As promised, here’s the game between the new CrazyStone DeepLearning first edition against Zenith Go 4. I decided to have full 5 matches even if one of them won first 3 matches. Many AI now trying to implement deep learning after the successful AlphaGo match against Lee Sedol (9p). I hope Zenith Go will release their own Deep Learning version. Before I show you the games, have you watched the game between Hajin Lee (3p) agains CrazyStone DeepLearning? it’s quite a disappointing match, where CrazyStone made quite a couple mistakes and chose the wrong joseki in the 3 corners, but that’s the weirdness of bot, whenever they feel behind they play like a kyu.
Here’s the game and the record of the match between Hajin Lee and CrazyStone DeepLearning.
The result is as expected, CrazyStone won the game however the first two matches were close. With all 5 games, Zen was building huge moyo but later destroyed by CrazyStone-DL, Zen seems hate and avoiding Ko, CrazyStone-DL attachments and attack was more humanely and Zen was not playing solid like the previous Zen agains CrazyStone-2013. Here’s the summary of the game:
- Game 1 – CrazyStone W+3.5
- Game 2 – Zen 4 W+4.5
- Game 3 – CrazyStone W+R
- Game 4 – CrazyStone B+R
- Game 5 – CrazyStone W+R