How the values work
Every number on FFPilot comes from one model. This page explains each step, the exact constants we use, where the inputs come from and what the model does not know. If something here disagrees with what a tool shows you, the tool has a bug: tell us.
1. Inputs
- Weekly projections for every week from the current one through week 17, from Sleeper's projections feed (Sleeper labels the provider as Rotowire). They include full stat lines (yards, touchdowns, receptions, turnovers), so we can rescore them for any format.
- Box scores for completed weeks, from Sleeper's stats feed. Used for season averages, recent form and defense-vs-position.
- Player details (team, age, depth chart, injury designation), from Sleeper's player list, once a day. Injury designations for projected players also refresh hourly with the projections.
- Trending: how many Sleeper leagues added or dropped each player in the last 24 hours and 7 days.
- Sleeper draft position (dynasty only), as a market check.
Kickers and team defenses appear in start/sit and on the waiver wire, but they do not get trade values: nobody trades for them, and their week-to-week scoring is mostly noise.
2. Scoring
We score every stat line ourselves: 1 point per 25 passing yards, 4 per passing touchdown, minus 1 per interception, 1 per 10 rushing or receiving yards, 6 per rushing or receiving touchdown, minus 2 per lost fumble, 2 per two-point conversion, plus 1, 0.5 or 0 per catch for PPR, half PPR and standard. TE premium adds 0.5 per tight end catch. Scoring is linear, so the sum of weekly projections scores the same as scoring each week.
3. Replacement level
A player's worth depends on what you could get for free. We fill every team's lineup (QB, 2 RB, 3 WR, TE, one RB/WR/TE flex, plus a second QB-eligible flex in Superflex) with the best rest-of-season projections, taking the best remaining player for each flex slot. The replacement level at a position is the average of the next three players after the last starter.
Right now, in a 12-team PPR league, replacement level over the rest of the season is about 148.5 points at QB, 55.0 at RB, 54.0 at WR and 74.7 at TE, with 27 running backs, 42 receivers and 15 tight ends in starting lineups league-wide.
4. From points to a 0-100 value
Points over replacement are divided by the best player's total and raised to the power 1.3 (1.2 in dynasty). The curve is convex on purpose: a player 50% as good as the best is worth about 40, not 50, which is how real leagues trade. Players below replacement keep a small value, at most 1.2, that shrinks with distance below the line. In redraft, Sleeper add/drop activity moves a value by at most 6%, because adds react to news before projections are updated. Finally, values are rescaled so the top player is 100.
5. Trades and consolidation
A trade compares the total value each side receives. When the sides receive a different number of players:
- The side receiving more players pays for the roster spots: each extra player (its lowest-valued ones) loses 60% of a waiver-level player's value, currently 6.8 in a 12-team league. Depth still has some use, so enough real starters will outweigh a stud, while piling on bench players barely moves the total.
- The side receiving fewer players earns a consolidation bonus if it gets the best player in the deal: 12% of the gap between the two sides' best players.
Verdicts: under 6% apart is a fair trade, 6 to 15% a slight edge, 15 to 32% a clear winner, more than that lopsided. The balancing suggestions are players whose value is close to the gap.
6. Dynasty
Dynasty value is this season's points over replacement plus four future seasons. For each future season we take the player's current per-game level (65% projection, 35% this season's actual average when he has played three or more games), multiply it by the change in his position's age curve, subtract replacement level per game, count 15 games, and discount by 20% per year. Growth for young players is capped at 25%.
| Age | 22 | 24 | 26 | 28 | 30 | 32 | 34 |
|---|---|---|---|---|---|---|---|
| QB | 80% | 91% | 98% | 100% | 100% | 100% | 95% |
| RB | 92% | 100% | 97% | 80% | 55% | 30% | 20% |
| WR | 86% | 97% | 100% | 97% | 83% | 62% | 40% |
| TE | 74% | 89% | 100% | 100% | 90% | 72% | 48% |
The curves are our own estimates of typical career arcs, not fitted to a single study. The model value is then blended with Sleeper dynasty draft position (30%, or 60% for players with almost no production to project from, such as injured rookies), converted to the same scale with an exponential curve.
7. Start/sit
This week's projection in your scoring, times a matchup factor of up to plus or minus 5% from defense-vs-position ranks, times availability: 0.9 for questionable, 0.3 for doubtful, 0 for out, injured reserve, suspended or on a bye.
8. Known limits
- Projections come from one provider. They are good, but every projection misses on touchdowns and on role changes that happen mid-week.
- Values do not know your roster. A fair trade can still be bad for a team that already has three tight ends.
- Rookie draft picks are not valued yet.
- We assume standard Sleeper-style scoring. Leagues with large yardage bonuses, six-point passing touchdowns or individual defensive players will see different numbers.
- Projections run to week 17. Leagues that play a week 18 championship lose a little rest-of-season value for players with a good week 18 matchup.
Snapshot: 2026 season, week 8, generated Tue, 27 Oct 2026 18:23:20 GMT.