Home > Publishing Schedule > Perpetual Chess Podcast

Perpetual Chess Podcast

Jacob was on the Perpetual Chess Podcast, released this week. At Quality Chess we have finally gotten around to sponsor the podcast and we strongly encourage everyone else to give a little to continue this great free product.
Jacob spoke about a lot of things in the programme, some of which will be opened up for debate here on the blog next week, starting with the “960 circus?!” and the future of chess.
When you are interviewed, there is a temptation to hedge your bets and not give any opinions, but this does not make interesting radio and makes little sense to Jacob. So, he said what he was thinking. Others may disagree, but rather than thinking that is a personal issue, we will have a civilised airing of differences here on the blog.
Btw. The next Podcast will be on Game Changer, the chess publication of the year, if you are to believe the hype. This will certainly not be one to miss!

Categories: Publishing Schedule Tags:
  1. Alex
    February 3rd, 2019 at 22:38 | #1

    As impressive as AlphaZero seems, it is important to remember that its purpose is marketing for their technology. The matches occurred behind closed doors and without impartial supervision. They’re more a demonstration of AZ’s style rather than proof of superiority. That would need to be done publicly and with independent testers. I’m guessing that AZ is way overhyped and nowhere near as strong as it claims compared to traditional chess engines. I will change my mind once they have independent testing. I can even take much weaker engines than SF10 and eventually show some nice wins.

  2. Jacob Aagaard
    February 4th, 2019 at 11:15 | #2

    @Alex
    I entirely agree with this assertion and have said the same in the past. Still I am looking forward to reading Matthew’s book a lot!

  3. Jay
    February 4th, 2019 at 13:16 | #3

    FWIW, I have started the book, and find it extremely interesting. AZ generates very original ideas in a wide range of positions. The way AZ trashes SF in some French and Classical QG positions is particularly impressive. Enjoy!

  4. Matt Fletcher
    February 4th, 2019 at 14:08 | #4

    @Alex
    It’s not quite AlphaZero but LC0 (LeelaChess) has just won its first major tournament (the TCEC Cup), and will be contesting the SuperFinal of the main TCEC event starting shortly against Stockfish 10.

    Regardless of the concerns around the setup and partiality of the AlphaZero-Stockfish 8 match, the LCZero experiment seems to be proving that an AlphaZero-style approach can produce an incredibly strong engine.

    I’ve read most of GameChanger – it’s definitely an interesting book with some fun games.

  5. RYV
    February 4th, 2019 at 15:45 | #5

    OK, computer & AI are playing chess better than humans. But what can we learn from them ?

    When they will be 4000 or 5000 Elo, no more use for humans !?

  6. Bulkington
    February 4th, 2019 at 16:06 | #6

    @RYV
    Isn`t it that the programs have already changed top player`s approach tremendously. Defence has improved, games are much more concrete, basically it is the machines that produce opening theory… and now AZ… exciting times !

  7. RYV
    February 4th, 2019 at 18:52 | #7

    Computers calculate millions of moves while we do a dozen.
    Not thé same game !
    Concrete play Just mean that judgement of à position do not only rely on well know rôles. We have to investigate candidates moves that are not considered valid at first guess.

  8. Gollum
    February 4th, 2019 at 19:22 | #8

    I got quite interested in the QCAcademy and I think I may be able to convince my wife to go together (she liked the hotel), so I hope there’s a third edition in november of 2019.

  9. Jacob Aagaard
    February 5th, 2019 at 14:09 | #9

    @Gollum
    If there is a third edition, it will be spring 2020.

  10. Alex
    February 5th, 2019 at 20:17 | #10

    Oh, I should have asked here what happened to 1. c4? I read Game Changers and they say that AZ prefers 1. Nf3 and 1. d4 almost exclusively. In the original research paper, the self-play graphs also showed it playing 1. c4 quite frequently.

  1. No trackbacks yet.

 Limit your comments to