ROS Rankings
Plan for the rest of the season, including the playoffs. Rank players by their simulated championship value. Compare trade prices and every remaining matchup at a glance.
League and playoff settings
Loading published rankings…
Baseline + matchup/context = points if active. Weekly projections also include the chance of playing. Tap Why? for that breakdown. Small championship-value gaps can be simulation noise.
| Loading rankings… |
How the defense model works, and what we've tested
D/ST points come from sacks (1), interceptions and recovered fumbles (2), touchdowns (6), safeties (2), and the points-allowed scoring tiers. We model sacks and turnovers using both the defense's recent rates and the opposing offense's rates. Points allowed come from the opponent's implied score.
For posted games we use sportsbook spreads and totals. For later weeks we estimate those lines from team offense and defense ratings, with last season as a prior. A future matchup is an estimate, not a future sportsbook line. The baseline keeps the defense's own rates and team ratings, replaces the opponent with league averages, and balances home and away.
Recorded historical backtests: weekly error 5.35 points RMSE and rank correlation 0.30 (2015–2025); ROS rank correlation 0.42 versus 0.32 for trailing D/ST scoring (2018–2025). Those tests support useful signal, but not precise ranks. Leave-one-season-out tests also aren't the same as a forward-only deployment test. We haven't validated the playoff championship ordering against real fantasy titles.
The championship simulation lets both sides stream from the same projected waiver pool. That measures the benefit of owning a defense over streaming, so it can differ from the order by total projected points. Waivers are a shared replacement approximation; the model doesn't simulate competing claims or transaction limits. Close title values should be read as ties.
How the ROS rankings are built
Championship value sets the order. We simulate the remaining regular season and playoffs, then replay each league with a player replaced by a waiver pickup. His value is the change in championship chance, in percentage points. The standard format has 12 teams in two randomized divisions, six playoff qualifiers and week-15 byes for the division winners. Semifinals are week 16 and the final is week 17.
Published rankings. Full PPR, half PPR and standard scoring, in 1QB and superflex, are calculated ahead of time and refreshed hourly. Every ranking and weekly projection on the page comes from the same snapshot. Custom roster and playoff settings are calculated on your device and saved for that snapshot.
Weekly points. Projections combine player usage, historical production, sportsbook props, future opponents and the chance of playing. Greener cells mean higher expected points compared with the same position. A bye has no points. Playoff weeks are marked in the header.
Trade value. This separate column shows what recent public redraft trades imply a player is worth, blended with our projection. Its scale stays on the published 12-team PPR format for 1QB or superflex. Trade prices do not set the order here.
What the simulation assumes. Teams start with equal records and rosters are randomized. Managers can stream waiver replacements. These rankings have not been shown to outperform trade-market rankings or calibrated against actual fantasy championships. The uncertainty shown with championship value measures simulation noise, not all forecast uncertainty.