Omaha 0.5.0
Released 2 September 2026
The first version measured end to end by walking forward through a full season, one week at a time, against what a reader could work out from a player's own season average.
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.
Accuracy by stat
| Stat | vs. season average | |
|---|---|---|
| Receiving touchdownsrare event | +15.1% | — |
| Rushing touchdownsrare event | +12.5% | — |
| Passing yards | +6.9% | — |
| Passing touchdownsrare event | +6.9% | — |
| Pass attempts | +2.0% | — |
| Completions | +1.3% | — |
| Receiving yards | −4.5% | — |
| Rushing yards | −12.4% | — |
| Receptions | −21.6% | — |
| Targets | −27.4% | — |
| Rush attempts | −41.5% | — |
| Average across all figures | −5.7% |
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.
Each change carries the range the measurement can actually support. A change whose range crosses zero is shown in neutral rather than as a gain: we cannot tell it apart from no change, and we are not claiming it as one.
What changed
Introduced the measurement these pages report. For each week of the 2024 season the model is rebuilt using only what was known at the time, then scored against what actually happened. The comparison is a player's own season average to date, scaled to how much game is left. The figure a reader could reasonably work out without us. Zero means we add nothing over that; positive means we beat it.
18 weeks, 269 games, roughly 324,000 scored player-situations per version.
A ruled-out player is given a zero share and the remaining players are redrawn, which promotes a backup from the data rather than from an assumption. An earlier approach redistributed the missing player's workload across everyone on the field, which sent an injured back's carries to the quarterback.
The players on the field are now drawn as a group that fills the slots the formation names, rather than each player being flipped independently. The previous approach produced formations with no running back on 29.8% of snaps.
What this version still gets wrong
- Rush attempts are the weakest figure at −41.5%, meaningfully worse than a player's own season average.
- Scoring runs about 9% low, and rushing touchdowns are under-predicted.
- Play calling does not respond to the score.