Skip to main content

Public validation record

2026 Football Model Performance

These results use official OSPR projections locked before kickoff and evaluated against final scores. Mirrored schedule records and live forecasts are excluded.

Evaluated through Oct 15, 2026

Accuracy calculated Oct 16, 2026 8:00 AM EDT

Reset

Performance summary

2,842 games evaluated with ospr-2026.1

Games evaluated

2,842

Winner accuracy

68.5%

1,947 correct

Average margin error

22.5

Lower is better

Average score error

13.3

Per-team points

Brier score

0.203

Probability accuracy

Close projection rate

14.1%

Correct and within 7

Regular season vs. playoffs

A direct comparison using the selected season, week, team, and OSPR model version.

Regular season

Games
2,842
Winner accuracy
68.5%
Margin error
22.5

Playoffs

Games
0
Winner accuracy
—
Margin error
—

Performance by week

Results for the selected model version and season phase.

Week Games Winner accuracy Margin error
Week 1 355 58.3% 22.0
Week 2 351 61.0% 23.3
Week 3 358 62.3% 21.9
Week 4 354 65.3% 23.5
Week 5 356 74.4% 22.6
Week 6 357 80.4% 22.7
Week 7 358 75.4% 22.0
Week 8 348 71.3% 21.8
Week 9 4 25.0% 26.2

Probability calibration

Observed win rate should approach the average predicted probability as sample sizes grow.

Probability Games Predicted Observed Gap Brier
0–9% 13 8.5% 7.7% -0.8 0.069
10–19% 178 15.9% 9.6% -6.4 0.091
20–29% 302 25.5% 23.2% -2.3 0.178
30–39% 445 35.1% 35.1% 0.0 0.227
40–49% 562 45.1% 43.6% -1.5 0.245
50–59% 478 54.9% 57.7% +2.9 0.247
60–69% 404 64.8% 71.0% +6.2 0.205
70–79% 325 74.8% 79.7% +4.9 0.162
80–89% 128 84.0% 86.7% +2.7 0.113
90–99% 7 91.9% 85.7% -6.1 0.127

Performance by prediction confidence

Newer snapshots use historical calibration and comparable sample size. Legacy snapshots fall back to rating-evidence confidence.

Confidence Games Winner accuracy Average margin error
High 2,368 69.6% 22.4 points
Medium 31 80.6% 23.7 points
Low 443 61.9% 22.6 points

Performance by risk indicator

A game may appear in more than one group because indicators describe different kinds of uncertainty.

Indicator Games Winner accuracy Average margin error
Limited Data 1,515 62.4% 22.9 points
Close Game 2,021 62.8% 22.3 points
Upset Possible 1,886 61.9% 22.1 points
Blowout Range 1 100.0% 23.8 points

Largest misses

The largest margin errors in this selection. A miss is a model result to learn from, not automatically a data error.

Date Matchup Projected Final Margin error
Sep 11, 2026 Farrell vs. Reynolds 25–24 84–0 82.5 points
Sep 25, 2026 Normandy vs. Rocky River 14–28 0–83 68.7 points
Sep 18, 2026 Trotwood-Madison vs. Taft 23–17 72–0 65.6 points
Sep 26, 2026 Bedford vs. Cleveland JFK 18–17 66–0 64.8 points
Sep 18, 2026 John Adams vs. Cleveland JFK 16–18 68–6 64.4 points

Evaluated games

The game-level record behind the summary above.

Oct 2, 2026

Willard vs. Mohawk

Correct
Projected
19–27
Final
0–44
Projected winner
Mohawk
Margin error
35.3

Oct 2, 2026

WS Northwestern vs. Hillsdale

Correct
Projected
17–27
Final
0–49
Projected winner
Hillsdale
Margin error
38.8

Oct 2, 2026

Lakota vs. Gibsonburg

Correct
Projected
21–22
Final
14–28
Projected winner
Gibsonburg
Margin error
12.4

Oct 2, 2026

Millbury Lake vs. Otsego

Missed
Projected
21–23
Final
35–33
Projected winner
Otsego
Margin error
3.6

Oct 2, 2026

Indian Lake vs. Springfield Shawnee

Correct
Projected
23–19
Final
46–14
Projected winner
Indian Lake
Margin error
27.7

Oct 2, 2026

Elmwood vs. Ada

Missed
Projected
24–20
Final
53–64
Projected winner
Elmwood
Margin error
14.2

Oct 2, 2026

Eastwood vs. Oak Harbor

Correct
Projected
22–21
Final
27–13
Projected winner
Eastwood
Margin error
13.0

Oct 2, 2026

Chippewa vs. Waynedale

Correct
Projected
20–19
Final
28–13
Projected winner
Chippewa
Margin error
14.9

Oct 2, 2026

Benjamin Logan vs. North Union

Correct
Projected
22–27
Final
26–28
Projected winner
North Union
Margin error
2.2

Oct 2, 2026

Lima Bath vs. Ottawa-Glandorf

Correct
Projected
20–24
Final
7–23
Projected winner
Ottawa-Glandorf
Margin error
12.4

Oct 2, 2026

Waterloo vs. McDonald

Correct
Projected
13–29
Final
0–1
Projected winner
McDonald
Margin error
14.6

Oct 2, 2026

Tuslaw vs. Toronto

Correct
Projected
20–24
Final
28–29
Projected winner
Toronto
Margin error
3.2

Oct 2, 2026

Grand Valley vs. Leetonia

Correct
Projected
23–16
Final
40–14
Projected winner
Grand Valley
Margin error
19.9

Oct 2, 2026

Garrettsville Garfield vs. Youngstown Liberty

Correct
Projected
27–24
Final
43–14
Projected winner
Garrettsville Garfield
Margin error
26.2

Oct 2, 2026

Garaway vs. Sandy Valley

Correct
Projected
28–15
Final
55–21
Projected winner
Garaway
Margin error
21.6

Oct 2, 2026

Crestwood vs. Kirtland

Correct
Projected
21–27
Final
0–32
Projected winner
Kirtland
Margin error
25.9

Oct 2, 2026

Columbiana Crestview vs. Newton Falls

Missed
Projected
17–23
Final
49–14
Projected winner
Newton Falls
Margin error
40.8

Oct 2, 2026

Conneaut vs. Fairview Park

Correct
Projected
26–16
Final
36–21
Projected winner
Conneaut
Margin error
4.4

Oct 2, 2026

Champion vs. Ravenna Southeast

Missed
Projected
20–19
Final
14–26
Projected winner
Champion
Margin error
12.7

Oct 2, 2026

Buckeye Local vs. Martins Ferry

Correct
Projected
14–29
Final
7–49
Projected winner
Martins Ferry
Margin error
27.1

Oct 2, 2026

Berkshire vs. Rootstown

Correct
Projected
22–21
Final
47–14
Projected winner
Berkshire
Margin error
32.1

Oct 2, 2026

Beachwood vs. Independence

Correct
Projected
19–20
Final
14–20
Projected winner
Independence
Margin error
5.5

Oct 2, 2026

Wyoming vs. Madeira

Correct
Projected
27–21
Final
41–7
Projected winner
Wyoming
Margin error
28.6

Oct 2, 2026

Waynesville vs. Carlisle

Missed
Projected
24–21
Final
13–38
Projected winner
Waynesville
Margin error
27.2

Oct 2, 2026

Waverly vs. Minford

Correct
Projected
21–19
Final
48–24
Projected winner
Waverly
Margin error
22.2

How to interpret these numbers

Winner accuracy shows whether OSPR favored the team that won. Average margin error shows how far the projected point margin was from the final margin.

Brier score measures probability quality from 0 to 1; lower is better. Calibration compares predicted win probabilities with observed outcomes.

Small groups can swing sharply, so every percentage is shown with its evaluated game count.

Injuries, roster changes, weather, and information not reflected in recorded results may affect individual games.