How the values work

Every number on this site comes from our own model, built on open data and rebuilt every night. This page explains each step, including where the model is weak. Current values: November 14, 2026, model version 0.1.0.

1. A projection for every player

For each quarterback, running back, wide receiver and tight end we project a per-game stat line: yards, touchdowns, receptions and turnovers. It is an average of the player's games over roughly the last three seasons, weighted so that a game's influence halves about every 14 game-weeks, with extra decay across each offseason.

Small samples are pulled toward a baseline. A veteran with few recent games is pulled toward the average for the position. A first-year player is pulled toward what rookies drafted in the same range have produced, so draft capital counts until real games replace it.

2. Points under your league's scoring

The site stores stat lines, not fantasy points. When you change PPR, tight end premium or any other scoring setting, every player is re-scored in your browser. That is why a reception-heavy back moves when you switch from PPR to standard.

3. Value over replacement

A player matters by how much better they are than someone you could get for nothing. We fill every starting lineup in your league (set positions first, then flex, then Superflex), add typical bench depth, and take the next player at each position as replacement level. Adding a Superflex slot or more teams pushes that level down, which is what raises quarterbacks in Superflex.

Quarterbacks get two extra rules. A backup who is not playing does not count as a free quarterback, however good his numbers would be if he started. And in Superflex, starting quarterbacks are scarce: we count how many are left over once every lineup in your league is full. A team that loses a quarterback has a spare one only that often, and otherwise has to start a running back, receiver or tight end in the Superflex slot. The fewer spares your league has, the more every starting quarterback is worth, which is why they top Superflex rankings and matter less in 1QB leagues.

The best player under your settings is set to 10,000 and everyone else is scaled to match.

4. Dynasty: future seasons and age

Redraft value counts the rest of this season. Dynasty value adds nine future seasons, each discounted 18% a year because the future is uncertain, and each adjusted by an age curve for the position. The curves come from how players' per-game production changed from one season to the next since 2010, counting players who dropped out of the league as zero.

5. A check against the market

Where current draft data exists for your format, a player's draft position is translated into a value and blended in at up to 35% weight. Old draft data counts for less: its weight halves every 45 days. When drafters and the projection disagree, the player's likely range gets wider.

6. Draft picks

For each draft class from 2011 on, we order the skill players roughly the way a rookie draft goes and measure what each slot's player went on to produce over six seasons. Over the same years we measure what established players went on to produce, by where they ranked going into that draft. A pick is worth what the equally productive veteran is worth today: historically the first pick of a Superflex rookie draft has produced like the 11th-ranked veteran, and a late first like one ranked around 60th. Future-year picks are discounted 12% a year, and picks carry the widest ranges of any asset.

7. Uneven trades and roster spots

Two players for one is not a plain sum, because the team taking two has to make room. Its weakest starter moves to the bench and its last bench player is dropped, while the team taking one gets an open spot to fill. So when one side receives more players and picks than it gives up, its least valuable extra pieces are counted for less: in full only for the part above your league's last starter, at half between the last rostered player and the last starter, and not at all below that.

Both levels come from your league's size and lineup, so there is nothing to tune. A smaller league has better players on waivers and a spot costs more. A draft pick counts as a roster spot too, because it turns into a player, so a player for a pick is an even swap. A trade with the same number of pieces on each side is a plain sum. The verdict shows how much came off the larger side.

8. The verdict and its confidence

Every asset has a likely range, not just a number. We treat each side's total as uncertain, and the confidence figure is the share of scenarios in which one side comes out ahead. A trade is called fair when the gap is under 5% or the leader is ahead in fewer than 60% of scenarios; it is a clear win at 75% or more.

How accurate is it?

We test the projection on past seasons using only what was known at the time, and compare it with the simple approach of using each player's average so far. Error is the average miss in PPR points per game.

The model is better in all three tests, by about 7% to 13%. For running backs mid-season it is no better than the simple average.

Known limits

• Weeks a player sits out do not age his earlier games, and missed games reduce future seasons by at most 40%, so a long-term injury lowers a dynasty value without erasing it.
• There is no injury report feed. A player's availability is inferred from recent games and roster status, so a star returning this week can look undervalued in redraft.
• Dynasty draft data is thin during the season, so dynasty values lean mostly on the projection and age curves right now.
• Coaching changes, depth chart moves and offseason trades only show up once games are played.
• Kickers, defenses and individual defensive players are not valued.

Data sources

• Player statistics, rosters, ages and draft history: nflverse, licensed CC-BY 4.0.
• Draft position data: Fantasy Football Calculator.
• Player ID matching for league import: the DynastyProcess ID table.