Observed arrival · 2026-11-17
NeuroDraughts: a draughts player that learns by playing itself
A playable draughts engine with in-browser training, where the learner improves by playing copies of itself.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Draughts players curious about machine learning
- Worth noticing
- Each saved champion plays 40 games against the built-in June 1997 network, providing a recurring comparison point.
Field notes
The training panel exposes look-ahead settings of 1, 3, 5, or 7 plies, plus learning rate, exploration, trace decay, and batch-size controls. A champion is cloned after reaching a configurable test score; the page says each saved champion plays 40 games against the built-in June 1997 network. Finished games can enter a public record under a player name, with an opt-out.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue