Not a fan. Bring back the depth. "3560 rating" is meaningless.
The new method should complete faster and be as good or better than depth was.
Not a fan. Bring back the depth. "3560 rating" is meaningless.
The new method should complete faster and be as good or better than depth was.
How could it possibly be better than depth? The engine is running for the same amount of time, no? I’d just suggest a clearer explanation than “3560” rating. I also liked having the depth analysis where I could set it to do a much longer and accurate evaluation but that’s gone.
I appreciated being able to run at depth 30.
I agree with TBM, the "rating" doesn't make sense. The difference between the "Standard" 3430 and the "Maximum" 3560 means nothing - the numbers are imaginary, and they might as well be the same.
However, knowing that the engine is evaluating to a 30 move depth gave me confidence that the game review is not overlooking deeper strategic lines. If chess.com had to remove it for capacity issues or whatever, they might as well just say so, but why pretend the new way is better?
How could it possibly be better than depth? The engine is running for the same amount of time, no? I’d just suggest a clearer explanation than “3560” rating. I also liked having the depth analysis where I could set it to do a much longer and accurate evaluation but that’s gone.
No with depth, it's running to specific number of ply on ever single move, regardless if it finds the best move early on or not. With nodes, some parts of the game will get a faster result and other parts may take a little longer.
Based on a comment by staff, my guess is that they benchmarked the differences and found that method to be faster and generate very similar or better results.
1 second? So the engine spends roughly 0.01-0.02 second on every move? That would probably be something random, not a 3270 rated analysis for sure.
1 second? So the engine spends roughly 0.01-0.02 second on every move? That would probably be something random, not a 3270 rated analysis for sure.
Depends on the hardware and if they use parallel processing on the positions. That loses some cache hits but fast and large enough hardware can make quick work of analysis. 1 second does seem a bit fast but I've never ran full have analysis on a big server system before to compare it to.
If you need help, please contact our Help and Support team.
Not a fan. Bring back the depth. "3560 rating" is meaningless.