Omaha 0.9.0
Released 3 September 2026
Players are now simulated for the team they play for, and a player who changed clubs in the off-season brings his role with him. Rush attempts improved by nineteen points.
How to read these numbers
Every figure below compares the simulation against a player’s own season average so far, scaled to how much of the game is left, roughly what a reader could work out without us. Zero means we add nothing over that. Positive means we beat it; negative means we are worse than it.
Measured across the 2024 season, weeks 1–18, rebuilding the model each week using only what was known at the time. 400 simulations per situation.
The change column is simply this version’s accuracy minus 0.8.0’s, each measured across everything that version covers, so the two columns you can see always account for it. This release also changed which situations can be measured at all, so a like-for-like comparison on shared situations would describe only part of what changed.
Accuracy by stat
| Stat | vs. season average | Change from 0.8.0 |
|---|---|---|
| Receiving touchdownsrare event | +20.4% | +2.01 |
| Rushing touchdownsrare event | +17.6% | +4.64 |
| Receiving yards | +8.9% | +5.68 |
| Passing yards | +7.9% | −2.00 |
| Passing touchdownsrare event | +7.7% | −0.13 |
| Pass attempts | +6.6% | −1.73 |
| Rushing yards | +4.4% | +8.85 |
| Completions | +3.9% | −2.89 |
| Receptions | −0.6% | +9.39 |
| Targets | −7.1% | +8.45 |
| Rush attempts | −8.0% | +18.88 |
| Average across all figures | +5.6% |
The average is an unweighted mean across the figures above. It is a summary of how we are doing, not a single score for the model: yards and counts are not measured in the same units, so combining them any more cleverly than this would just let passing yards decide the answer.
This release changed which situations can be measured at all, it produces a usable figure for players it previously had nothing to say about. Ranges are therefore not shown: the only honest comparison is between the two accuracy columns, each measured across everything that version covers.
What changed
The previous release ruled out players who could not play at all, but did not notice that a player could be perfectly available and simply playing for someone else. A quarter of the snaps in our history belong to players who have since moved, and we were still simulating them for the team they left. We now read each team's own game-day roster.
34.4% of the player-team pairs in the training window had moved, carrying 28.3% of historical snaps. The share of predicted carries reaching players who touched the ball rose from 67.6% to about 90%.
Fixing the above on its own made things worse, and the reason is worth stating: a departing player's workload was handed to whoever stayed, because the player who actually replaced him had no history with his new team and so was not simulated at all. A back's role is now recorded against the player rather than the club, so a signing arrives with his own workload from week one.
Measured on the opening week, the first approach predicted 32% more carries than actually occurred. With roles attached to players, that error is under 2%.
These depend entirely on the ball reaching the right player, so they gain the most.
Rush attempts improved 18.9 points, receptions 9.4, rushing yards 8.9, targets 8.5, receiving yards 5.7. Rushing yards are now better than a player's own season average (+4.4%), joining receiving yards (+8.9%).
Identifying a starting quarterback from a club's own roster rather than from two seasons of history fixes the case where a team changed quarterback in the off-season, previously we fielded last year's starter through the early weeks.
The correct starter is now simulated 83.7% of the time, against 65.2%.
What this version still gets wrong
- The four passing figures each fell by 0.1 to 2.9 points, and the reason is coverage rather than quality. We now project nine more starting quarterbacks, rookies and mid-season replacements we previously ignored entirely, and we project them less well than established starters. Compared on the quarterbacks both versions covered, passing is unchanged (between −0.8 and +0.6 points). We publish the figure including the harder cases because that is what a reader actually gets.
- Rush attempts remain the least accurate figure we publish at −8.0%, with targets at −7.1% and receptions at −0.6%. Five releases have improved rush attempts by 33.5 points in total.
- A player who moves clubs brings the role he had at his old one. That is a reasonable starting point and much better than nothing, but it will understate a player signed to do more than he used to, and overstate one signed to do less.
- Team quality still barely affects running plays. A better rushing offense does not reliably produce better runs.