Start / Sit Optimizer
Import your Sleeper or ESPN roster, or type one in, and get your best lineup
Read-only, and no login: Sleeper’s API is public, so a username is all it takes. Your league’s roster slots and scoring rules come across with it, which is what makes the advice yours rather than generic half-PPR.
How this works. Each player’s week projection starts from the JOAS season projection divided by 17, re-scored to your league’s rules rather than to a preset. It is then nudged by how the defense they face has handled that position — at half the raw strength of that signal, and capped at ±12%, because points allowed carries a defense’s own schedule inside it and deserves less confidence than it looks like it has. Byes score zero; players ruled out score zero; questionable and doubtful are discounted. The best lineup is solved exactly rather than filled greedily, which matters the moment your league has overlapping flex slots. Where the numbers are weakest: per-game is an average, not a forecast of this specific game — no snap-count news, no weather, no game script. Confidence percentages assume the two players are independent, which is wrong when they play each other. Kickers and defenses. Neither is in the projection engine, so both are projected from their team instead: a kicker from his own offense’s touchdowns and field goals, a defense mostly from how good the offense across from it is, which is the one thing that genuinely predicts a defense’s week. Treat the kicker number as “about eight” and nothing more — weekly kicker scoring is close to unpredictable, and since it’s a one-of-one slot there is no decision to make anyway. IDP is not projected and is left where you had it. Players whose game has already kicked off are locked in place, the same way your own platform locks them.