The Window. This season, once a week is final, is the team's official points a week against the league average (100). Each season ahead blends the roster projected forward (its best 9 players by points a game, the ones who start) with how the team actually scores, because a GM's lineup and lock calls carry over and a roster sum cannot see them: half and half next season, 70% roster two seasons out, since results say less the further out you look. Early in the year that second half leans on last season's results and moves to this season's in proportion to the games played, because the evidence was full seasons, not a week or two. Before any week is final, this season is built the same way from last season's results; a new GM has only his roster. The evidence: on this league's 2024 official scores, points so far predicted the rest of a season far better than the roster from the second week on; across 534 other Sleeper dynasty leagues (about 6,200 returning teams), those blends read each coming season better than the roster alone and better than the results alone, with every gain above zero at 95% confidence, and each mix is the best one measured or within its range. Rising, Steady or Closing follows the roster projection alone, so it is about the roster, not this season's luck or lineup calls; across those leagues the roster's projected change moved with the team's actual change in about four leagues of five.
This season. A player's points a game this season, steadied by last season's while the sample is small: his games so far count against 5 games of last season (last season first carried one year on the aging curve below, since he is a year older). Tested on 2025, that predicted the rest of each player's season better than the other weights tried (0 to 30 games). Only a player out for the season (a season-ending injury, or out indefinitely) counts nothing this season; a player missing a game or two keeps his value. If he played fewer than 5 games last season, only those count.
Seasons ahead. Each season, every player's points move with his age, the way Da League scoring actually moved from 2023-24 to 2024-25: about +15% at 19, +7% at 21, +3.5% at 22, level at 23, then a slow fade: about 4-6% a year through the late 20s, 6-8% in the early 30s. That is the typical path. Around a quarter of young players jump far more in a season, and a quarter fall back.
Rookies and picks. A rookie who has barely played (under 10 games this season and last), and every future pick a team holds, counts as what Da League rookies have actually scored: a top-three pick about 25 a game as a rookie and 31 after, other firsts about 20, seconds 10-12, thirds 5-6. A held pick counts as a top-three pick only if it is next year's and the team it comes from projects to finish in the bottom four. Most rookies don't crack a starting nine for two years, so picks matter here mostly as trade value.
Tiers. Super Team: a window index of 115 or more and at least 35% to win the title. With playoff odds on file, the odds decide who is contending (60% or better; before any odds, a window index of 100 or more). Among contenders: Last Dance for the one oldest lineup (starters averaging 29 or older, weighted by points) whose window is closing, then All In for a team in the bottom three for pick value, then Contender. Frisky is 35-60% odds. Below that: Trust the Process (window index 80 or lower, and rising or in the top two for pick value), Rebuilding (rising, or in the top three for pick value), and Stuck in the Middle for everyone else.
Last 60 days. Every completed trade, measured by what it did to the starting lineup this season (more bodies only count if they start) and to pick value. Buying gave up pick value for a better lineup now, Selling the reverse. The season-by-season path below each card uses the same rule for each whole season.
What the numbers can't see. Injury news after the last sync, real NBA trades and role changes, and which young player is about to break out.