Category: Games

The amazing Gukesh (17 years old!)  is half a point ahead with one round remaining today.  Three players — Nakamura, Caruana, and Nepo — trail him by half a point.  Naka is playing Gukesh, and of course Naka will try to win.  But what about Caruana vs. Nepo?  Yes, each must try to win (no real reward for second place), but which openings should they aim for?  You might think they both will steer the game in the direction of freewheeling, risky openings.  Game theory, however, points one’s attention in a different direction.  What if one of the two players opts for something truly drawish, say like the Petroff or (these days) the Berlin, or the Slav exchange variation?  Then the other player really needs to try to win, and to take some crazy chances in what is otherwise a quite even position.  Why not precommit to “drawish,” knowing the other player will have to go to extreme lengths to rebel against that?

Of course game theory probably is wrong in this case, but is this such a crazy notion?  I’ll guess we’ll find out at about 2:45 EST today.

For the first time in 75 years, Mattel is making a major change to the iconic board game Scrabble — and touting a “No More Scoring” gameplay option.

The new launch is a double-sided version of the famous board game — one side with the original game for those who want to stick to the long-time traditional version, and one side with a “less competitive” version to appeal to Gen Z gamers.

The flip side of the classic game, called Scrabble Together, will include helper cards, use a simpler scoring system, be quicker to play and allow people to play in teams.

“The makers of Scrabble found that younger people, Gen Z people, don’t quite like the competitive nature of Scrabble,” Gyles Brandreth, who co-hosts the language podcast Something Rhymes With Purple, told BBC Radio 4 Today. “They want a game where you can simply enjoy language, words, being together and having fun creating words.”

And:

So what’s new exactly? In addition to a dual-sided board, there are helper cards, which provide assistance, prompts and clues and can be selected to match the player’s challenge level of their choice.

In the new version, scoring is a thing of the past — now all one has to do is finish a goal and collect the goal card.

Goal cards include challenges such as: “Play a horizontal word,” “play a three-letter word” or “play a word that touches the edge of the board.”

Here is the full story.  I have read that 33 out of the world’s top 100 Scrabble players are Nigerian — I wonder what they think of this?

As you may know, a major tournament is going on right now, based on a variant of Fischer Random rules, sometimes misleadingly called “Freestyle.”  Subject to some constraints, the pieces are placed into the starting position randomly, so in Fischer Random chess opening preparation is useless.  You have to start thinking from move one.  This is a big advantage in a game where often the entire contest is absorbed into 20-30 moves of advance opening preparation, with little or no real sporting element appearing over the board.

Yet I don’t like Fischer Random, for a few hard to fix reasons:

1. Most of the time, at least prior to the endgame, I don’t understand what is going on.  Even with computer assistance.  I could put in five to ten minutes to study the position, and get a sense of the constraints, but as a spectator I don’t want to do that.  As a relatively high opportunity cost person, I am not going to do that.

1b. Classical chess sometimes generates positions where one does not really understand what is going on.  Then it is thrilling, precisely because it is occasional.  A perpetual “fog of war,” as we receive in Fischer Random, just isn’t that thrilling.  In the opening, for instance, I don’t even know if one player is attempting “a risky strategy.”  I am not sure the player knows either.  And I don’t feel that watching more Fischer Random would change that, as there are hundreds of different possible opening positions, mostly with different properties.

2. The younger players have a notable advantage, because they are better at calculating concrete variations and rely less on intuition.  (We already see this in the current results.)  Experience is simply worth much less in this very novel format.  For any one tournament, that is an interesting intrigue.  But over time it is a bore, as if only rookies and sophomores could win NBA titles.  In fact what spectators enjoy watching is Steph Curry going up against Lebron James, or the analogs in chess.  We want to see Magnus meet Fabiano again, not watch two eighteen-year-olds slug it out.  Sorry, Pragga!  You’ll have your day in the sun.

3. Fischer Random cuts off chess from the rest of its history.   That is otherwise a big advantage of chess over many other games and contests.  I like seeing that a player’s move is connected to say an idea from Tal in the early 1960s, or whatever.  I like “Oh, the Giuoco Piano is making a comeback at top levels,” or “today’s players are more willing to sacrifice the exchange than in the 1970s,” and so on.

4. I get frustrated seeing all those Kings sitting on F1, not able to castle in the traditional sense.  There are rules for castling in Fischer Random, but it feels more like pressing the “hyperspace” button in the old Space Invaders video game than anything else.  Who wants to see a Knight on C1 for twenty-five moves?  Not I.

5. I agree that current opening prep is insanely out of control.  I am fine with the remedy of 25-minutes per player Rapid games, or anything in that range, with increment of course.  Those contests are consistently exciting and they are not forced draws (you can play something weird against the Petroff, or to begin with) nor are they dominated by prep.

6. If you don’t want to watch Rapid, I would rather randomize the first few opening moves than the placement of the pieces.  If you don’t control the first three (seven? ten?) first moves, once again opening prep becomes much tougher.  So what if some games start with 1. b4 b6?  The resulting position is still playable for both sides and furthermore it still makes intuitive sense to chess spectators.  Of course the computers would restrict this randomization to sequences that still are playable for both sides.  The very exact nature of current chess opening prep in fact implies you need only a very small change in the rules to disrupt it, not the kind of huge change represented by Fischer Random.

That all said, I am all for experimentation, it’s just that some of them should be strangled in the crib.

Scaling seems to be working here:

The recent breakthrough successes in machine learning are mainly attributed to scale: namely large-scale attention-based architectures and datasets of unprecedented scale. This paper investigates the impact of training at scale for chess. Unlike traditional chess engines that rely on complex heuristics, explicit search, or a combination of both, we train a 270M parameter transformer model with supervised learning on a dataset of 10 million chess games. We annotate each board in the dataset with action-values provided by the powerful Stockfish 16 engine, leading to roughly 15 billion data points. Our largest model reaches a Lichess blitz Elo of 2895 against humans, and successfully solves a series of challenging chess puzzles, without any domain-specific tweaks or explicit search algorithms. We also show that our model outperforms AlphaZero’s policy and value networks (without MCTS) and GPT-3.5-turbo-instruct. A systematic investigation of model and dataset size shows that strong chess performance only arises at sufficient scale. To validate our results, we perform an extensive series of ablations of design choices and hyperparameters.

Here is the full paper by Anian Ruoss, et.al. I guess AI is likely to get better at other things too  — I’ve become numb to the miracles, frankly.