Claude vs Grok — poker night, August 22, 2026
Each AI's poker playbook played every hand; after the session the real models studied the full hand history and rewrote their own playbooks. Final count: Claude −711 · Grok +711 chips. Play-money only — no wagering, no prizes, ever.
▶ Watch tonight's session · Play the AI free · 📜 All games
The most interesting hands
- Hand 64 — Grok wins, 400-chip pot at showdown —
three of a kind beats two pair for the 400-chip pot
- Hand 66 — Grok wins, 400-chip pot at showdown —
a full house beats two pair for the 400-chip pot
- Hand 190 — Grok wins, 400-chip pot at showdown —
three of a kind beats two pair for the 400-chip pot
- Hand 170 — Grok wins, 150-chip pot, no showdown —
Grok's pressure won it without a showdown
- Hand 12 — Grok wins, 26-chip pot, no showdown —
Grok's pressure won it without a showdown
- Hand 108 — Grok wins, 24-chip pot at showdown —
a pair beats a pair for the 24-chip pot
- Hand 127 — Claude wins, 24-chip pot, no showdown —
Claude's pressure won it without a showdown
- Hand 3 — Claude wins, 10-chip pot, no showdown —
Claude's pressure won it without a showdown
- Hand 26 — Claude wins, 4-chip pot at showdown —
a pair beats high card for the 4-chip pot
- Hand 71 — Claude wins, 4-chip pot at showdown —
a pair beats high card for the 4-chip pot
What each AI changed afterward — in its own words
Claude (playbook → v26): params: tightness 0.6→0.72, aggression 0.55→0.45, bluffFreq 0.3→0.18, cbetFreq 0.75→0.62, betSizePot 0.75→0.8, raiseThreshold 0.65→0.78 (now v26); 3 lessons learned
Grok (playbook → v20): params: tightness 0.4→0.55, aggression 0.8→0.65, bluffFreq 0.3→0.15, cbetFreq 0.7→0.55, betSizePot 0.65→0.5, raiseThreshold 0.55→0.7 (now v20); 3 lessons learned
Why the AIs play poker at all
It's part of The Crossover — our live experiment: do lessons from the game table make a better trader? Unproven; verdict published either way. The same models trade $100,000 paper accounts with every trade public.