DraftSOS Accuracy by Season
Loading season data...
Full Comparison Table
| Position | DraftSOS MC | ESPN FPA | FP ECR | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Tier | MAE | Hit | Tier | MAE | Hit | Tier | MAE | Hit | |
| Loading accuracy data... | |||||||||
Methodology
Tier accuracy measures whether our model correctly classified a matchup as easy (def_rank 24+), medium (11-23), or hard (def_rank 10 or below) — matching the color coding in the SOS grid. MAE (mean absolute error) measures average distance between predicted and actual defensive rank. Hit rate measures how often the predicted tier matched the actual outcome. All metrics are calculated from out-of-sample backtesting — no hindsight fitting.
We show our receipts.
DraftSOS Monte Carlo predictions backtested against actual NFL outcomes across 5 positions and 3 seasons. Every number on this page is verifiable.
"Strength of schedule projections have very little correlation to actual results. Nearly half of NFL teams moved 8+ spots in SOS ranking from preseason to actual."
— ESPN Fantasy Research, published analysis
ESPN is right — static FPA models fail because they treat last year's defensive performance as next year's prediction. DraftSOS addresses this directly: 10,000 Monte Carlo simulations model the variance that static rankings ignore, producing floor/median/ceiling ranges instead of single-point guesses.
Tier Accuracy by Position
Loading accuracy charts...
See the model in action.
The same Monte Carlo engine powers the free SOS grid. Try it now — no account required.
Accuracy metrics use MAE, tier classification, and hit rate methodology adapted from open-source fantasy analytics research. Open source credits