Perfect chess opening: If AI solves chess | Hikaru Nakamura and Lex Fridman

Lex Fridman Podcast full episode:
Please support this podcast by checking out our sponsors:
– Mizzen+Main: and use code LEX to get $35 off
– InsideTracker: to get 20% off
– NetSuite: to get free product tour
– SimpliSafe:

GUEST BIO:
Hikaru Nakamura is a chess super grandmaster and is currently the #1 ranked blitz chess player in the world. He is also one of the top chess streamers on Twitch and YouTube.

PODCAST INFO:
Podcast website:
Apple Podcasts:
Spotify:
RSS:
Full episodes playlist:
Clips playlist:

SOCIAL:
– Twitter:
– LinkedIn:
– Facebook:
– Instagram:
– Medium:
– Reddit:
– Support on Patreon:

91 Comments

  1. Just a guess, but White can't gain any effective assymetry because no initial move forces a move on Black, i.e. there is no single purely advantageous move to open with.

  2. It is theorized that there are more possible chess moves than atoms in the universe.

    Atoms: 10^120

    Chess moves: ~ 10^123

  3. There’s a way to win chess and it’s to account for how every piece would move given the spaces that it’s allowed, I’m no mathematician but it takes crazy math but it’s possible

  4. Really love this. Great insight. Truly wonder if Chess will ever be hard-solved.

  5. This guy talks i keep thinking he is asking a question

  6. It doesn't really make sense to talk about something if the levels of expertise are so vastly different. Lex doesn't even understand what Nakamura talks about. It is like watching a 6 year old kid trying to talk to a professor about fusion.

  7. U CAN SOLVE SCHESS BUT U CANT SOLVE GUITAR COMPOSTITION

  8. I think getting enough advantage for a win with white should be impossible with perfect play. I agree with Hikaru.

  9. You can tell he’s disappointed but the truth is this is likely the reason chess has existed for so long and continues to fascinate.

  10. To me, it seems because there are literally quintillions of possible board states, the advantage that moving first gives you is going to be reduced to 0 as the limit of possible states approaches infinity. So even if the advantage were significant in any one state, with quintillions of possible states, I think it is not close enough to a value that is significantly different from 0.

    Games that have a set opening can lead to wins even for similar computers playing, but I think this is because the possible board states are drastically reduced so that significant advantage that happens to exist in a single state, where the openings end, can be maintained.

  11. If you are keeping asymmetry then you are handicapping 1 side and you aren't really solving anything. 1 side is always going to win, and if you pui 2 AI against each other who are equally smart the will draw.

  12. My prediction is that AI will end up showing that there is a mutual Zugzang at move 1. The player who moves a pawn first, will lose. That's why it's a draw after Nf3 Nf6 Ng1 Ng8 repetition.

  13. That is some high conjecture. How does symmetrical moves equate to a draw, not to mention symmetrical moves can only occur for so many moves before white wins.

    My conjecture would be white wins unless AI develops a perfect defense for black that yields in a draw. Mostly because the turn advantage is incredibly powerful and already breaks the symmetry Hikaru speaks about.

  14. What is this lunatic talking about? e4e5 is never a draw! Black always wins! You're not allowed to move twice in a row, so white is disqualified!

  15. Wouldn’t every game end with the first person to play winning rather than a draw?

  16. Chess could only be "solved" if there was a proven set of moves that white can play that always wins. This is impossible as the advantage of moving first is negated by the fact black has more information before every move. Therefore black can always respond in a way to white that maintains parity.

  17. for our (humans) reference, AI have already solved chess to beat any human 100-0, not even Magnus can win a single game from Stockfish .

  18. well i think it could also be like some rock paper scissor stuff, like each opening has a counter, but still many would prbbly lead to draws then

  19. Why does Hikaru sound half 14yo boy and half 45yo housewife

  20. In a perfect zero sum game doesn't the first move give an advantage?

  21. there might end up being like one or 2 openings that only white can win or something, with everything else being a draw

  22. wouldnt you think that White always wins, as he is basically one move ahead?

  23. I just say, 'Yes'. So, people may think I know what he's talking about.

  24. how cant it be solved? its fucking math… any math can be done.. I dont play but math is math..

  25. Plays E4
    Stockfish: this is just simply a draw in 80 moves

  26. I love chess…don't get me wrong, but I've outgrown it. I miss the magic of the game as a young boy. My father didn't teach me, I learned by watching him play against a computer chess board for countless hours and days, over several years. I played him when I was older and eventually I got to the point could beat him on occasion. I'm comforted in knowing that boys and girls are discovering the magic of chess right now. And, hopefully, they too will outgrow it.

  27. Can just make it 9×9 squares and add a new piece, or even let it 8×8 and replace one of the bishops or knights with a new piece, or just play chess 960.

  28. Even if it gets solved. There's still the requierment to remember all the sequences and variations to fully take advantage of it. Plus we're already confronted with many "engine-moves" that are very unnatural and difficult to understand why they are better. If anything many more moves like that will get revealed with quantum computing that need to be understood first which is by no means trivial. For 99.9999% of chess players the game will personaly always remain unsolved and there will be far more things they don't understand about than things they know about it and I would suppose that the number of people having "solved" the game to a greater level will remain in the single digits.

  29. I think Chess keeps its popularity (opposed to chess960 or other variants) because of the human history behind, and also literature and humanities connections. Although, it seems the last chapter on chess history is computers not a human feat – curiously this does not happens in the arts which keeps surpassing machines. So probably chess is not an art it is just a kind of limited but entertaining game. Of course one can say computers can't be compared to people, even in chess, because they are not on the same "living stage", and they haven´t a biological biography, and so on… To put it short: some GMs say they don't consider computers an opponent but an aid in their training. Whatever the case a mathematician can't be beaten by a cell phone in a decisive breakthrough in their field, neither a painter, but a chess player can.

  30. imagine chess is solved. people will memorize the perfect white moves, now black has an incentive not to play the best moves because they are the ones white memorized how to play against

  31. We will never solve chess, because there aren't enough particles in the observable universe. Sounds like a joke, but the full game tree (which we need to brute force the game) of a chess game is 10^120 nodes large, working to the Shannon number (a lower bound estimate of the moves of an average chess game, 30^80). The entire information of the game tree must be stored on a storage medium, the smallest medium that exists is a particle and there are only 10^80 of them in the observation universe. 10^80 is significantly smaller than 10^120, although we are not even calculating with the maximum length of a chess game

  32. also, even with the slight advantage white has at the start, there are a lot of draws for example a pawn at the end on left/right side of table or just one horse/bishop, where it is draw bc insufficient material
    if both computers play flawlessly, that advantage that white has will probably be moved to the end of the match, with a draw in white's favor

  33. Chess can never be solved because the number of possible legal positions in the middlegame is astronomical. There is also more than one sound response to any move or position. Your top 2 or 3 choices for moves, there is not an immense difference where it's not forced, there are just degrees of good and better. Past these top choices there is inaccurate then blunders. There is no "universal" move for opening because it depends on what you do after and what your opponent does. What you have are mainlines that are known, some variations for each and the rest are novelties that are not super common.

  34. Not happening in our lifetime. And even if it is, I doubt humans have the capability to reach every possible draw in every possible variation. We're not computers. The way it is now, is the way it will be for some time. Maybe for all time.

  35. I think if both players play perfectly, it will always be a draw , so the game it s to take advantage of your opponent's mistakes ‼️

  36. He didn't really answer what Lex was asking which was if White has an advantage in creating asymmetry because you start 1st. Or does he mean that White cannot do that with perfect play from black?

  37. If chess is "solved" then I think the highest levels will turn into stalemates or white wins every time.

  38. There are 10^120 chess games it will take more than age of universe to guess all games

  39. I think Hikaru has a really good point. However as far as I know we have only 'solved' the game with up to 7 pieces on the board. If our current technology isn't able to play 100% perfectly with even just 8 pieces on the board, how are we supposed to be certain about the rest of the game when its exponentially more complex the more pieces you add. Of course Hikaru knows way more about chess than I do I just feel that at the moment we can't really say for sure. Would be cool to see something unexpected ig.

  40. But it isn't perfectly symmetric because white moves first.

Leave a Reply

Your email address will not be published. Required fields are marked *