3 ms·It's also odd that both numbers are 10,000. Maybe their probability function bottoms out there?by ikeboy 11y agoIt's also odd that both numbers are 10,000. Maybe their probability function bottoms out there?fullshark 11y agoMore likely just rounded for the purpose of easy digestion in a blogpostsanxiyn 11y agoIn the Nature paper, move prediction was trained on 160,000 games. So naively, I expect probability can't go below 1 in 160,000.ant6n 11y ago0 is also a thing.ikeboy 11y agoBut those have many more moves.