-1.2 vs Hawai'i's -3.5, last four games) and the edge in yards per play (5.4 vs 6.3 favors UNLV's opponent-relative efficiency).
- Over their last five meetings, UNLV has outscored Hawai'i by an average of 12.4 points, and UNLV carries a 50-point ELO‡ edge (1519 vs 1569).
- With the model and market numbers this close, there's no real lean here in either direction.
Per-model breakdown (home win probability)
Logistic Regression44%
Random Forest43%
XGBoost49%
LightGBM48%
Extra Trees50%
Final: Hawai'i 6 - 21 UNLV — moneyline pick was correct, spread pick covered the spread
Central Michigan @ New Mexico
Sat Sep 5, 10:00 PM EDT
Model: New Mexico by +11.3 (81% win prob) · Market: New Mexico by 10.8
Moneyline pick: New Mexico (81% win prob) HIGH Confidence†
Spread pick: New Mexico -10.8 (-110) LOW Confidence†
New Mexico is favored by 10.8 in the market, and the model's margin lands almost exactly there at 11.3 -- essentially no edge, so this pick carries only a 51% cover probability at low confidence.
- New Mexico has the better recent scoring margin (5.8 vs Central Michigan's -3.5, last four games) and a 6.7-point opponent-adjusted margin edge.
- Central Michigan has actually posted the better yards-per-play number recently (5.0 vs 4.7) and has covered the spread in 60% of its last five games.
- With the model and market numbers this close, there's no real lean here in either direction.
Per-model breakdown (home win probability)
Logistic Regression80%
Random Forest76%
XGBoost78%
LightGBM77%
Extra Trees78%
Final: New Mexico 38 - 7 Central Michigan — moneyline pick was correct, spread pick covered the spread
Mississippi Valley State @ Sacramento State
Sat Sep 5, 10:00 PM EDT
Model: Sacramento State by +27.6 (88% win prob) · Market: Sacramento State by 37.5
Moneyline pick: Sacramento State (88% win prob) HIGH Confidence†
Spread pick: Mississippi Valley State +37.5 (-110*) HIGH Confidence†
Mississippi Valley State has only 2 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 9.9-point gap between the model's margin and the market's line.
Sacramento State is favored by 37.5 in the market, but the model's margin has it by 27.6 -- a 9.9-point gap that makes Mississippi Valley State the pick, at high confidence (73% cover probability).
- Sacramento State has the better recent scoring margin (-12.0 vs -56.5, last four games), though pre-game ELO‡ gives it only a 55% win probability.
- Mississippi Valley State actually rates ahead once schedule strength is factored in (-11.0 vs 0.0) and carries a 33-point ELO edge (1466 vs 1499).
- Mississippi Valley State has only 2 tracked games in the database as a recent FBS entrant, an extremely thin sample -- both this and Sacramento State's own limited history make the market's 37.5-point line look overconfident.
Per-model breakdown (home win probability)
Logistic Regression85%
Random Forest83%
XGBoost92%
LightGBM97%
Extra Trees87%
Final: Sacramento State 52 - 0 Mississippi Valley State — moneyline pick was correct, spread pick did not cover the spread
Northern Arizona @ Arizona
Sat Sep 5, 9:30 PM EDT
Model: Arizona by +24.6 (90% win prob) · Market: Arizona by 31.5
Moneyline pick: Arizona (90% win prob) HIGH Confidence†
Spread pick: Northern Arizona +31.5 (-110*) MEDIUM Confidence†
Northern Arizona has only 15 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 6.9-point gap between the model's margin and the market's line.
Arizona is favored by 31.5 in the market, but the model's margin has it by 24.6 -- a 6.9-point gap that makes Northern Arizona the pick, at medium confidence (67% cover probability).
- Arizona still rates as clearly the better team: a 75% ELO‡ win probability, a 13.0-point opponent-adjusted margin advantage, and the edge in recent turnover margin (2.5 vs 0.0).
- Northern Arizona has actually faced tougher recent competition than Arizona (average opponent rating +7.0 vs +4.7).
- Northern Arizona has only 15 tracked games in the database as a recent FBS entrant -- its recent-form numbers are a small sample, likely contributing to this gap.
Per-model breakdown (home win probability)
Logistic Regression92%
Random Forest94%
XGBoost93%
LightGBM91%
Extra Trees89%
Final: Arizona 35 - 7 Northern Arizona — moneyline pick was correct, spread pick covered the spread
Portland State @ San Diego State
Sat Sep 5, 9:30 PM EDT
Model: San Diego State by +37.8 (92% win prob) · Market: San Diego State by 40.5
Moneyline pick: San Diego State (92% win prob) HIGH Confidence†
Spread pick: Portland State +40.5 (-110*) LOW Confidence†
San Diego State is favored by 40.5 in the market, and the model's margin lands close behind at 37.8 -- a 2.7-point gap that makes Portland State the technical pick, at low confidence given how thin that edge is (57% cover probability).
- San Diego State still rates as clearly the better team: a 74% ELO‡ win probability, a 19.6-point opponent-adjusted margin advantage, and the better recent scoring margin (6.0 vs -42.8).
- Portland State has covered the spread in 60% of its last five games and has actually faced tougher recent competition than San Diego State (+1.8 vs -1.0).
- With the model and market numbers this close, there's not a strong lean here in either direction.
Per-model breakdown (home win probability)
Logistic Regression97%
Random Forest97%
XGBoost99%
LightGBM99%
Extra Trees95%
Final: San Diego State 53 - 20 Portland State — moneyline pick was correct, spread pick covered the spread
Mercyhurst @ New Mexico State
Sat Sep 5, 9:00 PM EDT
Model: New Mexico State by +15.9 (69% win prob) · Market: New Mexico State by 28.5
Moneyline pick: New Mexico State (69% win prob) MEDIUM Confidence†
Spread pick: Mercyhurst +28.5 (-110*) HIGH Confidence†
Mercyhurst has only 0 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 12.6-point gap between the model's margin and the market's line.
New Mexico State is favored by 28.5 in the market, but the model's margin has it by just 15.9 -- a 12.6-point gap that makes Mercyhurst the pick, at high confidence (79% cover probability).
- Mercyhurst actually rates ahead once schedule strength is factored in, a 17.0-point opponent-adjusted margin edge (-17.0 vs 0.0), and carries an 87-point ELO‡ advantage (1413 vs 1500).
- New Mexico State has covered the spread in 80% of its last five games, the one real factor working against this pick.
- Mercyhurst has only 0 tracked games in the database as a recent FBS entrant -- essentially no usable recent-form history, so this large gap is more about missing data on Mercyhurst than a fully confident read.
Per-model breakdown (home win probability)
Logistic Regression55%
Random Forest68%
XGBoost70%
LightGBM73%
Extra Trees63%
Final: New Mexico State 51 - 14 Mercyhurst — moneyline pick was correct, spread pick did not cover the spread
Utah Tech @ No. 14 BYU
Sat Sep 5, 8:00 PM EDT
Model: BYU by +32.9 (90% win prob) · Market: BYU by 51.5
Moneyline pick: BYU (90% win prob) HIGH Confidence†
Spread pick: Utah Tech +51.5 (-110*) HIGH Confidence†
Utah Tech has only 3 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 18.6-point gap between the model's margin and the market's line.
BYU is a massive favorite (51.5 in the market), and the model still likes BYU big at +32.9 -- an 18.6-point gap that makes Utah Tech the technical pick, at high confidence (88% cover probability).
- BYU still rates as clearly the better team: a 79% ELO‡ win probability, a 169-point ELO edge (1645 vs 1476), and tougher recent competition (+6.8 vs -0.3).
- Utah Tech has covered the spread in 0% of its one tracked game and has only 3 tracked games in the database overall as a recent FBS entrant -- an extremely thin sample, likely the main driver of this large gap rather than any real value.
- BYU has outscored opponents by 2.2 points per game on average recently, against Utah Tech's -35.0.
Per-model breakdown (home win probability)
Logistic Regression94%
Random Forest92%
XGBoost96%
LightGBM95%
Extra Trees90%
Final: BYU 63 - 7 Utah Tech — moneyline pick was correct, spread pick did not cover the spread
Hampton @ Maryland
Sat Sep 5, 8:00 PM EDT
Model: Maryland by +28.3 (89% win prob) · Market: Maryland by 45.5
Moneyline pick: Maryland (89% win prob) HIGH Confidence†
Spread pick: Hampton +45.5 (-110*) HIGH Confidence†
Hampton has only 5 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 17.2-point gap between the model's margin and the market's line.
Maryland is favored by 45.5 in the market, but the model's margin has it by 28.3 -- a 17.2-point gap that makes Hampton the pick, at high confidence (86% cover probability).
- Maryland still projects ahead: a 55% ELO‡ win probability and tougher recent competition (+6.1 vs -0.9), though Hampton actually carries a 31-point ELO edge (1463 vs 1494).
- Hampton has covered the spread in just 20% of its last five games and has only 5 tracked games in the database as a recent FBS entrant -- an extremely thin sample, likely the main driver of this gap.
- Maryland has outscored opponents by an average of -17.0 points recently, better than Hampton's -41.8, but not by enough to support a 45.5-point market spread.
Per-model breakdown (home win probability)
Logistic Regression89%
Random Forest88%
XGBoost95%
LightGBM95%
Extra Trees85%
Final: Maryland 62 - 0 Hampton — moneyline pick was correct, spread pick did not cover the spread
South Dakota State @ Northwestern
Sat Sep 5, 8:00 PM EDT
Model: Northwestern by +17.4 (83% win prob) · Market: Northwestern by 13.5
Moneyline pick: Northwestern (83% win prob) HIGH Confidence†
Spread pick: Northwestern -13.5 (-109) MEDIUM Confidence†
South Dakota State has only 10 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 3.9-point gap between the model's margin and the market's line.
Northwestern is favored by 13.5 in the market, and the model likes Northwestern more, by 17.4 -- a 3.9-point gap in Northwestern's favor, making Northwestern the pick at medium confidence (60% cover probability).
- Northwestern has the edge in recent scoring margin (5.2 vs -4.0, last four games) and rates ahead once schedule strength is factored in (+4.5 vs 0.0).
- South Dakota State has covered the spread in 80% of its last five games, though it has only 10 tracked games in the database as a recent FBS entrant -- a thin sample behind that trend.
- Pre-game ELO‡ gives Northwestern a 65% win probability, a real but not overwhelming edge.
Per-model breakdown (home win probability)
Logistic Regression82%
Random Forest77%
XGBoost84%
LightGBM83%
Extra Trees76%
Final: Northwestern 34 - 18 South Dakota State — moneyline pick was correct, spread pick covered the spread
Lamar @ Louisiana
Sat Sep 5, 8:00 PM EDT
Model: Louisiana by +22.0 (89% win prob) · Market: Louisiana by 20.5
Moneyline pick: Louisiana (89% win prob) HIGH Confidence†
Spread pick: Louisiana -20.5 (-112) LOW Confidence†
Lamar has only 16 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 1.5-point gap between the model's margin and the market's line.
Louisiana is favored by 20.5 in the market, and the model's margin lands close behind at 22.0 -- a 1.5-point gap that makes Louisiana the technical pick, though this is a low-confidence, essentially coin-flip edge.
- Louisiana has the edge in recent scoring margin (0.8 vs Lamar's -26.2, last four games) and a 12.1-point opponent-adjusted margin advantage.
- Lamar has only 16 tracked games in the database as a recent FBS entrant, a thin sample behind its numbers.
- With the model and market numbers this close, there's no strong lean here beyond what the market already reflects.
Per-model breakdown (home win probability)
Logistic Regression87%
Random Forest93%
XGBoost94%
LightGBM91%
Extra Trees90%
Final: Louisiana 38 - 7 Lamar — moneyline pick was correct, spread pick covered the spread
Florida Atlantic @ Florida
Sat Sep 5, 7:45 PM EDT
Model: Florida by +15.2 (80% win prob) · Market: Florida by 26.5
Moneyline pick: Florida (80% win prob) HIGH Confidence†
Spread pick: Florida Atlantic +26.5 (-108) HIGH Confidence†
Florida is favored by 26.5 in the market, but the model's margin has Florida by 15.2 -- an 11.3-point gap that makes Florida Atlantic the pick, at high confidence (76% cover probability).
- Florida Atlantic has covered the spread in 80% of its last five games, the strongest recent ATS trend in this matchup.
- Florida still projects as the better team: pre-game ELO‡ gives it a 70% win probability, and a 12.1-point opponent-adjusted margin edge (+5.9 vs -6.2), backed by tougher recent competition (+10.1 average opponent rating vs -0.1).
- Florida Atlantic has actually outscored its own opponents by a closer margin recently than the gap suggests (-8.5 vs Florida's -10.5 over the last four games).
Per-model breakdown (home win probability)
Logistic Regression79%
Random Forest75%
XGBoost79%
LightGBM81%
Extra Trees68%
Final: Florida 66 - 21 Florida Atlantic — moneyline pick was correct, spread pick did not cover the spread
Clemson @ No. 11 LSU
Sat Sep 5, 7:30 PM EDT
Model: LSU by +2.0 (57% win prob) · Market: LSU by 10.0
Moneyline pick: LSU (57% win prob) LOW Confidence†
Spread pick: Clemson +10.0 (-110) HIGH Confidence†
LSU is favored by 10.0 in the market, but the model's margin has LSU by just 2.0 -- an 8.0-point gap that makes Clemson the pick, at high confidence (69% cover probability).
- Clemson carries a 38-point ELO‡ edge over LSU (1564 vs 1602) and has the better recent scoring margin (9.5 vs -0.8, last four games).
- LSU still rates marginally ahead once schedule strength is factored in, a 1.5-point opponent-adjusted margin edge (+6.7 vs +5.2), and has converted third downs at a higher rate recently (35% vs 29%).
- Pre-game ELO sees this as close to a coin flip, giving LSU only a 54% win probability -- not the kind of gap that supports a 10-point market spread.
Per-model breakdown (home win probability)
Logistic Regression58%
Random Forest54%
XGBoost53%
LightGBM56%
Extra Trees53%
Final: LSU 51 - 10 Clemson — moneyline pick was correct, spread pick did not cover the spread
UL Monroe @ Mississippi State
Sat Sep 5, 7:30 PM EDT
Model: Mississippi State by +18.1 (86% win prob) · Market: Mississippi State by 28.5
Moneyline pick: Mississippi State (86% win prob) HIGH Confidence†
Spread pick: UL Monroe +28.5 (-111) HIGH Confidence†
Mississippi State is favored by 28.5 in the market, but the model's margin has it by 18.1 -- a 10.4-point gap that makes UL Monroe the pick, at high confidence (74% cover probability).
- Mississippi State still rates as clearly the stronger team: a 71% ELO‡ win probability, a 16.4-point opponent-adjusted margin edge (+3.9 vs -12.4), a 89-point ELO advantage, and tougher recent competition (+9.1 vs -3.4).
- UL Monroe has actually outscored its opponents by a closer margin recently than the gap implies, -14.2 versus Mississippi State's -18.8 over the last four games.
- The market's 28.5-point spread looks wider than the underlying opponent-adjusted numbers support, which is what's pulling the pick toward UL Monroe.
Per-model breakdown (home win probability)
Logistic Regression84%
Random Forest87%
XGBoost89%
LightGBM91%
Extra Trees80%
Final: Mississippi State 62 - 13 UL Monroe — moneyline pick was correct, spread pick did not cover the spread
VMI @ Virginia Tech
Sat Sep 5, 7:30 PM EDT
Model: Virginia Tech by +45.9 (91% win prob) · Market: Virginia Tech by 55.5
Moneyline pick: Virginia Tech (91% win prob) HIGH Confidence†
Spread pick: VMI +55.5 (-110*) HIGH Confidence†
VMI has only 17 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 9.6-point gap between the model's margin and the market's line.
Virginia Tech is a heavy favorite (55.5 in the market), and the model still likes Virginia Tech big at +45.9 -- a 9.6-point gap that makes VMI the technical pick, at high confidence (73% cover probability).
- Virginia Tech still rates as clearly the better team: a 60% ELO‡ win probability, a 21.3-point opponent-adjusted margin advantage, and much tougher recent competition (+15.7 vs -4.4).
- VMI has covered the spread in just 20% of its last five games and has only 17 tracked games in the database as a recent FBS entrant -- a thin sample behind its numbers.
- The market's 55.5-point spread is wider than the model's own rating gap supports, which is the core driver of this pick.
Per-model breakdown (home win probability)
Logistic Regression94%
Random Forest95%
XGBoost99%
LightGBM99%
Extra Trees92%
Final: Virginia Tech 73 - 3 VMI — moneyline pick was correct, spread pick did not cover the spread
Western Michigan @ No. 16 Michigan
Sat Sep 5, 7:30 PM EDT
Model: Michigan by +10.8 (76% win prob) · Market: Michigan by 27.5
Moneyline pick: Michigan (76% win prob) HIGH Confidence†
Spread pick: Western Michigan +27.5 (-105) HIGH Confidence†
Michigan is favored by 27.5 in the market, but the model's margin has it by just 10.8 -- a 16.7-point gap that makes Western Michigan the pick, at high confidence (85% cover probability).
- Michigan still rates ahead once schedule strength is factored in, an 11.6-point opponent-adjusted margin edge (+10.2 vs -1.4), and pre-game ELO‡ gives it a 61% win probability.
- Western Michigan has actually posted the better recent scoring margin (17.8 vs Michigan's -1.2, last four games) and the edge in yards per play (6.1 vs 5.2).
- Over their last three meetings, Michigan has outscored Western Michigan by an average of 34.3 points -- a real head-to-head trend, though not one that closes this large a gap against a 27.5-point market spread.
Per-model breakdown (home win probability)
Logistic Regression78%
Random Forest66%
XGBoost81%
LightGBM75%
Extra Trees61%
Final: Michigan 13 - 12 Western Michigan — moneyline pick was correct, spread pick covered the spread
Northwestern State @ Louisiana Tech
Sat Sep 5, 7:30 PM EDT
Model: Louisiana Tech by +37.8 (92% win prob) · Market: Louisiana Tech by 44.5
Moneyline pick: Louisiana Tech (92% win prob) HIGH Confidence†
Spread pick: Northwestern State +44.5 (-110*) MEDIUM Confidence†
Louisiana Tech is a massive favorite (44.5 in the market), and the model still likes Louisiana Tech big at +37.8 -- a 6.7-point gap that makes Northwestern State the technical pick, at medium confidence (66% cover probability).
- Louisiana Tech still rates as clearly the better team: a 73% ELO‡ win probability, a 28.3-point opponent-adjusted margin advantage, and a much higher third-down conversion rate (43% vs 16%).
- Northwestern State has covered the spread in just 20% of its last five games.
- The market's 44.5-point spread runs slightly ahead of what Louisiana Tech's own strong underlying ratings support.
Per-model breakdown (home win probability)
Logistic Regression98%
Random Forest96%
XGBoost98%
LightGBM100%
Extra Trees94%
Final: Louisiana Tech 80 - 6 Northwestern State — moneyline pick was correct, spread pick did not cover the spread
Missouri State @ No. 8 Texas A&M
Sat Sep 5, 7:00 PM EDT
Model: Texas A&M by +22.1 (87% win prob) · Market: Texas A&M by 40.5
Moneyline pick: Texas A&M (87% win prob) HIGH Confidence†
Spread pick: Missouri State +40.5 (-110) HIGH Confidence†
Texas A&M is favored by 40.5 in the market, but the model's margin has it by 22.1 -- an 18.4-point gap that makes Missouri State the pick, at high confidence (88% cover probability).
- Texas A&M still rates as clearly the better team: a 79% ELO‡ win probability, a 17.5-point opponent-adjusted margin advantage (+12.0 vs -5.6), and tougher recent competition (+3.8 vs -5.0).
- Missouri State has actually posted the better yards-per-play number recently, 6.7 versus Texas A&M's 5.7 over the last four games.
- The market's 40.5-point spread is wider than any of the model's underlying ratings support, which is the core driver of this pick.
Per-model breakdown (home win probability)
Logistic Regression89%
Random Forest84%
XGBoost91%
LightGBM89%
Extra Trees80%
Final: Texas A&M 50 - 0 Missouri State — moneyline pick was correct, spread pick did not cover the spread
Austin Peay @ Vanderbilt
Sat Sep 5, 7:00 PM EDT
Model: Vanderbilt by +29.6 (91% win prob) · Market: Vanderbilt by 29.5
Moneyline pick: Vanderbilt (91% win prob) HIGH Confidence†
Spread pick: Vanderbilt -29.5 (-103) LOW Confidence†
Vanderbilt is favored by 29.5 in the market, and the model's own margin lands almost exactly there at 29.6 -- essentially no edge, so this pick carries only a 50% cover probability at low confidence.
- Vanderbilt has the clear recent-form edge: a 12.2 vs -20.2 scoring margin, 531.2 vs 246.0 total yards, and 8.0 vs 3.7 yards per play over the last four games.
- Over their last two meetings, Vanderbilt has outscored Austin Peay by an average of 37.5 points.
- With the model and market numbers this close together, there's genuinely no lean here -- this is a pass-quality spread, not a real edge in either direction.
Per-model breakdown (home win probability)
Logistic Regression96%
Random Forest93%
XGBoost97%
LightGBM97%
Extra Trees93%
Final: Vanderbilt 28 - 9 Austin Peay — moneyline pick was correct, spread pick did not cover the spread
Abilene Christian @ No. 12 Texas Tech
Sat Sep 5, 7:00 PM EDT
Model: Texas Tech by +33.7 (91% win prob) · Market: Texas Tech by 42.5
Moneyline pick: Texas Tech (91% win prob) HIGH Confidence†
Spread pick: Abilene Christian +42.5 (-110*) HIGH Confidence†
Texas Tech is favored by 42.5 in the market, but the model's margin has it by 33.7 -- an 8.8-point gap that makes Abilene Christian the pick, at high confidence (71% cover probability).
- Texas Tech still rates as clearly the better team: an 85% ELO‡ win probability, a 235-point ELO edge (1683 vs 1449), and a 30.2-point opponent-adjusted margin advantage.
- Abilene Christian has covered the spread in 80% of its last five games, the strongest recent ATS trend in this matchup.
- Texas Tech has outscored opponents by 23.0 points per game on average recently, against Abilene Christian's -19.5.
Per-model breakdown (home win probability)
Logistic Regression97%
Random Forest95%
XGBoost97%
LightGBM98%
Extra Trees93%
Final: Texas Tech 33 - 10 Abilene Christian — moneyline pick was correct, spread pick covered the spread
Nicholls @ Kansas State
Sat Sep 5, 7:00 PM EDT
Model: Kansas State by +29.1 (90% win prob) · Market: Kansas State by 40.5
Moneyline pick: Kansas State (90% win prob) HIGH Confidence†
Spread pick: Nicholls +40.5 (-110*) HIGH Confidence†
Kansas State is favored by 40.5 in the market, but the model's margin has it by 29.1 -- an 11.4-point gap that makes Nicholls the pick, at high confidence (76% cover probability).
- Kansas State still rates as clearly the better team: a 75% ELO‡ win probability, a 21.7-point opponent-adjusted margin advantage, and tougher recent competition (+6.2 vs -0.5).
- Nicholls has covered the spread in 80% of its last five games, the strongest recent ATS trend in this matchup.
- Kansas State's recent scoring margin edge is real (-2.2 vs -20.2, last four games) but doesn't close a 40.5-point market spread.
Per-model breakdown (home win probability)
Logistic Regression94%
Random Forest94%
XGBoost94%
LightGBM90%
Extra Trees91%
Final: Kansas State 71 - 3 Nicholls — moneyline pick was correct, spread pick did not cover the spread
Idaho State @ Utah State
Sat Sep 5, 7:00 PM EDT
Model: Utah State by +11.7 (81% win prob) · Market: Utah State by 14.0
Moneyline pick: Utah State (81% win prob) HIGH Confidence†
Spread pick: Idaho State +14.0 (-110*) LOW Confidence†
Utah State is favored by 14.0 in the market, and the model's margin lands close behind at 11.7 -- a 2.3-point gap that makes Idaho State the technical pick, at low confidence given how thin that edge is (56% cover probability).
- Utah State still rates ahead: a 64% ELO‡ win probability, the better recent scoring margin (-1.5 vs -22.5, last four games), and a 3.2-point opponent-adjusted margin edge.
- Over their last three meetings, Utah State has outscored Idaho State by an average of 36.0 points -- a lopsided head-to-head history working against this pick.
- Idaho State has covered the spread in 80% of its last five games, the main factor keeping this close.
Per-model breakdown (home win probability)
Logistic Regression84%
Random Forest72%
XGBoost78%
LightGBM85%
Extra Trees73%
Final: Utah State 17 - 29 Idaho State — moneyline pick was incorrect, spread pick covered the spread
Arkansas State @ Memphis
Sat Sep 5, 7:00 PM EDT
Model: Memphis by +9.4 (75% win prob) · Market: Memphis by 11.2
Moneyline pick: Memphis (75% win prob) HIGH Confidence†
Spread pick: Arkansas State +11.2 (-110) LOW Confidence†
Memphis is favored by 11.2 in the market, and the model's margin lands close behind at 9.4 -- a 1.9-point gap that makes Arkansas State the technical pick, though this is a low-confidence, near-coin-flip edge.
- Memphis still rates as the better team once schedule strength is accounted for, a 12.3-point opponent-adjusted margin edge, and carries an 89-point ELO‡ advantage.
- Arkansas State has actually posted the better recent scoring margin (-0.8 vs Memphis's -8.2, last four games) and has covered the spread in 60% of its last five games.
- With the model and market numbers this close, there's no strong lean here in either direction.
Per-model breakdown (home win probability)
Logistic Regression70%
Random Forest75%
XGBoost72%
LightGBM72%
Extra Trees68%
Final: Memphis 42 - 24 Arkansas State — moneyline pick was correct, spread pick did not cover the spread
Houston Christian @ Rice
Sat Sep 5, 7:00 PM EDT
Model: Rice by +29.0 (88% win prob) · Market: Rice by 22.5
Moneyline pick: Rice (88% win prob) HIGH Confidence†
Spread pick: Rice -22.5 (-114) MEDIUM Confidence†
Houston Christian has only 14 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 6.5-point gap between the model's margin and the market's line.
The market favors Rice by 22.5, and the model favors Rice even more, by 29.0 -- a 6.5-point gap in Rice's favor, making Rice the pick at medium confidence (66% cover probability).
- Rice rates ahead once schedule strength is factored in, an 11.8-point opponent-adjusted margin edge (-8.4 vs -20.2), and has the edge in recent yards per play (4.0 vs 3.7).
- Houston Christian actually carries a 24-point ELO‡ edge (1427 vs 1451), despite pre-game ELO giving Rice only a 56% win probability.
- Houston Christian has only 14 tracked games in the database as a recent FBS entrant, a thin sample behind its numbers.
Per-model breakdown (home win probability)
Logistic Regression84%
Random Forest88%
XGBoost94%
LightGBM93%
Extra Trees86%
Final: Rice 31 - 3 Houston Christian — moneyline pick was correct, spread pick covered the spread
Florida International @ South Florida
Sat Sep 5, 7:00 PM EDT
Model: South Florida by +23.3 (87% win prob) · Market: South Florida by 14.5
Moneyline pick: South Florida (87% win prob) HIGH Confidence†
Spread pick: South Florida -14.5 (-110) HIGH Confidence†
The market favors South Florida by 14.5, and the model favors South Florida even more, by 23.3 -- an 8.8-point gap in South Florida's favor, making South Florida the pick at high confidence (71% cover probability).
- South Florida has the edge in recent scoring margin (15.5 vs 4.0, last four games) and rates ahead once schedule strength is factored in (+9.0 vs -4.5).
- Florida International has covered the spread in 80% of its last five games, the one factor working against this pick.
- Pre-game ELO‡ gives South Florida a 73% win probability, backing the model's lean here.
Per-model breakdown (home win probability)
Logistic Regression91%
Random Forest86%
XGBoost89%
LightGBM89%
Extra Trees81%
Final: South Florida 19 - 9 Florida International — moneyline pick was correct, spread pick did not cover the spread
Murray State @ Middle Tennessee
Sat Sep 5, 7:00 PM EDT
Model: Middle Tennessee by +26.5 (90% win prob) · Market: Middle Tennessee by 20.5
Moneyline pick: Middle Tennessee (90% win prob) HIGH Confidence†
Spread pick: Middle Tennessee -20.5 (-122) MEDIUM Confidence†
The market favors Middle Tennessee by 20.5, and the model favors Middle Tennessee even more, by 26.5 -- a 6.0-point gap in Middle Tennessee's favor, making Middle Tennessee the pick at medium confidence (65% cover probability).
- Middle Tennessee has the edge in recent scoring margin (-5.2 vs Murray State's -36.0, last four games) and a 12.2-point opponent-adjusted margin advantage.
- Murray State actually carries a 38-point ELO‡ edge over Middle Tennessee (1395 vs 1433), despite that scoring-margin gap.
- Murray State has also faced tougher recent competition than Middle Tennessee (-6.2 vs -9.7).
Per-model breakdown (home win probability)
Logistic Regression87%
Random Forest92%
XGBoost96%
LightGBM96%
Extra Trees91%
Final: Middle Tennessee 38 - 14 Murray State — moneyline pick was correct, spread pick covered the spread
Eastern Kentucky @ Jacksonville State
Sat Sep 5, 7:00 PM EDT
Model: Jacksonville State by +31.1 (91% win prob) · Market: Jacksonville State by 20.5
Moneyline pick: Jacksonville State (91% win prob) HIGH Confidence†
Spread pick: Jacksonville State -20.5 (-113) HIGH Confidence†
The market favors Jacksonville State by 20.5, and the model favors Jacksonville State even more, by 31.1 -- a 10.6-point gap in Jacksonville State's favor, making Jacksonville State the pick at high confidence (75% cover probability).
- Jacksonville State has the edge in recent scoring margin (-5.8 vs Eastern Kentucky's -36.2, last four games) and carries a 71-point ELO‡ advantage.
- Eastern Kentucky actually rates ahead once schedule strength is factored in (-26.0 vs -15.2), the one real factor cutting against this pick.
- Jacksonville State has also faced tougher recent competition than Eastern Kentucky (+4.1 vs +0.7).
Per-model breakdown (home win probability)
Logistic Regression92%
Random Forest93%
XGBoost97%
LightGBM97%
Extra Trees93%
Final: Jacksonville State 49 - 7 Eastern Kentucky — moneyline pick was correct, spread pick covered the spread
Charleston Southern @ Georgia Southern
Sat Sep 5, 7:00 PM EDT
Model: Georgia Southern by +30.1 (91% win prob) · Market: Georgia Southern by 25.5
Moneyline pick: Georgia Southern (91% win prob) HIGH Confidence†
Spread pick: Georgia Southern -25.5 (-110) MEDIUM Confidence†
The market favors Georgia Southern by 25.5, and the model favors Georgia Southern even more, by 30.1 -- a 4.6-point gap in Georgia Southern's favor, making Georgia Southern the pick at medium confidence (61% cover probability).
- Georgia Southern has a large recent scoring margin edge (-1.5 vs Charleston Southern's -34.5, last four games) and the edge in yards per play (6.3 vs 3.1).
- Charleston Southern has actually faced tougher recent competition than Georgia Southern (+6.4 vs -4.6), part of why this edge stays moderate.
- Georgia Southern also holds a big total-yards edge recently, 460.8 versus Charleston Southern's 159.0 per game.
Per-model breakdown (home win probability)
Logistic Regression90%
Random Forest95%
XGBoost97%
LightGBM98%
Extra Trees96%
Final: Georgia Southern 31 - 0 Charleston Southern — moneyline pick was correct, spread pick covered the spread
SE Louisiana @ South Alabama
Sat Sep 5, 7:00 PM EDT
Model: South Alabama by +24.2 (90% win prob) · Market: South Alabama by 12.5
Moneyline pick: South Alabama (90% win prob) HIGH Confidence†
Spread pick: South Alabama -12.5 (-115) HIGH Confidence†
The market favors South Alabama by 12.5, and the model favors South Alabama by considerably more, 24.2 -- an 11.7-point gap in South Alabama's favor, making South Alabama the pick at high confidence (77% cover probability).
- South Alabama has a large recent scoring margin edge (-3.2 vs SE Louisiana's -36.8, last four games) and the edge in turnover margin (0.2 vs -1.2).
- SE Louisiana has actually faced tougher recent competition than South Alabama (+0.8 vs -6.4).
- SE Louisiana has covered the spread in just 40% of its last five games.
Per-model breakdown (home win probability)
Logistic Regression88%
Random Forest95%
XGBoost96%
LightGBM96%
Extra Trees92%
Final: South Alabama 39 - 14 SE Louisiana — moneyline pick was correct, spread pick covered the spread
Sam Houston @ Troy
Sat Sep 5, 7:00 PM EDT
Model: Troy by +10.1 (83% win prob) · Market: Troy by 16.5
Moneyline pick: Troy (83% win prob) HIGH Confidence†
Spread pick: Sam Houston +16.5 (-109) MEDIUM Confidence†
Troy is favored by 16.5 in the market, but the model's margin has it by 10.1 -- a 6.4-point gap that makes Sam Houston the pick, at medium confidence (66% cover probability).
- Troy still rates as the better team: a 75% ELO‡ win probability, a 130-point ELO edge (1520 vs 1391), and the better recent scoring margin (0.2 vs -11.8).
- Sam Houston has actually posted the better yards-per-play number recently, 4.4 versus Troy's 4.8.
- Sam Houston has also faced tougher recent competition than Troy (-6.6 vs -8.9), part of why this gap stays moderate.
Per-model breakdown (home win probability)
Logistic Regression85%
Random Forest79%
XGBoost75%
LightGBM78%
Extra Trees80%
Final: Troy 24 - 21 Sam Houston — moneyline pick was correct, spread pick covered the spread
Wyoming @ Colorado State
Sat Sep 5, 6:00 PM EDT
Model: Colorado State by -3.4 (58% win prob) · Market: Colorado State by 3.0
Moneyline pick: Colorado State (58% win prob) LOW Confidence†
Spread pick: Wyoming +3.0 (-110) MEDIUM Confidence†
The market favors Colorado State by 3.0, but the model actually favors Wyoming, +3.4 -- a 6.4-point swing that makes Wyoming the pick, at medium confidence (66% cover probability).
- Wyoming has the better recent scoring margin (-16.0 vs Colorado State's -21.0, last four games) and carries a 41-point ELO‡ edge (1407 vs 1448).
- Over their last five meetings, Wyoming has outscored Colorado State by an average of 7.6 points -- a real head-to-head trend backing this pick.
- Wyoming has covered the spread in just 20% of its last five games, the one factor working against this lean.
Per-model breakdown (home win probability)
Logistic Regression52%
Random Forest63%
XGBoost50%
LightGBM45%
Extra Trees62%
Final: Colorado State 35 - 13 Wyoming — moneyline pick was correct, spread pick did not cover the spread
Norfolk State @ Old Dominion
Sat Sep 5, 6:00 PM EDT
Model: Old Dominion by +30.7 (92% win prob) · Market: Old Dominion by 40.5
Moneyline pick: Old Dominion (92% win prob) HIGH Confidence†
Spread pick: Norfolk State +40.5 (-110*) HIGH Confidence†
Norfolk State has only 17 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 9.8-point gap between the model's margin and the market's line.
Old Dominion is favored by 40.5 in the market, but the model's margin has it by 30.7 -- a 9.8-point gap that makes Norfolk State the pick, at high confidence (73% cover probability).
- Old Dominion still rates as clearly the better team: a 79% ELO‡ win probability, a 24.0-point opponent-adjusted margin advantage, and a big edge in total yards per game (461.2 vs 196.5).
- Norfolk State has covered the spread in 0% of its last five games and has only 17 tracked games in the database as a recent FBS entrant -- a thin sample behind its numbers.
- The market's 40.5-point spread is wider than even Old Dominion's own strong underlying ratings support, which is the core driver of this pick.
Per-model breakdown (home win probability)
Logistic Regression97%
Random Forest97%
XGBoost99%
LightGBM99%
Extra Trees98%
Final: Old Dominion 31 - 10 Norfolk State — moneyline pick was correct, spread pick covered the spread
Alcorn State @ Southern Miss
Sat Sep 5, 5:00 PM EDT
Model: Southern Miss by +28.0 (90% win prob) · Market: Southern Miss by 20.5
Moneyline pick: Southern Miss (90% win prob) HIGH Confidence†
Spread pick: Southern Miss -20.5 (-102) HIGH Confidence†
Alcorn State has only 15 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 7.5-point gap between the model's margin and the market's line.
The market favors Southern Miss by 20.5, and the model favors Southern Miss by considerably more, 28.0 -- a 7.5-point gap in Southern Miss's favor, making Southern Miss the pick at high confidence (68% cover probability).
- Southern Miss has a large recent scoring margin edge (-13.8 vs Alcorn State's -45.5, last four games) and a 22.9-point opponent-adjusted margin advantage.
- Alcorn State has covered the spread in 0% of its last five games and has only 15 tracked games in the database as a recent FBS entrant -- a thin sample behind its numbers.
- Alcorn State has actually faced tougher recent competition than Southern Miss (+1.0 vs -4.2), the one factor working against this pick's size.
Per-model breakdown (home win probability)
Logistic Regression91%
Random Forest93%
XGBoost95%
LightGBM96%
Extra Trees88%
Final: Southern Miss 49 - 3 Alcorn State — moneyline pick was correct, spread pick covered the spread
North Alabama @ Arkansas
Sat Sep 5, 4:15 PM EDT
Model: Arkansas by +38.1 (91% win prob) · Market: Arkansas by 40.5
Moneyline pick: Arkansas (91% win prob) HIGH Confidence†
Spread pick: North Alabama +40.5 (-110*) LOW Confidence†
North Alabama has only 7 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 2.4-point gap between the model's margin and the market's line.
Arkansas is a big favorite either way, but the model's margin (+38.1) sits almost on top of the market's own line (+40.5) -- a 2.4-point gap that makes North Alabama the technical pick, though this is a low-confidence, essentially coin-flip edge.
- Arkansas rates as the clearly stronger team once schedule strength is factored in, a 32.8-point opponent-adjusted margin edge (+4.8 vs -27.9), and pre-game ELO‡ gives it a 63% win probability.
- Both teams have struggled against the number lately -- North Alabama and Arkansas have each covered just 20% of their last five games.
- North Alabama has only 7 tracked games in the database as a recent FBS entrant, so its recent-form numbers are a thin sample -- treat this edge as noise more than signal.
Per-model breakdown (home win probability)
Logistic Regression95%
Random Forest96%
XGBoost98%
LightGBM97%
Extra Trees95%
Final: Arkansas 31 - 14 North Alabama — moneyline pick was correct, spread pick covered the spread
Northern Illinois @ No. 22 Iowa
Sat Sep 5, 4:15 PM EDT
Model: Iowa by +22.2 (88% win prob) · Market: Iowa by 31.8
Moneyline pick: Iowa (88% win prob) HIGH Confidence†
Spread pick: Northern Illinois +31.8 (-110) HIGH Confidence†
Iowa is favored by 31.8 in the market, but the model's margin has it by 22.2 -- a 9.5-point gap that makes Northern Illinois the pick, at high confidence (73% cover probability).
- Iowa still rates as clearly the better team: an 82% ELO‡ win probability, a 19.9-point opponent-adjusted margin advantage, and a 197-point ELO edge (1644 vs 1447).
- Iowa has also faced much tougher recent competition than Northern Illinois (+9.1 vs -6.2).
- Iowa's recent scoring margin edge (7.2 vs -4.2, last four games) is real but doesn't fully support a 31.8-point market spread.
Per-model breakdown (home win probability)
Logistic Regression92%
Random Forest89%
XGBoost90%
LightGBM91%
Extra Trees85%
Final: Iowa 40 - 0 Northern Illinois — moneyline pick was correct, spread pick did not cover the spread
Oklahoma State @ Tulsa
Sat Sep 5, 3:45 PM EDT
Model: Tulsa by +4.9 (63% win prob) · Market: Oklahoma State by 13.5
Moneyline pick: Tulsa (63% win prob) MEDIUM Confidence†
Spread pick: Tulsa +13.5 (-105) HIGH Confidence†
The market favors Oklahoma State by 13.5, but the model actually favors Tulsa by 4.9 -- an 18.4-point swing that makes Tulsa the pick, at high confidence (88% cover probability).
- Pre-game ELO‡ backs this lean, giving Tulsa a 62% win probability, and Tulsa has the better recent scoring margin too (-2.0 vs -8.8, last four games).
- Oklahoma State has faced tougher recent competition than Tulsa (average opponent rating +4.1 vs -5.0) and has covered the spread in 80% of its last five games.
- Over their last five meetings, Oklahoma State has outscored Tulsa by an average of 12.2 points -- a real head-to-head trend that cuts against this pick and is worth weighing against the model's lean.
Per-model breakdown (home win probability)
Logistic Regression61%
Random Forest62%
XGBoost57%
LightGBM49%
Extra Trees66%
Final: Tulsa 24 - 10 Oklahoma State — moneyline pick was correct, spread pick covered the spread
Baylor @ Auburn
Sat Sep 5, 3:30 PM EDT
Model: Auburn by +0.7 (67% win prob) · Market: Auburn by 7.2
Moneyline pick: Auburn (67% win prob) MEDIUM Confidence†
Spread pick: Baylor +7.2 (-112) MEDIUM Confidence†
Auburn is favored by 7.2 in the market, but the model's own margin has Auburn by just 0.7 -- a 6.5-point gap that makes Baylor the pick, at medium confidence (66% cover probability).
- Auburn has the edge in recent scoring margin (6.0 vs -7.8, last four games) and in yards per play (6.4 vs 5.4).
- Pre-game ELO‡ sees this as close to even, giving Auburn only a 54% win probability.
- Baylor has faced tougher recent competition than Auburn (average opponent rating +6.8 vs +3.0), even while covering just 20% of its last five games.
Per-model breakdown (home win probability)
Logistic Regression66%
Random Forest67%
XGBoost56%
LightGBM55%
Extra Trees66%
Final: Auburn 17 - 16 Baylor — moneyline pick was correct, spread pick covered the spread
Furman @ No. 20 Tennessee
Sat Sep 5, 3:30 PM EDT
Model: Tennessee by +34.5 (92% win prob) · Market: Tennessee by 46.5
Moneyline pick: Tennessee (92% win prob) HIGH Confidence†
Spread pick: Furman +46.5 (-110*) HIGH Confidence†
Furman has only 16 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 12.0-point gap between the model's margin and the market's line.
Tennessee is favored by 46.5 in the market, but the model's margin has it by 34.5 -- a 12.0-point gap that makes Furman the pick, at high confidence (77% cover probability).
- Tennessee still rates as clearly the better team: a 75% ELO‡ win probability, a 129-point ELO edge (1598 vs 1469), and a 20.5-point opponent-adjusted margin advantage.
- Furman has faced tougher recent competition than Tennessee (average opponent rating +6.9 vs +2.3), and has only 16 tracked games in the database as a recent FBS entrant -- a thin sample behind its recent-form numbers.
- Tennessee has outscored opponents by 7.5 points per game on average recently, against Furman's -40.0.
Per-model breakdown (home win probability)
Logistic Regression96%
Random Forest96%
XGBoost98%
LightGBM99%
Extra Trees94%
Final: Tennessee 56 - 9 Furman — moneyline pick was correct, spread pick did not cover the spread
Texas State @ No. 5 Texas
Sat Sep 5, 3:30 PM EDT
Model: Texas by +16.1 (80% win prob) · Market: Texas by 29.5
Moneyline pick: Texas (80% win prob) HIGH Confidence†
Spread pick: Texas State +29.5 (-110) HIGH Confidence†
Texas is favored by 29.5 in the market, but the model's margin has Texas by 16.1 -- a 13.4-point gap that makes Texas State the pick, at high confidence (80% cover probability).
- Texas still projects as the stronger team: a 71% ELO‡ win probability, a 93-point ELO edge (1659 vs 1566), and an 11.4-point opponent-adjusted margin advantage, against much tougher recent competition (+10.0 vs -8.1).
- Texas State has actually posted the better recent scoring margin of the two, 24.5 versus Texas's 3.5 over the last four games.
- The market's 29.5-point spread doesn't line up with how close these two teams' recent form actually is, which is what's driving the pick toward Texas State.
Per-model breakdown (home win probability)
Logistic Regression74%
Random Forest76%
XGBoost84%
LightGBM85%
Extra Trees69%
Final: Texas 59 - 7 Texas State — moneyline pick was correct, spread pick did not cover the spread
Boston College @ Cincinnati
Sat Sep 5, 3:30 PM EDT
Model: Cincinnati by +11.4 (75% win prob) · Market: Cincinnati by 7.5
Moneyline pick: Cincinnati (75% win prob) MEDIUM Confidence†
Spread pick: Cincinnati -7.5 (-110) MEDIUM Confidence†
Cincinnati is favored by 7.5 in the market, and the model likes Cincinnati slightly more, by 11.4 -- a 3.9-point gap in Cincinnati's favor, making Cincinnati the pick at medium confidence (60% cover probability).
- Cincinnati carries a 59-point ELO‡ advantage over Boston College (1522 vs 1464), and pre-game ELO gives it a 67% win probability.
- Boston College has covered the spread in 80% of its last five games and has actually posted the better recent scoring margin (-6.8 vs Cincinnati's -15.5), a real trend working against this pick.
- Boston College has also faced tougher recent competition than Cincinnati (+5.6 vs +3.9), which is part of why this edge stays modest.
Per-model breakdown (home win probability)
Logistic Regression76%
Random Forest70%
XGBoost72%
LightGBM68%
Extra Trees66%
Final: Cincinnati 34 - 15 Boston College — moneyline pick was correct, spread pick covered the spread
Tulane @ Duke
Sat Sep 5, 3:30 PM EDT
Model: Duke by +0.6 (64% win prob) · Market: Duke by 7.5
Moneyline pick: Duke (64% win prob) MEDIUM Confidence†
Spread pick: Tulane +7.5 (-110) MEDIUM Confidence†
Duke is favored by 7.5 in the market, but the model's margin has Duke by just 0.6 -- a 6.9-point gap that makes Tulane the pick, at medium confidence (67% cover probability).
- The two teams' recent scoring margins are nearly identical (Duke 8.5, Tulane 8.2, last four games), and Tulane actually has the edge in yards per play (5.7 vs 5.2).
- Duke has faced tougher recent competition than Tulane (average opponent rating +9.8 vs +1.0), which is the main thing keeping Duke favored at all.
- Pre-game ELO‡ gives Duke only a 59% win probability -- not the kind of gap that supports a 7.5-point market spread.
Per-model breakdown (home win probability)
Logistic Regression65%
Random Forest60%
XGBoost52%
LightGBM57%
Extra Trees62%
Final: Duke 17 - 3 Tulane — moneyline pick was correct, spread pick did not cover the spread
Boise State @ No. 2 Oregon
Sat Sep 5, 3:30 PM EDT
Model: Oregon by +13.4 (82% win prob) · Market: Oregon by 24.8
Moneyline pick: Oregon (82% win prob) HIGH Confidence†
Spread pick: Boise State +24.8 (-110) HIGH Confidence†
Oregon is favored by 24.8 in the market, but the model's margin has it by 13.4 -- an 11.3-point gap that makes Boise State the pick, at high confidence (76% cover probability).
- Oregon still rates as clearly the better team: a 76% ELO‡ win probability, a 13.1-point opponent-adjusted margin advantage, and much tougher recent competition (+13.6 vs -0.2).
- The two teams' recent scoring margins are actually identical (4.5 each over the last four games).
- Oregon carries a 131-point ELO advantage (1707 vs 1577), but the market's 24.8-point spread outruns what the underlying numbers support.
Per-model breakdown (home win probability)
Logistic Regression83%
Random Forest77%
XGBoost84%
LightGBM82%
Extra Trees73%
Final: Oregon 34 - 27 Boise State — moneyline pick was correct, spread pick covered the spread
Marshall @ No. 18 Penn State
Sat Sep 5, 3:30 PM EDT
Model: Penn State by +17.0 (87% win prob) · Market: Penn State by 23.5
Moneyline pick: Penn State (87% win prob) HIGH Confidence†
Spread pick: Marshall +23.5 (-110) MEDIUM Confidence†
Penn State is favored by 23.5 in the market, but the model's margin has it by 17.0 -- a 6.5-point gap that makes Marshall the pick, at medium confidence (66% cover probability).
- Penn State still rates as the better team: a 72% ELO‡ win probability, a 14.3-point opponent-adjusted margin advantage, and the edge in yards per play (7.1 vs 6.6).
- Penn State has also faced tougher recent competition than Marshall (+4.2 vs -4.9).
- Penn State's recent scoring margin edge (15.2 vs -1.8, last four games) is real but doesn't fully close the gap against a 23.5-point market spread.
Per-model breakdown (home win probability)
Logistic Regression88%
Random Forest84%
XGBoost89%
LightGBM90%
Extra Trees83%
Final: Penn State 45 - 0 Marshall — moneyline pick was correct, spread pick did not cover the spread
The Citadel @ Charlotte
Sat Sep 5, 3:30 PM EDT
Model: Charlotte by +20.8 (83% win prob) · Market: Charlotte by 20.5
Moneyline pick: Charlotte (83% win prob) HIGH Confidence†
Spread pick: Charlotte -20.5 (-114) LOW Confidence†
The Citadel has only 18 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 0.3-point gap between the model's margin and the market's line.
Charlotte is favored by 20.5 in the market, and the model's margin lands almost exactly there at 20.8 -- essentially no edge, so this pick carries only a 51% cover probability at low confidence.
- The Citadel has actually covered the spread in 60% of its last five games and carries a 117-point ELO‡ edge (1359 vs 1476), despite pre-game ELO giving Charlotte only a 43% win probability.
- Charlotte has faced tougher recent competition than The Citadel (+5.3 vs +0.6), the main thing keeping it favored at all.
- The Citadel has only 18 tracked games in the database as a recent FBS entrant -- with the model and market numbers this close, there's no real lean here either way.
Per-model breakdown (home win probability)
Logistic Regression79%
Random Forest74%
XGBoost88%
LightGBM92%
Extra Trees74%
Final: Charlotte 41 - 43 The Citadel — moneyline pick was incorrect, spread pick did not cover the spread
Towson @ Navy
Sat Sep 5, 3:30 PM EDT
Model: Navy by +37.6 (92% win prob) · Market: Navy by 30.5
Moneyline pick: Navy (92% win prob) HIGH Confidence†
Spread pick: Navy -30.5 (-112) HIGH Confidence†
Towson has only 15 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 7.1-point gap between the model's margin and the market's line.
The market favors Navy by 30.5, and the model favors Navy even more, by 37.6 -- a 7.1-point gap in Navy's favor, making Navy the pick at high confidence (67% cover probability).
- Navy has the edge in recent scoring margin (9.2 vs -33.8, last four games) and rates well ahead once schedule strength is factored in (+0.8 vs -11.6).
- Navy also carries a 146-point ELO‡ advantage over Towson (1610 vs 1463).
- Towson has only 15 tracked games in the database as a recent FBS entrant, a thin sample behind its numbers -- consistent with the model seeing a wider gap here than the market's own line does.
Per-model breakdown (home win probability)
Logistic Regression97%
Random Forest94%
XGBoost98%
LightGBM99%
Extra Trees95%
Final: Navy 42 - 15 Towson — moneyline pick was correct, spread pick did not cover the spread
UT Rio Grande Valley @ UTSA
Sat Sep 5, 3:30 PM EDT
Model: UTSA by +25.2 (83% win prob) · Market: UTSA by 28.5
Moneyline pick: UTSA (83% win prob) HIGH Confidence†
Spread pick: UT Rio Grande Valley +28.5 (-110*) MEDIUM Confidence†
UT Rio Grande Valley has only 0 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 3.3-point gap between the model's margin and the market's line.
UTSA is favored by 28.5 in the market, but the model's margin has it by 25.2 -- a 3.3-point gap that makes UT Rio Grande Valley the technical pick, at medium confidence (58% cover probability).
- UTSA still rates ahead once schedule strength is factored in (+1.3 vs 0.0) and carries an 81-point ELO‡ advantage (1581 vs 1500).
- UT Rio Grande Valley has only 0 tracked games in the database as a recent FBS entrant -- essentially no usable recent-form history, so this gap is more about missing data than a real edge.
- Both teams enter on a full 7 days of rest, a neutral factor here.
Per-model breakdown (home win probability)
Logistic Regression70%
Random Forest84%
XGBoost88%
LightGBM87%
Extra Trees80%
Final: UTSA 45 - 16 UT Rio Grande Valley — moneyline pick was correct, spread pick did not cover the spread
Fordham @ North Dakota State
Sat Sep 5, 3:30 PM EDT
Model: North Dakota State by +28.8 (87% win prob) · Market: North Dakota State by 41.5
Moneyline pick: North Dakota State (87% win prob) HIGH Confidence†
Spread pick: Fordham +41.5 (-110*) HIGH Confidence†
North Dakota State has only 8 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 12.7-point gap between the model's margin and the market's line.
North Dakota State is favored by 41.5 in the market, but the model's margin has it by 28.8 -- a 12.7-point gap that makes Fordham the pick, at high confidence (79% cover probability).
- North Dakota State still rates as clearly the better team: a 69% ELO‡ win probability and a massive 57.0-point opponent-adjusted margin advantage (+26.0 vs -31.0).
- Fordham has covered the spread in 40% of its last five games.
- North Dakota State has only 8 tracked games in the database itself as a recent FBS entrant -- both sides of this game carry thin recent-form samples, which is likely inflating the size of the market's own 41.5-point line.
Per-model breakdown (home win probability)
Logistic Regression92%
Random Forest86%
XGBoost90%
LightGBM82%
Extra Trees82%
Final: North Dakota State 38 - 0 Fordham — moneyline pick was correct, spread pick covered the spread
Maine @ App State
Sat Sep 5, 3:30 PM EDT
Model: App State by +21.0 (85% win prob) · Market: App State by 19.5
Moneyline pick: App State (85% win prob) HIGH Confidence†
Spread pick: App State -19.5 (-112) LOW Confidence†
App State is favored by 19.5 in the market, and the model's margin lands close behind at 21.0 -- a 1.5-point gap that makes App State the pick, though this is a low-confidence, essentially coin-flip edge.
- App State rates ahead once schedule strength is factored in, an 11.1-point opponent-adjusted margin edge (-7.5 vs -18.6), and has the edge in yards per play (4.7 vs 4.4).
- Maine has covered the spread in 60% of its last five games.
- With the model and market numbers this close, there's no strong lean here beyond what the market already reflects.
Per-model breakdown (home win probability)
Logistic Regression81%
Random Forest84%
XGBoost86%
LightGBM87%
Extra Trees79%
Final: App State 55 - 3 Maine — moneyline pick was correct, spread pick covered the spread
Tennessee State @ No. 3 Georgia
Sat Sep 5, 3:00 PM EDT
Model: Georgia by +39.7 (92% win prob) · Market: Georgia by 46.5
Moneyline pick: Georgia (92% win prob) HIGH Confidence†
Spread pick: Tennessee State +46.5 (-110*) MEDIUM Confidence†
Tennessee State has only 8 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 6.8-point gap between the model's margin and the market's line.
Georgia is a heavy favorite (46.5 in the market), and the model agrees directionally at +39.7 -- a 6.8-point gap that makes Tennessee State the technical pick, at medium confidence (66% cover probability).
- Georgia rates as clearly the better team by every measure here: an 83% ELO‡ win probability, a 211-point ELO edge (1698 vs 1488), and a 13.0-point opponent-adjusted margin advantage.
- Tennessee State has covered the spread in 40% of its last five games, and has outscored opponents by an average of 13.8 points less than Georgia over the same four-game window.
- Tennessee State has only 8 tracked games in the database as a recent FBS entrant -- its recent-form numbers are a small sample, likely contributing to this gap rather than reflecting a real edge.
Per-model breakdown (home win probability)
Logistic Regression98%
Random Forest95%
XGBoost98%
LightGBM99%
Extra Trees94%
Final: Georgia 63 - 3 Tennessee State — moneyline pick was correct, spread pick did not cover the spread
Rhode Island @ Temple
Sat Sep 5, 2:00 PM EDT
Model: Temple by +27.6 (82% win prob) · Market: Temple by 14.5
Moneyline pick: Temple (82% win prob) HIGH Confidence†
Spread pick: Temple -14.5 (-118) HIGH Confidence†
Rhode Island has only 14 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 13.1-point gap between the model's margin and the market's line.
The market favors Temple by 14.5, and the model favors Temple even more, by 27.6 -- a 13.1-point gap in Temple's favor, making Temple the pick at high confidence (79% cover probability).
- Temple rates ahead once schedule strength is factored in, a 14.4-point opponent-adjusted margin edge (-3.5 vs -17.9), and has the better recent scoring margin (-20.8 vs Rhode Island's -27.2).
- Rhode Island has actually posted the better yards-per-play number recently (5.1 vs 4.5) and has covered the spread in 60% of its last five games.
- Rhode Island has only 14 tracked games in the database as a recent FBS entrant, a thin sample likely contributing to why the model sees a wider gap here than the market's own line does.
Per-model breakdown (home win probability)
Logistic Regression80%
Random Forest77%
XGBoost84%
LightGBM81%
Extra Trees74%
Final: Temple 38 - 14 Rhode Island — moneyline pick was correct, spread pick covered the spread
Youngstown State @ Kentucky
Sat Sep 5, 1:00 PM EDT
Model: Kentucky by +31.3 (90% win prob) · Market: Kentucky by 22.5
Moneyline pick: Kentucky (90% win prob) HIGH Confidence†
Spread pick: Kentucky -22.5 (-110) HIGH Confidence†
Youngstown State has only 13 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 8.8-point gap between the model's margin and the market's line.
Kentucky is favored by 22.5 in the market, and the model actually likes Kentucky more, by 31.3 -- an 8.8-point gap in Kentucky's favor, making Kentucky the spread pick at high confidence (71% cover probability).
- Kentucky has the edge in recent scoring margin (-1.5 vs -33.0, last four games) and carries a 44-point ELO‡ advantage over Youngstown State.
- Youngstown State has faced tougher recent competition than Kentucky (average opponent rating +7.6 vs +2.9) and has covered the spread in 60% of its last five games.
- Youngstown State has only 13 tracked games in the database as a recent FBS entrant, a thin sample behind its recent-form numbers.
Per-model breakdown (home win probability)
Logistic Regression89%
Random Forest93%
XGBoost97%
LightGBM96%
Extra Trees87%
Final: Kentucky 45 - 13 Youngstown State — moneyline pick was correct, spread pick covered the spread
Southeast Missouri State @ Iowa State
Sat Sep 5, 1:00 PM EDT
Model: Iowa State by +28.4 (90% win prob) · Market: Iowa State by 29.5
Moneyline pick: Iowa State (90% win prob) HIGH Confidence†
Spread pick: Southeast Missouri State +29.5 (-110*) LOW Confidence†
Southeast Missouri State has only 14 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 1.1-point gap between the model's margin and the market's line.
Iowa State is favored by 29.5 in the market, and the model's own margin lands almost exactly there at 28.4 -- essentially no edge, so this pick carries only a 53% cover probability at low confidence.
- Iowa State rates as clearly the stronger team by every measure: a 73% ELO‡ win probability, a 110-point ELO edge, and a 21.0-point opponent-adjusted margin advantage.
- Southeast Missouri State has covered the spread in 60% of its last five games, though it has only 14 tracked games in the database as a recent FBS entrant -- a thin sample behind that trend.
- With the model and market numbers this close together, there's genuinely no lean here -- this is a pass-quality spread, not a real edge in either direction.
Per-model breakdown (home win probability)
Logistic Regression93%
Random Forest93%
XGBoost95%
LightGBM90%
Extra Trees88%
Final: Iowa State 38 - 10 Southeast Missouri State — moneyline pick was correct, spread pick covered the spread
Duquesne @ Air Force
Sat Sep 5, 1:00 PM EDT
Model: Air Force by +24.9 (91% win prob) · Market: Air Force by 28.5
Moneyline pick: Air Force (91% win prob) HIGH Confidence†
Spread pick: Duquesne +28.5 (-110*) MEDIUM Confidence†
Duquesne has only 13 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 3.6-point gap between the model's margin and the market's line.
Air Force is favored by 28.5 in the market, but the model's margin has it by 24.9 -- a 3.6-point gap that makes Duquesne the technical pick, at medium confidence (59% cover probability).
- Air Force still rates as clearly the better team: a 70% ELO‡ win probability, a 23.0-point opponent-adjusted margin advantage, and a much higher third-down conversion rate (50% vs 21%).
- Duquesne has covered the spread in 0% of its last five games.
- Duquesne has only 13 tracked games in the database as a recent FBS entrant, a thin sample behind its numbers -- the gap here is modest, matching the medium confidence tier.
Per-model breakdown (home win probability)
Logistic Regression94%
Random Forest95%
XGBoost98%
LightGBM98%
Extra Trees94%
Final: Air Force 34 - 0 Duquesne — moneyline pick was correct, spread pick did not cover the spread
Kent State @ South Carolina
Sat Sep 5, 12:45 PM EDT
Model: South Carolina by +23.3 (87% win prob) · Market: South Carolina by 36.0
Moneyline pick: South Carolina (87% win prob) HIGH Confidence†
Spread pick: Kent State +36.0 (-110) HIGH Confidence†
South Carolina is favored by 36.0 in the market, but the model's margin has it by 23.3 -- a 12.7-point gap that makes Kent State the pick, at high confidence (79% cover probability).
- South Carolina still rates as clearly the better team: a 75% ELO‡ win probability, a 14.7-point opponent-adjusted margin edge (+5.5 vs -9.2), and tougher recent competition (+5.0 vs -8.4).
- Kent State has covered the spread in 60% of its last five games.
- South Carolina's recent scoring margin edge (3.2 vs -1.2, last four games) is real but modest next to a 36-point market spread.
Per-model breakdown (home win probability)
Logistic Regression87%
Random Forest87%
XGBoost92%
LightGBM89%
Extra Trees81%
Final: South Carolina 57 - 0 Kent State — moneyline pick was correct, spread pick did not cover the spread
Miami (OH) @ Pittsburgh
Sat Sep 5, 12:30 PM EDT
Model: Pittsburgh by +11.8 (77% win prob) · Market: Pittsburgh by 16.5
Moneyline pick: Pittsburgh (77% win prob) HIGH Confidence†
Spread pick: Miami (OH) +16.5 (-110) MEDIUM Confidence†
Pittsburgh is favored by 16.5 in the market, but the model's margin has it by 11.8 -- a 4.7-point gap that makes Miami (OH) the pick, at medium confidence (61% cover probability).
- Pittsburgh still rates as the better team: a 64% ELO‡ win probability, an 8.6-point opponent-adjusted margin advantage, and tougher recent competition (+9.6 vs -5.2).
- Miami (OH) has actually posted the better recent scoring margin (3.2 vs Pittsburgh's -11.2, last four games) and the edge in yards per play (5.0 vs 4.6).
- The market's 16.5-point spread looks wider than the underlying numbers fully support, which is what's pulling the pick toward Miami (OH).
Per-model breakdown (home win probability)
Logistic Regression71%
Random Forest71%
XGBoost79%
LightGBM78%
Extra Trees68%
Final: Pittsburgh 59 - 14 Miami (OH) — moneyline pick was correct, spread pick did not cover the spread
Ball State @ No. 1 Ohio State
Sat Sep 5, 12:30 PM EDT
Model: Ohio State by +33.0 (92% win prob) · Market: Ohio State by 50.8
Moneyline pick: Ohio State (92% win prob) HIGH Confidence†
Spread pick: Ball State +50.8 (-110) HIGH Confidence†
Ohio State is a massive favorite (50.8 in the market), and the model still likes Ohio State big at +33.0 -- a 17.8-point gap that makes Ball State the technical pick, at high confidence (87% cover probability).
- Ohio State still rates as clearly the better team: a 90% ELO‡ win probability, a 318-point ELO edge (1719 vs 1401), and a 28.8-point opponent-adjusted margin advantage.
- Ohio State has also faced much tougher recent competition than Ball State (+11.7 vs +0.7).
- The market's 50.8-point spread is wider than even the model's own large rating gap supports, which is the core driver of this pick.
Per-model breakdown (home win probability)
Logistic Regression97%
Random Forest97%
XGBoost98%
LightGBM98%
Extra Trees93%
Final: Ohio State 56 - 3 Ball State — moneyline pick was correct, spread pick did not cover the spread
East Carolina @ No. 13 Alabama
Sat Sep 5, 12:00 PM EDT
Model: Alabama by +9.0 (72% win prob) · Market: Alabama by 28.2
Moneyline pick: Alabama (72% win prob) MEDIUM Confidence†
Spread pick: East Carolina +28.2 (-110) HIGH Confidence†
Alabama is favored by 28.2 in the market, but the model's own margin has Alabama by just 9.0 -- a 19.3-point gap that makes East Carolina the pick, with an 89% cover probability (high confidence).
- East Carolina has outscored opponents by 3.8 points per game on average over its last four, versus Alabama's -9.8 -- and out-gained them on a yards-per-play basis too, 5.7 to 4.1.
- Alabama still rates as the better team on paper: pre-game ELO‡ gives it a 66% win probability, and it holds an 8.6-point opponent-adjusted margin edge (+12.3 vs +3.7).
- Alabama has also faced tougher recent competition, an average opponent rating of +12.9 versus East Carolina's +1.8.
- The market's 28.2-point spread looks out of proportion to how close the underlying numbers actually are, which is what's driving the pick toward East Carolina.
Per-model breakdown (home win probability)
Logistic Regression75%
Random Forest63%
XGBoost69%
LightGBM68%
Extra Trees62%
Final: Alabama 48 - 10 East Carolina — moneyline pick was correct, spread pick did not cover the spread
Oregon State @ No. 23 Houston
Sat Sep 5, 12:00 PM EDT
Model: Houston by +13.2 (86% win prob) · Market: Houston by 21.0
Moneyline pick: Houston (86% win prob) HIGH Confidence†
Spread pick: Oregon State +21.0 (-115) HIGH Confidence†
Houston is favored by 21.0 in the market, but the model's margin has it by 13.2 -- a 7.8-point gap that makes Oregon State the pick, at high confidence (69% cover probability).
- Houston still rates as the better team: a 76% ELO‡ win probability, a 136-point ELO edge (1554 vs 1418), and tougher recent competition (+1.3 vs -3.6).
- Oregon State has covered the spread in 40% of its last five games.
- Houston's recent scoring margin edge (2.5 vs -10.5, last four games) is real but doesn't fully support a 21-point market spread.
Per-model breakdown (home win probability)
Logistic Regression85%
Random Forest86%
XGBoost85%
LightGBM83%
Extra Trees85%
Final: Houston 33 - 20 Oregon State — moneyline pick was correct, spread pick covered the spread
Coastal Carolina @ West Virginia
Sat Sep 5, 12:00 PM EDT
Model: West Virginia by +21.4 (87% win prob) · Market: West Virginia by 21.0
Moneyline pick: West Virginia (87% win prob) HIGH Confidence†
Spread pick: West Virginia -21.0 (-110) LOW Confidence†
West Virginia is favored by 21.0 in the market, and the model's margin lands almost exactly there at 21.4 -- essentially no edge, so this pick carries only a 51% cover probability at low confidence.
- West Virginia has the better recent scoring margin (-8.5 vs -26.8, last four games) and the edge in yards per play (4.8 vs 4.4).
- West Virginia has also faced tougher recent competition than Coastal Carolina (+6.5 vs +0.6), and pre-game ELO‡ gives it a 61% win probability.
- Coastal Carolina has covered the spread in 40% of its last five games -- with the model and market numbers this close, there's no real lean here.
Per-model breakdown (home win probability)
Logistic Regression83%
Random Forest86%
XGBoost89%
LightGBM91%
Extra Trees84%
Final: West Virginia 31 - 24 Coastal Carolina — moneyline pick was correct, spread pick did not cover the spread
New Hampshire @ Syracuse
Sat Sep 5, 12:00 PM EDT
Model: Syracuse by +14.4 (72% win prob) · Market: Syracuse by 33.5
Moneyline pick: Syracuse (72% win prob) MEDIUM Confidence†
Spread pick: New Hampshire +33.5 (-110*) HIGH Confidence†
New Hampshire has only 14 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 19.1-point gap between the model's margin and the market's line.
Syracuse is favored by 33.5 in the market, but the model's margin has it by just 14.4 -- a 19.1-point gap that makes New Hampshire the pick, at high confidence (88% cover probability).
- Pre-game ELO‡ sees this as close to even, giving Syracuse only a 56% win probability, and New Hampshire actually carries a 25-point ELO edge (1442 vs 1467).
- New Hampshire has the better recent scoring margin too, -24.8 versus Syracuse's -32.5 over the last four games.
- New Hampshire has only 14 tracked games in the database as a recent FBS entrant -- a thin sample behind its recent-form numbers, likely contributing to the size of this gap alongside the market's own wide 33.5-point line.
Per-model breakdown (home win probability)
Logistic Regression71%
Random Forest66%
XGBoost65%
LightGBM70%
Extra Trees67%
Final: Syracuse 66 - 3 New Hampshire — moneyline pick was correct, spread pick did not cover the spread
North Texas @ No. 6 Indiana
Sat Sep 5, 12:00 PM EDT
Model: Indiana by +17.7 (85% win prob) · Market: Indiana by 40.8
Moneyline pick: Indiana (85% win prob) HIGH Confidence†
Spread pick: North Texas +40.8 (-110) HIGH Confidence†
Indiana is favored by a massive 40.8 in the market, but the model's margin has it by just 17.7 -- a 23.1-point gap that makes North Texas the pick, at high confidence (93% cover probability, the largest edge in this week's slate).
- Indiana still rates as clearly the better team: a 78% ELO‡ win probability, a 13.4-point opponent-adjusted margin advantage, and much tougher recent competition (+15.2 vs -1.6).
- North Texas has actually posted the better yards-per-play number recently, 8.4 versus Indiana's 5.6 over the last four games.
- The market's 40.8-point spread is far wider than any of the model's underlying ratings support, which is the core driver of this large edge.
Per-model breakdown (home win probability)
Logistic Regression87%
Random Forest81%
XGBoost91%
LightGBM82%
Extra Trees75%
Final: Indiana 52 - 16 North Texas — moneyline pick was correct, spread pick covered the spread
Ohio @ Nebraska
Sat Sep 5, 12:00 PM EDT
Model: Nebraska by +9.4 (66% win prob) · Market: Nebraska by 23.5
Moneyline pick: Nebraska (66% win prob) MEDIUM Confidence†
Spread pick: Ohio +23.5 (-106) HIGH Confidence†
Nebraska is favored by 23.5 in the market, but the model's margin has it by just 9.4 -- a 14.1-point gap that makes Ohio the pick, at high confidence (81% cover probability).
- Ohio actually carries a 72-point ELO‡ edge over Nebraska (1521 vs 1593) and has the better recent scoring margin (9.0 vs Nebraska's -16.5, last four games).
- Pre-game ELO sees this as close to even, giving Nebraska only a 49% win probability -- not the kind of gap that supports a 23.5-point market spread.
- Nebraska has faced tougher recent competition than Ohio (+9.3 vs -8.2), which is the main thing keeping it favored at all.
Per-model breakdown (home win probability)
Logistic Regression58%
Random Forest61%
XGBoost75%
LightGBM71%
Extra Trees52%
Final: Nebraska 49 - 21 Ohio — moneyline pick was correct, spread pick did not cover the spread
Lafayette @ UConn
Sat Sep 5, 12:00 PM EDT
Model: UConn by +30.6 (91% win prob) · Market: UConn by 19.5
Moneyline pick: UConn (91% win prob) HIGH Confidence†
Spread pick: UConn -19.5 (-114) HIGH Confidence†
Lafayette has only 8 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 11.1-point gap between the model's margin and the market's line.
The market favors UConn by 19.5, but the model favors UConn even more, by 30.6 -- an 11.1-point gap in UConn's favor, making UConn the pick at high confidence (76% cover probability).
- UConn has the edge in recent scoring margin (-2.2 vs -25.8, last four games) and in yards per play (6.6 vs 3.6).
- UConn also rates ahead once schedule strength is factored in, an 18.2-point opponent-adjusted margin edge, and carries a 116-point ELO‡ advantage.
- Lafayette has only 8 tracked games in the database as a recent FBS entrant, a thin sample behind its numbers -- consistent with the model seeing more separation here than the market's own line does.
Per-model breakdown (home win probability)
Logistic Regression92%
Random Forest96%
XGBoost96%
LightGBM93%
Extra Trees94%
Final: UConn 56 - 7 Lafayette — moneyline pick was correct, spread pick covered the spread
Bryant @ Army
Sat Sep 5, 12:00 PM EDT
Model: Army by +23.5 (88% win prob) · Market: Army by 36.5
Moneyline pick: Army (88% win prob) HIGH Confidence†
Spread pick: Bryant +36.5 (-110*) HIGH Confidence†
Bryant has only 5 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 13.0-point gap between the model's margin and the market's line.
Army is favored by 36.5 in the market, but the model's margin has it by 23.5 -- a 13.0-point gap that makes Bryant the pick, at high confidence (79% cover probability).
- Army still rates as clearly the better team: a 72% ELO‡ win probability, a 16.4-point opponent-adjusted margin advantage, and tougher recent competition (-1.4 vs -11.4).
- Bryant has covered the spread in 60% of its last five games and has only 5 tracked games in the database as a recent FBS entrant -- a thin sample likely contributing to this gap.
- The market's 36.5-point spread is wider than Army's own underlying ratings support, which is the core driver of this pick.
Per-model breakdown (home win probability)
Logistic Regression90%
Random Forest86%
XGBoost91%
LightGBM92%
Extra Trees81%
Final: Army 59 - 3 Bryant — moneyline pick was correct, spread pick did not cover the spread
Tarleton State @ Bowling Green
Sat Sep 5, 12:00 PM EDT
Model: Bowling Green by +16.8 (80% win prob) · Market: Bowling Green by 2.5
Moneyline pick: Bowling Green (80% win prob) HIGH Confidence†
Spread pick: Bowling Green -2.5 (-112) HIGH Confidence†
Tarleton State has only 4 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 14.3-point gap between the model's margin and the market's line.
The market favors Bowling Green by just 2.5, but the model likes Bowling Green far more, by 16.8 -- a 14.3-point gap that makes Bowling Green the pick, at high confidence (82% cover probability).
- Bowling Green has a large recent scoring margin edge (-0.8 vs Tarleton State's -29.8, last four games).
- Tarleton State actually rates ahead once schedule strength is factored in (-7.8 vs +1.0) and carries a 26-point ELO‡ edge, despite that scoring-margin gap.
- Tarleton State has only 4 tracked games in the database as a recent FBS entrant, an extremely thin sample -- likely why the market's own line is so much tighter than the model's read here.
Per-model breakdown (home win probability)
Logistic Regression79%
Random Forest71%
XGBoost80%
LightGBM81%
Extra Trees73%
Final: Bowling Green 13 - 20 Tarleton State — moneyline pick was incorrect, spread pick did not cover the spread
Liberty @ James Madison
Sat Sep 5, 12:00 PM EDT
Model: James Madison by +16.8 (87% win prob) · Market: James Madison by 6.0
Moneyline pick: James Madison (87% win prob) HIGH Confidence†
Spread pick: James Madison -6.0 (-110) HIGH Confidence†
The market favors James Madison by 6.0, but the model likes James Madison considerably more, by 16.8 -- a 10.8-point gap in James Madison's favor, making James Madison the pick at high confidence (75% cover probability).
- James Madison has the edge in recent scoring margin (13.2 vs Liberty's -5.8, last four games) and carries a 171-point ELO‡ advantage (1666 vs 1495).
- James Madison also rates ahead once schedule strength is factored in, an 11.9-point opponent-adjusted margin edge.
- Liberty has covered the spread in just 20% of its last five games.
Per-model breakdown (home win probability)
Logistic Regression86%
Random Forest88%
XGBoost86%
LightGBM85%
Extra Trees86%
Final: James Madison 20 - 13 Liberty — moneyline pick was correct, spread pick covered the spread
No. 7 Miami @ Stanford
Fri Sep 4, 9:00 PM EDT
Model: Stanford by -7.6 (33% win prob) · Market: Miami by 24.2
Moneyline pick: Miami (67% win prob) MEDIUM Confidence†
Spread pick: Stanford +24.2 (-114) HIGH Confidence†
Miami is favored by 24.2 in the market, but the model actually favors Stanford by 7.6 -- a 16.7-point swing that makes Stanford the pick, at high confidence (85% cover probability).
- Miami carries a large 216-point ELO‡ advantage (1478 vs 1694) and has covered the spread in 100% of its last five games, both real factors working against this pick.
- Miami has faced tougher recent competition than Stanford (average opponent rating +15.5 vs +2.6), and has the better recent scoring margin (3.8 vs -0.8, last four games).
- Despite Miami's edge on the surface, the model's own rating gap between these teams isn't large enough to support a 24.2-point market spread, which is what's driving this pick toward Stanford.
Per-model breakdown (home win probability)
Logistic Regression37%
Random Forest37%
XGBoost33%
LightGBM31%
Extra Trees36%
Final: Stanford 6 - 45 Miami — moneyline pick was correct, spread pick did not cover the spread
Fresno State @ No. 14 USC
Fri Sep 4, 9:00 PM EDT
Model: USC by +8.0 (81% win prob) · Market: USC by 22.0
Moneyline pick: USC (81% win prob) HIGH Confidence†
Spread pick: Fresno State +22.0 (-110) HIGH Confidence†
USC is favored by 22.0 in the market, but the model's margin has it by 8.0 -- a 14.0-point gap that makes Fresno State the pick, at high confidence (81% cover probability).
- USC still rates as clearly the better team: a 67% ELO‡ win probability, a 17.2-point opponent-adjusted margin advantage, and a much higher third-down conversion rate recently (55% vs 29%).
- Fresno State has actually posted the better recent scoring margin, 13.0 versus USC's 4.2 over the last four games.
- Over their last four meetings, USC has outscored Fresno State by an average of 25.0 points -- a real trend, though not one that closes this gap against a 22-point market spread.
Per-model breakdown (home win probability)
Logistic Regression84%
Random Forest74%
XGBoost78%
LightGBM80%
Extra Trees70%
Final: USC 39 - 0 Fresno State — moneyline pick was correct, spread pick did not cover the spread
UTEP @ No. 10 Oklahoma
Fri Sep 4, 8:00 PM EDT
Model: Oklahoma by +22.9 (88% win prob) · Market: Oklahoma by 41.0
Moneyline pick: Oklahoma (88% win prob) HIGH Confidence†
Spread pick: UTEP +41.0 (-110) HIGH Confidence†
Oklahoma is favored by 41.0 in the market, but the model's margin has Oklahoma by 22.9 -- an 18.1-point gap that makes UTEP the pick, at high confidence (87% cover probability).
- Oklahoma still projects as clearly the stronger team: an 83% ELO‡ win probability, a 213-point ELO edge (1613 vs 1400), and a 21.0-point opponent-adjusted margin advantage (+11.5 vs -9.6), against much tougher recent competition (+10.0 vs -13.9).
- Over their last three meetings, Oklahoma has outscored UTEP by an average of 32.7 points -- a real head-to-head trend, though not one that closes an 18-point gap against this specific market number.
- The market's 41-point spread is wider than any of the model's underlying ratings support, which is the core driver of this pick.
Per-model breakdown (home win probability)
Logistic Regression91%
Random Forest87%
XGBoost90%
LightGBM94%
Extra Trees85%
Final: Oklahoma 51 - 0 UTEP — moneyline pick was correct, spread pick did not cover the spread
Long Island University @ Kansas
Fri Sep 4, 8:00 PM EDT
Model: Kansas by +36.2 (90% win prob) · Market: Kansas by 41.5
Moneyline pick: Kansas (90% win prob) HIGH Confidence†
Spread pick: Long Island University +41.5 (-110*) MEDIUM Confidence†
Long Island University has only 10 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 5.3-point gap between the model's margin and the market's line.
Kansas is favored by 41.5 in the market, but the model's margin has it by 36.2 -- a 5.3-point gap that makes Long Island University the technical pick, at medium confidence (63% cover probability).
- Kansas still rates as clearly the better team: a 67% ELO‡ win probability, a 54-point ELO edge (1516 vs 1462), and a 17.4-point opponent-adjusted margin advantage.
- Long Island University has covered the spread in 80% of its last five games, though it has only 10 tracked games in the database as a recent FBS entrant -- a thin sample behind that trend.
- Kansas has the better recent scoring margin too, -5.2 versus Long Island University's -29.5 over the last four games, despite both being negative.
Per-model breakdown (home win probability)
Logistic Regression95%
Random Forest94%
XGBoost97%
LightGBM95%
Extra Trees88%
Final: Kansas 51 - 6 Long Island University — moneyline pick was correct, spread pick did not cover the spread
Toledo @ Michigan State
Fri Sep 4, 8:00 PM EDT
Model: Michigan State by -1.5 (41% win prob) · Market: Michigan State by 10.0
Moneyline pick: Toledo (59% win prob) LOW Confidence†
Spread pick: Toledo +10.0 (-110) HIGH Confidence†
The market favors Michigan State by 10.0, but the model actually favors Toledo, +1.5 -- an 11.5-point swing that makes Toledo the pick, at high confidence (76% cover probability).
- Toledo has the better recent scoring margin (15.8 vs Michigan State's -3.5, last four games), and the two teams grade out nearly even in opponent-adjusted margin (+2.9 vs +3.7).
- Toledo has covered the spread in 80% of its last five games.
- Pre-game ELO‡ actually gives Michigan State only a 41% win probability -- Toledo is favored by the model's own numbers here despite being the market underdog.
Per-model breakdown (home win probability)
Logistic Regression43%
Random Forest40%
XGBoost43%
LightGBM43%
Extra Trees40%
Final: Michigan State 30 - 20 Toledo — moneyline pick was incorrect, spread pick push the spread
Indiana State @ Purdue
Fri Sep 4, 7:00 PM EDT
Model: Purdue by +23.1 (87% win prob) · Market: Purdue by 35.5
Moneyline pick: Purdue (87% win prob) HIGH Confidence†
Spread pick: Indiana State +35.5 (-110*) HIGH Confidence†
Indiana State has only 18 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 12.4-point gap between the model's margin and the market's line.
Purdue is favored by 35.5 in the market, but the model's margin has it by 23.1 -- a 12.4-point gap that makes Indiana State the pick, at high confidence (78% cover probability).
- Purdue still rates ahead once schedule strength is factored in, a 17.1-point opponent-adjusted margin edge (+0.8 vs -16.3), and has outscored opponents by more recently (-29.5 vs Indiana State's -48.5).
- Over their last four meetings, Purdue has outscored Indiana State by an average of 33.8 points -- a lopsided head-to-head history.
- Indiana State has covered the spread in 0% of its last five games and has only 18 tracked games in the database as a recent FBS entrant -- a thin sample likely contributing to this gap alongside the market's own wide line.
Per-model breakdown (home win probability)
Logistic Regression87%
Random Forest87%
XGBoost93%
LightGBM88%
Extra Trees83%
Final: Purdue 44 - 19 Indiana State — moneyline pick was correct, spread pick covered the spread
North Carolina A&T @ Georgia State
Fri Sep 4, 7:00 PM EDT
Model: Georgia State by +26.6 (87% win prob) · Market: Georgia State by 28.5
Moneyline pick: Georgia State (87% win prob) HIGH Confidence†
Spread pick: North Carolina A&T +28.5 (-110*) LOW Confidence†
North Carolina A&T has only 11 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 1.9-point gap between the model's margin and the market's line.
Georgia State is favored by 28.5 in the market, and the model's margin lands close behind at 26.6 -- a 1.9-point gap that makes North Carolina A&T the technical pick, though this is a low-confidence, essentially coin-flip edge.
- Georgia State rates ahead once schedule strength is factored in, an 18.7-point opponent-adjusted margin edge (-11.1 vs -29.8), despite pre-game ELO‡ seeing this as a dead-even 50% win probability.
- North Carolina A&T actually carries a 63-point ELO edge (1386 vs 1448) and has only 11 tracked games in the database as a recent FBS entrant, a thin sample behind its numbers.
- With the model and market numbers this close, there's no strong lean here in either direction.
Per-model breakdown (home win probability)
Logistic Regression83%
Random Forest85%
XGBoost94%
LightGBM94%
Extra Trees80%
Final: Georgia State 59 - 10 North Carolina A&T — moneyline pick was correct, spread pick did not cover the spread
San José State @ Eastern Michigan
Fri Sep 4, 6:30 PM EDT
Model: Eastern Michigan by +24.3 (89% win prob) · Market: Eastern Michigan by 3.0
Moneyline pick: Eastern Michigan (89% win prob) HIGH Confidence†
Spread pick: Eastern Michigan -3.0 (-110) HIGH Confidence†
The market favors Eastern Michigan by just 3.0, but the model likes Eastern Michigan far more, by 24.3 -- a 21.3-point gap that makes Eastern Michigan the pick, at high confidence (91% cover probability, one of the largest edges this week).
- Eastern Michigan has the edge in recent scoring margin (5.5 vs -27.5, last four games) and a 27.0-point opponent-adjusted margin advantage (+11.0 vs -16.0).
- San José State has covered the spread in just 20% of its last five games.
- San José State has faced tougher recent competition than Eastern Michigan (+2.8 vs -7.7), which is part of why the market's own line is so much tighter than the model's read on this game.
Per-model breakdown (home win probability)
Logistic Regression84%
Random Forest93%
XGBoost92%
LightGBM93%
Extra Trees89%
Final: Eastern Michigan 21 - 27 San José State — moneyline pick was incorrect, spread pick did not cover the spread
Idaho @ No. 21 Utah
Thu Sep 3, 9:00 PM EDT
Model: Utah by +30.4 (91% win prob) · Market: Utah by 36.5
Moneyline pick: Utah (91% win prob) HIGH Confidence†
Spread pick: Idaho +36.5 (-108) MEDIUM Confidence†
Utah is favored by 36.5 in the market, but the model's margin has it by 30.4 -- a 6.1-point gap that makes Idaho the pick, at medium confidence (65% cover probability).
- Utah still projects as clearly the better team: an 81% ELO‡ win probability, a 185-point ELO edge (1673 vs 1488), and a 17.3-point opponent-adjusted margin advantage.
- Idaho has covered the spread in 100% of its last five games, the strongest recent ATS trend in this matchup.
- Utah's recent scoring margin edge (15.8 vs -3.0, last four games) is real, but the gap between the model and market numbers is what's actually driving this pick.
Per-model breakdown (home win probability)
Logistic Regression95%
Random Forest92%
XGBoost96%
LightGBM96%
Extra Trees92%
Final: Utah 66 - 14 Idaho — moneyline pick was correct, spread pick did not cover the spread
UAB @ Illinois
Thu Sep 3, 9:00 PM EDT
Model: Illinois by +18.9 (86% win prob) · Market: Illinois by 27.5
Moneyline pick: Illinois (86% win prob) HIGH Confidence†
Spread pick: UAB +27.5 (-110) HIGH Confidence†
Illinois is favored by 27.5 in the market, but the model's margin has it by 18.9 -- an 8.6-point gap that makes UAB the pick, at high confidence (71% cover probability).
- Illinois still rates as clearly the better team: a 77% ELO‡ win probability, a 15.7-point opponent-adjusted margin advantage, and a 141-point ELO edge (1571 vs 1430).
- UAB has covered the spread in just 20% of its last five games.
- Illinois's recent scoring margin edge (2.5 vs -14.8, last four games) is real but doesn't fully support a 27.5-point market spread.
Per-model breakdown (home win probability)
Logistic Regression85%
Random Forest87%
XGBoost86%
LightGBM86%
Extra Trees83%
Final: Illinois 42 - 23 UAB — moneyline pick was correct, spread pick covered the spread
Arkansas-Pine Bluff @ No. 25 Missouri
Thu Sep 3, 8:00 PM EDT
Model: Missouri by +47.9 (92% win prob) · Market: Missouri by 55.5
Moneyline pick: Missouri (92% win prob) HIGH Confidence†
Spread pick: Arkansas-Pine Bluff +55.5 (-110*) HIGH Confidence†
Arkansas-Pine Bluff has only 11 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 7.6-point gap between the model's margin and the market's line.
Missouri is a heavy favorite (55.5 in the market), and the model agrees at +47.9 -- a 7.6-point gap that makes Arkansas-Pine Bluff the technical pick, at high confidence (68% cover probability).
- Missouri rates as clearly the better team across the board: a 77% ELO‡ win probability, a 149-point ELO edge (1588 vs 1439), a 21.5-point opponent-adjusted margin advantage, and tougher recent competition (+11.6 vs +2.5).
- Arkansas-Pine Bluff has only 11 tracked games in the database as a recent FBS entrant -- its recent-form numbers are a small, likely unrepresentative sample, which is probably contributing to this gap more than any real value.
- Missouri has outscored opponents by 4.8 points per game on average recently, against Arkansas-Pine Bluff's -55.2.
Per-model breakdown (home win probability)
Logistic Regression99%
Random Forest98%
XGBoost99%
LightGBM99%
Extra Trees97%
Final: Missouri 54 - 14 Arkansas-Pine Bluff — moneyline pick was correct, spread pick covered the spread
Colorado @ Georgia Tech
Thu Sep 3, 8:00 PM EDT
Model: Georgia Tech by +16.0 (85% win prob) · Market: Georgia Tech by 6.5
Moneyline pick: Georgia Tech (85% win prob) HIGH Confidence†
Spread pick: Georgia Tech -6.5 (-110) HIGH Confidence†
Georgia Tech is favored by 6.5 in the market, and the model likes Georgia Tech considerably more, by 16.0 -- a 9.5-point gap in Georgia Tech's favor, making Georgia Tech the pick at high confidence (72% cover probability).
- Georgia Tech has the edge in recent scoring margin (-5.8 vs -19.2, last four games) and in yards per play (6.0 vs 4.4).
- Georgia Tech also rates ahead once schedule strength is factored in, a 5.3-point opponent-adjusted margin edge (+4.7 vs -0.5), and pre-game ELO‡ gives it a 73% win probability.
- Colorado has covered the spread in just 20% of its last five games.
Per-model breakdown (home win probability)
Logistic Regression86%
Random Forest85%
XGBoost83%
LightGBM81%
Extra Trees84%
Final: Georgia Tech 13 - 14 Colorado — moneyline pick was incorrect, spread pick did not cover the spread
Eastern Illinois @ Minnesota
Thu Sep 3, 8:00 PM EDT
Model: Minnesota by +26.7 (91% win prob) · Market: Minnesota by 42.5
Moneyline pick: Minnesota (91% win prob) HIGH Confidence†
Spread pick: Eastern Illinois +42.5 (-110*) HIGH Confidence†
Eastern Illinois has only 17 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 15.8-point gap between the model's margin and the market's line.
Minnesota is favored by 42.5 in the market, but the model's margin has it by 26.7 -- a 15.8-point gap that makes Eastern Illinois the pick, at high confidence (84% cover probability).
- Minnesota still rates as clearly the better team: a 72% ELO‡ win probability, an 18.0-point opponent-adjusted margin advantage, and the edge in turnover margin (1.2 vs -1.2).
- Eastern Illinois has covered the spread in 40% of its last five games and has only 17 tracked games in the database as a recent FBS entrant -- a thin sample behind its numbers, likely contributing to this gap.
- The market's 42.5-point spread is wider than the model's underlying ratings support, which is the core driver of this pick.
Per-model breakdown (home win probability)
Logistic Regression93%
Random Forest96%
XGBoost97%
LightGBM95%
Extra Trees93%
Final: Minnesota 59 - 7 Eastern Illinois — moneyline pick was correct, spread pick did not cover the spread
Bethune-Cookman @ UCF
Thu Sep 3, 7:00 PM EDT
Model: UCF by +37.5 (91% win prob) · Market: UCF by 42.5
Moneyline pick: UCF (91% win prob) HIGH Confidence†
Spread pick: Bethune-Cookman +42.5 (-110*) MEDIUM Confidence†
UCF is favored by 42.5 in the market, but the model's margin has it by 37.5 -- a 5.0-point gap that makes Bethune-Cookman the technical pick, at medium confidence (62% cover probability).
- UCF still rates as the better team once schedule strength is accounted for, a 14.9-point opponent-adjusted margin edge (+0.7 vs -14.2), and pre-game ELO‡ gives it a 69% win probability.
- Over their last two meetings, UCF has outscored Bethune-Cookman by an average of 41.5 points -- a lopsided head-to-head history in the favorite's direction.
- Bethune-Cookman has covered the spread in 80% of its last five games, though the small gap here means this isn't a strong lean either way.
Per-model breakdown (home win probability)
Logistic Regression96%
Random Forest94%
XGBoost97%
LightGBM96%
Extra Trees91%
Final: UCF 73 - 6 Bethune-Cookman — moneyline pick was correct, spread pick did not cover the spread
Akron @ Wake Forest
Thu Sep 3, 7:00 PM EDT
Model: Wake Forest by +22.6 (87% win prob) · Market: Wake Forest by 27.5
Moneyline pick: Wake Forest (87% win prob) HIGH Confidence†
Spread pick: Akron +27.5 (-110) MEDIUM Confidence†
Wake Forest is favored by 27.5 in the market, but the model's margin has it by 22.6 -- a 4.9-point gap that makes Akron the pick, at medium confidence (62% cover probability).
- Wake Forest still rates as the better team: a 72% ELO‡ win probability, a 12.8-point opponent-adjusted margin advantage, and tougher recent competition (+1.5 vs -10.6).
- Akron has covered the spread in 60% of its last five games.
- Wake Forest's recent scoring margin edge is real but close (12.8 vs 9.5, last four games), not the kind of gap that fully supports a 27.5-point market spread.
Per-model breakdown (home win probability)
Logistic Regression86%
Random Forest86%
XGBoost90%
LightGBM89%
Extra Trees81%
Final: Wake Forest 38 - 16 Akron — moneyline pick was correct, spread pick covered the spread
Merrimack @ Delaware
Thu Sep 3, 7:00 PM EDT
Model: Delaware by +13.2 (84% win prob) · Market: Delaware by 31.5
Moneyline pick: Delaware (84% win prob) HIGH Confidence†
Spread pick: Merrimack +31.5 (-110*) HIGH Confidence†
Merrimack has only 5 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 18.3-point gap between the model's margin and the market's line.
Delaware is favored by 31.5 in the market, but the model's margin has it by 13.2 -- an 18.3-point gap that makes Merrimack the pick, at high confidence (87% cover probability).
- Delaware still rates ahead once schedule strength is factored in, though only by 2.9 points (-7.1 vs -10.0), a much thinner edge than the market's spread implies.
- Merrimack has covered the spread in 80% of its last five games and has actually faced tougher recent competition than Delaware (-3.8 vs -6.8).
- Merrimack has only 5 tracked games in the database as a recent FBS entrant, a thin sample likely contributing to the size of this gap alongside the market's own wide line.
Per-model breakdown (home win probability)
Logistic Regression85%
Random Forest81%
XGBoost77%
LightGBM75%
Extra Trees86%
Final: Delaware 42 - 7 Merrimack — moneyline pick was correct, spread pick did not cover the spread
West Georgia @ Kennesaw State
Thu Sep 3, 7:00 PM EDT
Model: Kennesaw State by +21.9 (82% win prob) · Market: Kennesaw State by 22.5
Moneyline pick: Kennesaw State (82% win prob) HIGH Confidence†
Spread pick: West Georgia +22.5 (-110*) LOW Confidence†
West Georgia has only 0 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 0.6-point gap between the model's margin and the market's line.
Kennesaw State is favored by 22.5 in the market, and the model's margin lands almost exactly there at 21.9 -- essentially no edge, so this pick carries only a 51% cover probability at low confidence.
- West Georgia actually rates slightly ahead once schedule strength is factored in (-1.8 vs 0.0), despite Kennesaw State carrying a small 9-point ELO‡ edge (1509 vs 1500).
- West Georgia has only 0 tracked games in the database as a recent FBS entrant -- essentially no usable recent-form history behind either side of this line.
- Both teams enter on a full 7 days of rest -- with the model and market numbers this close, there's no real lean here.
Per-model breakdown (home win probability)
Logistic Regression70%
Random Forest82%
XGBoost84%
LightGBM86%
Extra Trees76%
Final: Kennesaw State 47 - 0 West Georgia — moneyline pick was correct, spread pick did not cover the spread
UAlbany @ Buffalo
Thu Sep 3, 7:00 PM EDT
Model: Buffalo by +13.8 (82% win prob) · Market: Buffalo by 17.5
Moneyline pick: Buffalo (82% win prob) HIGH Confidence†
Spread pick: UAlbany +17.5 (-110*) MEDIUM Confidence†
UAlbany has only 11 tracked games in our database (a recent FBS entrant, or a rarely-tracked crossover opponent -- our data only covers games involving an FBS team) -- its recent-form stats are a small, possibly unrepresentative sample, likely contributing to the 3.7-point gap between the model's margin and the market's line.
Buffalo is favored by 17.5 in the market, but the model's margin has it by 13.8 -- a 3.7-point gap that makes UAlbany the pick, at medium confidence (59% cover probability).
- UAlbany actually rates ahead once schedule strength is factored in, a 6.2-point opponent-adjusted margin edge (-7.5 vs -1.3), and has faced tougher recent competition than Buffalo (-0.7 vs -4.4).
- UAlbany has covered the spread in 60% of its last five games.
- Buffalo still has the better recent scoring margin (-4.0 vs -19.2, last four games), which is the main thing keeping it favored at all.
Per-model breakdown (home win probability)
Logistic Regression82%
Random Forest79%
XGBoost77%
LightGBM74%
Extra Trees79%
Final: Buffalo 21 - 17 UAlbany — moneyline pick was correct, spread pick covered the spread
Massachusetts @ Rutgers
Thu Sep 3, 6:00 PM EDT
Model: Rutgers by +31.5 (91% win prob) · Market: Rutgers by 28.8
Moneyline pick: Rutgers (91% win prob) HIGH Confidence†
Spread pick: Rutgers -28.8 (-112) LOW Confidence†
Rutgers is favored by 28.8 in the market, and the model likes Rutgers slightly more, by 31.5 -- a 2.8-point gap in Rutgers's favor, making Rutgers the pick at medium confidence (57% cover probability).
- Rutgers rates as clearly the better team once schedule strength is factored in, a 20.9-point opponent-adjusted margin edge (+3.0 vs -17.9), and pre-game ELO‡ gives it an 83% win probability.
- Rutgers has also faced much tougher recent competition than Massachusetts (+9.6 vs -6.9) and carries a 212-point ELO advantage.
- Rutgers has the better recent scoring margin too (-11.0 vs -33.8, last four games), though this edge is close enough to the market's own number to stay at medium confidence.
Per-model breakdown (home win probability)
Logistic Regression95%
Random Forest97%
XGBoost97%
LightGBM95%
Extra Trees93%
Final: Rutgers 21 - 37 Massachusetts — moneyline pick was incorrect, spread pick did not cover the spread