Hi, I've been wondering how accurate the lichess chess engine is, or more specifically if I should even use it for analysis of my games/studies. I was looking at a game in Alexey Suetin's 3 Steps to Chess Mastery book, with the game in question heading to an endgame, but for some reason when Suetin makes a point that one side was slightly better or something along those lines, the engine thought it was pure equality. I'm not sure if its a matter of device or depth, an incorrect analysis by Suetin, etc. Can someone please clear this up? Thanks.
Well, if you're reading an instructional book, I don't see much reason in corroborating that with an engine.
The game in question was by Karpov vs Miklayev, forgot the rest of the details.
MrPushwood now that you mention it, you're probably right. Although in some cases the games are played by GMs and because I'm not a very tactical guy, I like to see the crazy tactics the engines give in some of the positions or other ideas.
One of my own personal favorites ! Three Steps To Chess Mastery by A Suetin . Of course the book wasn't written yesterday but when I studied it it was preety recent . The Ideas Suetin expresses can be used in different settings which makes the book worthwhile' . Using Lichess Analysis with while looking at Chess Books is very useful & looking at games also as it spots many errors & has a rating Stockfish does around 3200 , It may be difficult to grasp everything but it's great to have such an Analysis partner , @YourBroThePro A good book to read also would be Modern Chess Strategy by L Pachman probably before tackling this book & possibly check out some new books on Chess Strategy by J Aagard
This is basically because chess engines evaluate things very differently to humans. Chess engines will look at the very best line they can see and if that line comes out with a drawn endgame, it will quite rightly conclude that the game is tied.
However, a human cannot calculate close to that of a chess engine so humans take shortcuts in their analysis and look for themes and general heuristics to allow them to asses the position overall. If a human can see a number of lines a few moves in that seem to always favor one side or the other, a human will, also quite rightly assess that this side is better because humans will assume that no one follows correct play always. If side A) has to find 10 best moves in a row to draw and otherwise side B is ahead, then side B is definitely better from a human point of view as unless you are incredibly lucky or maybe Magnus Carlsen or some other super GM it is unlikely that side A will find it. The human will assess that its better to be the guy on side B but the computer seeing the absolute best play from both sides will think it is a draw.
A common real world example is a situation where you can get two rooks for a queen. Now the computer will often (not always) assess that this is good (in theory 2 rooks are worth 10 and the queen is only worth 9) if it can see a line where the slightly higher amount of material can get over the fact that the queen is so powerful. The reality is that in almost all those games there are very few lines where the guy without the queen is better and many where they are worse, but in the best line he might well be better so the computer thinks that sacrificing the queen for the two rooks is the right thing to do (and it is) because it can already see the best line having looked 20 moves in. However, the vast majority of humans will not take that swap knowing that they will have to demonstrate almost perfect play to make that swap worth it.
When one is rated 1300 most engines are probably quite accurate.
ENgines have no fear and are only ice-cold calculation ... people are different than that ...
This topic has been archived and can no longer be replied to.