Loading...
No backtesting. No cherry-picked results. Live accuracy tracking from Week 1: every call logged, every result verified.
The 2026 NFL season kicks off in September. From Week 1, every SnapCount ranking and Start/Sit call is recorded before kickoff and scored once the results are in. No retroactive adjustments.
Take every pair of players where we rated one above the other. How often did the higher-rated player actually outscore the lower one? Ties and same-rating pairs are excluded rather than counted as wins.
How far each weekly projection landed from the actual result, in fantasy points. Lower is better. Averaged across every call we made.
Average points actually scored by each tier. Elite should outscore Starter, Starter should outscore Flex, and so on. If the ladder inverts, the tiers are not doing their job.
How closely the full ranking order matched the order players actually finished in that week. +1 is perfect, 0 is random, negative means inverted.
Every weekly ranking and Start/Sit call is recorded automatically on Thursday morning, before any game kicks off. What gets scored is exactly what was published.
After each week, actual fantasy points are compared against those recorded calls. No cherry-picking; every call counts. Players on bye or who did not play are excluded, not scored as zero.
Season totals update as each week is scored. Nothing is ever revised after the fact. A recorded call cannot be edited once its games have started.
Any model can be tuned to look good on historical data. Backtested accuracy is meaningless because you can always adjust parameters until the past looks great. We only track live, forward-looking predictions made before games start. That's the only honest way to measure accuracy.
Sign up free and make your own Start/Sit picks. Your accuracy gets tracked too. See how you stack up on the leaderboard.
Start Free