Experience file:
I did a thorough test with the SugaR AI 2.50 learning file on Playchess.
A 500 MB learning file (link, see SugaA AI 2.50 topic) was posted here in the forum. I had one of my own with 50 MB. When I merged both learning files, the new size was 540 MB. I used it to play about 200 games on PlayChess with my old Intel laptop (only 860 kns). The result wasn't bad, but it was worse than my current test with a current stockfish. With the learning file and SugaR AI I had a slightly negative balance. With the same (own) book and the new Stockfish 220122 (own compile) on my old laptop (860 kns in the starting position) I am after more than 300 games in a row with only one loss and two wins! Thats fantastic.
You don't need a learning file, a current engine is better than an old engine like SugaR AI with a learning file.
I did a thorough test with the SugaR AI 2.50 learning file on Playchess.
A 500 MB learning file (link, see SugaA AI 2.50 topic) was posted here in the forum. I had one of my own with 50 MB. When I merged both learning files, the new size was 540 MB. I used it to play about 200 games on PlayChess with my old Intel laptop (only 860 kns). The result wasn't bad, but it was worse than my current test with a current stockfish. With the learning file and SugaR AI I had a slightly negative balance. With the same (own) book and the new Stockfish 220122 (own compile) on my old laptop (860 kns in the starting position) I am after more than 300 games in a row with only one loss and two wins! Thats fantastic.
You don't need a learning file, a current engine is better than an old engine like SugaR AI with a learning file.
