purechess wrote: ↑Mon Jan 17, 2022 3:22 pm
You use Marvin for analysis? As a sparring partner it's a bit to strong and can't be weaken
No, for sparring I usually use the Maia engines in Lucas chess. Find the one that is not too weak and start from there. Maia is tactically weak but positionally way above a player rated 2000, so chose complicated positions if you want to win, or simple positions when you want to improve your technique. Another solution is too try Zahak 9. It could be even better than Maia.
I play sequences of about ten moves against a strong engine from time to time. Since I always miss some of its moves, it helps fighting against negative emotions, frustrations, and helps refocus. I play only middlegames, mostly technical positions, but still, it is painful most of time. A better solution is probably to play against Maia 1700-1900, at least there is a solid chance of winning a won position.
Engines, in general are useful for two things: one, develop the ability to calculate in sharp positions; two: by exploring a position to learn what are the best plans of play and what the true nature of the position is. While learning calculation is relatively straightforward (exercise, exercise, exercise), learning how to play some positions is a daunting task. It is sometimes paradoxically difficult -- e.g. I have been playing for years the French Defense as black, but I still have problems playing it as white. I pray for my opponent to choose 3... c5 in the Tarrasch, and avoid typical French structures that I adore playing as black. Recently I switched from the exchange var. in the Caro-Kann to the advanced variation, and I was completely lost until I read an old Karpov's book about it. I think it is extremely important to discover what's wrong in a position -- and NN engines are very good at refuting wrong moves and bad plans.
Finally, I found that 30-40 mn + 30s is enough for me to play some quality chess and calculate extensively in the few critical positions when deep calculation is unavoidable.
).
