February 12, 2026

Opposing pawns, opposing opinions

Partial 8-Piece Tablebase

Lichess releases a massive endgame answer book—fans cheer, purists panic

TLDR: Lichess released perfect answers for many 8-piece endgames with opposing pawns, now in its analysis tools and API. Commenters are split between praising a study superpower that covers about half of practical endings and fretting about lost artistry, cheating fears, and the “partial” label

The chess internet just got loud. Lichess rolled out a partial 8‑piece “tablebase,” basically a perfect endgame answer book for positions where one white pawn and one black pawn block each other—called op1. It’s live on the analysis board and mobile app, open to developers, and there’s even a jaw‑dropping 63‑terabyte download. Commenters split fast: half shouting “this is a free masterclass,” the other half clutching pearls about the “end of endgame artistry.” Purists grumble that excluding wild, lopsided pawn endings is cherry‑picking; pragmatists clap back that op1 covers roughly half of real 8‑piece endgames, which is where people actually get stuck. Old‑school cred landed too: Marc Bourzutschky (a legend in endgame theory) and Niklas Fiekas got their flowers, while meme lords declared “OP1 = OverPowered One.” Jokes flew about 63 TiB being “bigger than my life’s photos,” and someone asked if their phone would melt—reply: it’s server‑backed, you’re safe. Fans revisited classics, like Kasparov–Karpov and Petrosian–Korchnoi, sparking “human vs machine” banter when engines show wins humans didn’t find. In short, it’s not “all 8 pieces solved,” but it’s a big practical leap—and the comments are absolutely on fire

Key Points

  • Lichess released a partial 8-piece endgame tablebase covering op1 positions with opposing pawn pairs.
  • The dataset (63 TiB) is available for download, via a developer API, and within Lichess’s analysis board and mobile app.
  • Op1 includes 8-piece positions with at least one opposing pawn pair on the same file; extreme one-pawn imbalances are excluded.
  • Empirical scans (ChessBase Megabase, CCRL database, Lichess data) indicate about half of practical 8-piece endgames are op1.
  • Examples from Kasparov–Karpov (1990), Petrosian–Korchnoi (1976), and Korchnoi–Ponomariov (2001) illustrate op1 evaluations.

Hottest takes

“Congrats, but did we just delete endgame drama forever?” — rook_rolls
“63 terabytes? Finally a use for my crypto miner rig” — hashbrown
“Partial is fine—half of real endgames > perfection we never use” — practical_patzer
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.