{"name":"christaylor-ai","version":"3.25.4","notes":"The engine plays about +22 Elo stronger at depth 8, on an interval of +2 to +42 that excludes an even result, over 600 games from random openings. Two changes, and the second only works because of the first. Captures are now ordered by static exchange evaluation, what a capture is worth once both sides have taken everything they can afford to, rather than by what was taken, and quiescence no longer follows a capture the exchange already says loses material; completing depth 7 on a crowded middlegame costs 4.2 million nodes instead of 11.6. With that in place the pruning that was switched off in 3.14.0 for winning no games is switched on for the one search whose answer is a move, at depth 8 only, and off again for everything that publishes a number. The engine also refuses the third repetition while it thinks it is ahead, which is how a won queen ending in the archive was drawn.","engine":"zig-wasm","mcpProtocolVersions":["2024-11-05","2025-03-26","2025-06-18","2026-07-28"],"endpoints":{"mcp":"https://christaylor.ai/mcp","a2a":"https://christaylor.ai/a2a"}}