Week of Sep 28 to Oct 4, 2026 (2026-W40) · model clean_v2 · generated Oct 5, 2026
| Moneyline split | Record | n |
|---|---|---|
| Model's side was the home team | 4–2 | 6 |
| Model's side was the road team | not scored: below the n=5 floor (n=0) | |
| Model's side was the market favorite | 3–2 | 5 |
| Model's side was the market underdog | not scored: below the n=5 floor (n=1) | |
| Coin-flip calls (50-55%) | not scored: below the n=5 floor (n=2) | |
| Moderate calls (55-65%) | not scored: below the n=5 floor (n=3) | |
| Strongest calls (65%+) | not scored: below the n=5 floor (n=1) | |
| Games with a total under 8.5 | 4–2 | 6 |
| Games with a total 8.5-9.5 | not scored: no opening total captured for these games (n=0) | |
| Games with a total above 9.5 | not scored: no opening total captured for these games (n=0) | |
| Totals split | Our avg miss | Market's | n |
|---|---|---|---|
| Projection sat above the line | 3.72 | 2.42 | 6 |
| Projection sat below the line | not scored: no opening total captured for these games (n=0) | ||
| Lines under 8.5 | 3.72 | 2.42 | 6 |
| Lines 8.5-9.5 | not scored: no opening total captured for these games (n=0) | ||
| Lines above 9.5 | not scored: no opening total captured for these games (n=0) | ||
| Confidence bucket | We said | It happened | n |
|---|---|---|---|
| Coin-flip calls (50-55%) | not scored: below the n=5 floor (n=2) | ||
| Moderate calls (55-65%) | not scored: below the n=5 floor (n=3) | ||
| Strongest calls (65%+) | not scored: below the n=5 floor (n=1) | ||
League base rates for the same games: home teams went 4–2 (n=6); market favorites went 3–3 (50.0%, n=6); overs went 1–4 against the line (n=6). Every model result above sits inside that backdrop.
The three largest totals misses account for 74.2% of the week's total absolute totals error (16.6 of 22.3 runs across 6 games).
SD @ MIL (Oct 3): we had the home side at 60%; the home side won.
ATL @ LAD (Oct 4): we had the home side at 65%; the away side won. It stays in the record.
One week is a small sample. These are counts of where the results came from, not conclusions about skill; the season-long record is the number that matters.
The point of a public track record is that the misses stay in. Our three largest totals misses this week:
| Game | Projected | Actual | Off by |
|---|---|---|---|
| NYY @ TB | 8.2 | 1 | 7.2 |
| CWS @ CLE | 8.6 | 3 | 5.6 |
| ATL @ LAD | 8.8 | 5 | 3.8 |
| Game | Proj total | Actual | Miss | Home win% | ML side |
|---|---|---|---|---|---|
| PHI @ ATL | 8.9 | 8 | 0.9 | 0.52 | Hit |
| SD @ MIL | 8.3 | 5 | 3.3 | 0.60 | Hit |
| CWS @ CLE | 8.6 | 3 | 5.6 | 0.50 | Miss |
| NYY @ TB | 8.2 | 1 | 7.2 | 0.59 | Hit |
| ATL @ LAD | 8.8 | 5 | 3.8 | 0.65 | Miss |
| SD @ MIL | 8.6 | 7 | 1.6 | 0.57 | Hit |
Across 11,972 final regular-season games with a standard run line, the moneyline favorite laying −1.5 covered 42.8% of the time and won outright 58.3%. We publish no run-line projection, because the market prices this one efficiently, so this is the market's record, not ours.
For our moneyline probabilities, Brier squares the gap between the win probability we published and what happened (1 or 0), then averages it. 0.250 is a coin flip; lower is better. Reported live, never smoothed.
How our projections did against the market. One email a week, no picks, no spam, unsubscribe anytime.
We store only your email to send the recap. Analytics, not picks. Nothing here is betting advice.