The trust page

We grade ourselves every morning.

Every set of lineups we deliver gets scored against real box scores — best lineup, set average, top-3 average. Postponements, busts, engine upgrades: it all stays on the board. No screenshots of winners only.

Lineup sets

DateSetLineupsBestMeanTop-3 AvgNotes
2026-06-11early42_v6_delivered42164.1115.7160.8
2026-06-11early42_v7_retro42209.5149.4200.1v7 wins
2026-06-11night11_61511136.686.3126.8
2026-06-11night11_afterhours_PLAYED11149.1100.3136.8Kay PPD zeros x4
2026-06-11replay20_131393_night20146.3102.8136.9replay v10 engine
2026-06-11replay20_131399_early(top20of42)20178.8144.6174.5replay v10 engine
2026-06-10replay20_131366_main20139.197.6134.9replay v10 engine
2026-06-10replay20_131367_2game13185.7144.8180.0replay v10 engine
2026-06-09replay20_131322_main20169.7121.6155.2replay v10 engine

Daily model report card

DateTypePlayersMAEBiasRank Corr
2026-06-11Pitchers616.23-16.23-0.551
2026-06-11Hitters546.59+0.320.400
2026-06-10Pitchers2812.48+1.480.307
2026-06-10Hitters2327.79+1.650.093

Engine upgrades are validated retroactively against the boards we actually delivered — when v7 beat the delivered v6 set 209.5 to 164.1 on the same slate, that comparison went here too. Updated daily by the pipeline.