Fantasy Football Trade Calculator

1. Points per game. Redraft uses each player's 2026 points per game in the scoring you pick, from the same stats as our player pages. To keep a big game or two from swinging things, that average is blended with 3 extra games at replacement level. Dynasty starts from this 2026 number too, then mixes in prior seasons as described below.
2. Points above replacement. Replacement level is about the best player you could find on waivers in a 12-team league with 15-man rosters: the 20th QB (32nd in superflex), 48th RB, 60th WR and 18th TE by points per game (at least 3 games played). Players at or below that line are worth 0.
3. Redraft multiplies points above replacement per game by the games each player has left through Week 17, with byes removed. A player on IR is counted as missing at least 4 games.
4. Dynasty rate. Instead of using 2026 alone, we build an established rate from 2023, 2024 and 2025 (only seasons with at least 6 games). More recent years count more, and we lean toward the player's better seasons so one down year doesn't erase a track record. That rate is blended with 2026, and the current season only counts as heavily as the number of games played, so a 7-game slump doesn't outweigh three prior years. For players with at least two solid seasons (their better years clearly above the waiver wire), we also keep a floor based on those better seasons, discounted, so a proven starter doesn't fall to zero. A single quiet rookie year is ignored, and players with no useful prior seasons, including most rookies, stay on the 2026 rate. Injured players still lose games from this season only (IR is at least 4 games), not from future years.
5. Dynasty projection. That per-game number is then used over three seasons, weighted 100%, 75% and 55%. Value drops each year past a peak age (RB 26, WR 28, TE 29, QB 32), and rookies and second-year players drafted in the first two rounds get a 10% to 20% bump.
6. Rookie draft picks (dynasty only). We looked back at the 2021 through 2025 NFL drafts in nflverse's draft data. In each class, the QBs, RBs, WRs and TEs were put in NFL draft order and treated as a 12-team, 4-round rookie draft, so the first skill player taken is pick 1.01. A pick is worth the average of the better half of those outcomes: what our model values the top 50% of players drafted in that group at today, counting 0 for anyone without 2026 stats. Using the better half, rather than the plain average of every outcome, leans toward each slot's hits. Picks are grouped into early (1 to 4), mid (5 to 8) and late (9 to 12) in each round, about 20 players per group. If a later group came out higher, the two are averaged together so a pick is never worth more than the one before it. Because an earlier pick can take any player a later one could, we also require at least a 15% drop from one group to the next when the sample doesn't show one, so an early 1st stays a clear tier above a mid 1st. For rounds 2 through 4, a five-class sample is full of zeros, which understates the chance of a hit and the value of a cheap rookie contract. Those picks are floored at the late 1st, stepped down 15% at each later group, unless the historical value is already higher. 2028 picks count 85% and 2029 picks 72%, because they're further away. "2027 1st (slot unknown)" is the average of the early, mid and late 1st. Picks are hidden in redraft.
7. Scale. In each setting the top player is 10,000. When one side has more than one player, its second-best piece counts 65%, its third about 42%, and so on, because one great player is worth more than two good ones who take two roster spots.

Limits: these aren't market prices, rookies and players with only a few games are hard to value, a full down year still counts, so a star coming off two quiet seasons won't jump back to the top of the board, pick values rest on five draft classes, so a couple of hit or miss classes can swing a slot, and role changes show up after the next stats refresh.