He talked about the history of chess, and how there used to be a lot more variants of the game (some even being a 4 player game with dice!), and over time competitive players naturally will want to remove random elements from the game.
But on the other hand, some amount of wild unpredictability is important to attract players - there's a softening of skill gaps.
Pool has little randomness, and therefore it is very difficult to beat a player who is better than you. The best players want to eliminate the possibility of that happening by making longer races, racking their own balls, winner breaks, things like that. Pool is dying for it.
Meanwhile poker has a large amount of short term variance (luck) and it keeps bad players interested for years and years. The worst player in the world can sit and beat the best players in the world at any given moment. Poker is still going as strong as ever-- maybe more strongly than ever at this point. People are coming out of the woodwork this year itching to play.
I think most of the greatest, longest- lived games in the modern era will need a high amount of randomness, because of computers doing analysis. Even more, with the absent of solvers and the like, many poker variants cannot be solved in real time and all-encompassing strategies cannot be developed. More computational power could change that in the future, I guess.
Cepheus http://poker.srv.ualberta.ca/about is an approximately perfectly strategy for Heads Up Limit Hold Em. That is, the two player game of Texas Hold Em poker with fixed bet sizes. There are almost certainly other approximately perfect strategies, which would break even against Cepheus over the long term, but you can't beat it. The best an opponent could hope for is to merely get lucky briefly, for which you might just as well play Roulette.
In principle if you could memorise Cepheus you could play the same strategy, but it's basically a vast number of fractions/ percentages so you're not going to -- and it's important to note that while the strategy is unbeatable it is not the best way to extract money from weaker players. If you want to grind money playing poker you need to focus on taking $1000 from that holidaymaker playing $5/$10 before anybody else realises they're soft, not on trying to break even with a machine.
Heads Up No Limit which was still being played a fair amount not so many years back, is crushed by AI. Pluribus beat the best players in the world, comprehensively. Unlike Cepheus, Pluribus doesn't have an incredibly boring yet precise strategy mapped out that you could copy, it's a result of AI learning. Its bet sizing feels a bit weird to humans, but it ends up taking their money, so, whatever.
There is no AI that can play well, for instance, in a 9-handed game with varying stack sizes, while itself and some competent players are 600BB deep and some other players are 40bb deep.
It takes a specific, narrow ruleset to tailor an AI to be able to play it at such a high level. Or more compute than we currently have in real-time.
Pluribus' matches against pros had each hand reset to 100BB.
Also, notably, the bot can still lose in the short term to terrible players, which was the thrust of my post. In fact, given its bluffing frequency, it might actually do worse against weaker players than it did against pros. Additionally, no human can realistically implement an AI strategy, meaning the AI is not a big detriment to the actual game of poker, as it's currently played.
To be clear about why this might be true for someone who doesn't know much of poker strategy--the AI is (I presume) going to optimize for strategies that can't be beaten by changing one's own distribution of plays (for (simplified) example, if the AI raises a given hand 20% of the time and folds 80%, it's doing so because even if opponent were to call 100%, fold 100%, or something in-between; it wouldn't change the overall expected value of that hand with that distribution of plays). Professional human players, on the other hand, will absolutely cotton onto a player whose distribution of plays is not optimal and can therefore be exploited for profit.
The AI notoriously bluffs a lot, which many bad player types accidentally counter by calling too often.
Good human players will also specifically exploit bad players, as you note, while the AI attempts instead to itself be unexploitable.
Yes, if it's very chip poor, Cepheus could run out of chips before it stops being unlucky and takes your money, but that's unavoidably true anyway in Poker due to it being a game of chance as well as skill.
Cepheus is actually a good Rorschach test for players who say "GTO" a lot. If you actually grasp what game theory optimality is about, you see Cepheus and you're like "OK, I'm done playing Heads Up Limit" (and indeed although it was once somewhat common you won't find any professionals playing this for some years now). But if you're the sort of person for whom it's just another phrase to sling around, like complaining you had a "bad beat" when you were never better than evens, chances are you see Cepheus and you think "I could beat that". Oh dear.
> Also, notably, the bot can still lose in the short term to terrible players, which was the thrust of my post.
This is exactly the useless observation I'm complaining about. Poker is a game of chance so of course anybody might get lucky in the very short term. If what you value is the fact you could win despite being terrible at it, just play roulette and save everybody else the trouble.
> In fact, given its bluffing frequency, it might actually do worse against weaker players than it did against pros.
It isn't about "doing worse" here and that's yet more of the mindset that got you into trouble here. The machine doesn't care about "doing worse" it just wins statistically over time. A grinder needs to make rent and buy groceries, so for them being up $50 after a full day feels like losing, it is losing. But the machine doesn't need money.
Again, if it's just about being able to get lucky and get some sort of temporary "High" from that, I'd commend roulette over poker, or, if you like cards, try blackjack.
No, there aren't any academics showing off bots that play full ring. Of course, there weren't any academics showing off that they could count Blackjack. At least not until after they were run out of Vegas with a fortune of the casino's money... Grinding online full ring with AI is not academically interesting, but it would be profitable. The former would cause you to publish (like Cepheus) the latter not so much.
I'm not sure about the relation between Pluribus and Libratus- I think Pluribus is a newer version of Libratus, essentially?
And this topic frequently comes up in wargaming circles (frequently enough to be annoying :-)). Some feel that nondeterminism is a crutch for the low-skilled while others feel that it is the only reasonable way to handle a low-fidelity model of reality or that it teaches the valuable skill of how to deal with the bag of rotten lemons that the universe periodically hands you.
You can tell he's going a little from memory, but the points are all still there. His arguments that skill and luck are not opposite sides of the same spectrum is quite good.
Charutaji: https://en.wikipedia.org/wiki/Chaturaji
That would really suck.
Medieval chess was a very different game. For the most stark example, the queen did not get her modern move until around 1450.
The game has since been stable for about 500.
And yet something as basic as, "white moves first" was first suggested in 1857.
Left-handed players may prefer Black goes first.
Materially, no, though it probably makes it less complex for humans by slightly simplifying the pattern recognition issues, particularly in the opening.
First player wins more than second in chess (because there's a slight advantage).
White goes first in chess.
Therefore, white wins more than black in chess.
However, that conclusion (what you present) is valueless without the context. If you reversed which color starts the game then we'd end up with "black wins more than white in Chess" as a conclusion and your statement would be false. And if there was no specific color which always started, we'd be left with just the first statement: First player wins more than second in chess.
If you reversed the king and queen placement for both players would the advantage be as great?
There is no asymmetry of moves other than the starting position of the king and queen, so all strategy will simply be mirrored if the king and queen are mirrored.
In this sense, it also includes the conventions, skill-level titles, playing etiquette, etc.
This is awesome :)
I had no idea the game had middle eastern origins for instance. The rooks used to be war elephants hence how they "charge" across the board in straight lines (they were adapted into rooks as the game was Europeanized). Also the reason you never capture the king, which used to be the shah, and resign instead is because killing a rival shah was a big no-no!
So many interesting tidbits in that book. Highly recommend.
https://en.wikipedia.org/wiki/History_of_chess#Origin
Chess originated in India, not the Middle East.
In "Do Dice Play God", another great book, I learned that the earliest dice (probably used initially for diving the future and only later for gambling) had rectangular sides instead of square ones.
I wonder if (a) that was because their creators didn't understand even the very basics of probability, or (b) if the idea of fairness and each number being rolled with equal frequency just wasn't important to them. Not sure.
Rook = tower. Knight = horse. Bishop = alfil, maybe from Arabic for "elephant".
It’s going to be fascinating to see. I can imagine games getting “frozen” with hard coded rules and clear Official Rules too. I expect that to happen to word spellings, for example, with most everything we write having a layer of autocorrect in the loop. Digital games could do the same, when you can’t make house rules without programming your own variant.
Or go the other way, as low-/no-code customization tools and online distribution make it easier to make and share variants than it is to do so at any scale with physical games, subject to the openness (both in design and social factors like IP status) of the base game.
Shogi is really neat in that captured pieces can be returned to the board by the capturer. You don't have different colored pieces, but directional pieces to show which side they belong to.
Xiangqi is my favorite of the two. To me, it feels like a better depiction of war than Chess. The equivalent of Chess's king stays in a small area, there's a river separating the two sides of the board which some pieces can't cross, there's a catapult for interesting ranged attacks. Maybe I've just grown a bit bored of Chess over all the years and Xiangqi is just relatively newer to me, but Xiangqi feels a lot more fun to play, IMO.
Xianqi has sixteen pieces on a side just like chess; Shogi has 20 on a side.
There were a few patches to chess in the mid to late 20th century that disallowed promoting a pawn to an enemy piece (this was useful for forcing a smothered mate), and to prevent vertically castling to a pawn that had been promoted into a rook on the 8th rank (notated as 0-0-0-0-0-0).
Not exactly fundamental changes, but still amusing that they needed to be made after so many centuries.
About Senet, a few rules have been suggested: https://ludii.games/details.php?keyword=Senet
Humans will never be replaced by anything less than humans.