3

I'm writing an essay related to AI and chess. However I am not a chess player, but a Go player. There are many AI chess engines nowadays that significantly overwhelm the chess ability of human.

Go has many powerful AI engines. The one thing I think about the disadvantage of AI is that so many openings and variations are reduced. For example, there are many openings in Go (i.e. some of the 3-4 Point openings) that were evaluated to be not so good (maybe unconvincing for amateurs) by AI engines, so they become less involved in many Go games.

My question is, did Chess follow the similar trend? Did AI chess engines really have an impact on the chess openings or variations? If so, it would be really interesting to show supporting data that quantify how much reduction has happened. Showing some examples would be nice to!

Improve this question
1
  • "The one thing I think about the disadvantage of AI is that so many openings and variations are reduced." << Regarding go, I've observed the exact opposite of what you describe. Before AI, we had joseki and opening sequences that people would follow blindly. Then AI came along and said "stop playing standard sequences, use strategy and imagination instead" and brought a lot of diversity to the opening.
    – Stef
    Aug 24 at 16:04

2 Answers 2

Reset to default
4

A lot depends on what you mean by 'negative impact'. Certainly some things changed, different middlegame strategies have been re-evaluated, and people focus on openings that the AIs approve of.

The details are technical, e.g. one "new" strategy has been to rush the h-pawn down the board to h6 or h3, cramping the opponent's king position and fixing a weakness on h7 (or h2). This means for example that the Grunfeld with 3. h4 is currently in fashion. It's overall a complex topic, and as a starting point I'd suggest these two recent articles dealing with this very question.

Improve this answer
3
  • Thanks for your detailed explanation. Negative impact on my question means the reduction of the number of variations. I love to see how each boardgames have individual openings and variations. Nowadays, AI re-evaluated existing openings and variations on Go game. Many of them were disappeared since it was not recommended by AI engine and lots of game became standardized according to AI. I'd like to know if there are some similar trends in chess also.
    – Arete
    Jan 18, 2022 at 6:08
  • 1
    @Arete if you mean whether some openings declined after chess AI, the answer is yes. The King's Gambit is an example. I don't know if it's a "trend" though. I doubt the number of viable openings is declining.
    – Allure
    Jan 18, 2022 at 6:30
  • AI has done more for middlegame and endgame, in most AI vs AI game they have them play a few different opening from both sides since given free choice they would end up in a draw every time.
    – Styxsksu
    Aug 23 at 20:19
1

In short, no. I can think of three possible reasons for this.

First, most of the advanced chess-level games are draws. While still important, opening is not the most crucial part of modern chess studies. As many have already mastered opening skills ( there are some basic principles to follow, like developing pieces, pawn structures, etc.), endgames are the real battleground of Grand Masters. Magnus Carlsen, one of the best players of all time ( best of the current generation), is the best ( not one of ) endgame player of all time.

Second, most openings are being studied at least 30 to 40 steps into the game. When AI analyzed it, it was just all zeros ( indicating a draw), which is unhelpful at all.

While indeed some lines are discovered by AIs, those are often so-called "weird" openings that players rarely play in actual games. Moreover, when playing, we can always choose a different line that is more thoroughly studied.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .