Board Game Arena

I guess Quarto is the "perfect" game for making an AI to win...

Of course with the level of programming AIs for games these days... Quarto would be like a "walk in the park".

I stuided the only(!) match between two of the highest ranked players here, watched the replay, and guess what happend. They played the advanced mode and both did a perfect match. The match was decided by the last piece left. Totally flawless play.
I stuided another match, basic rules this time, between high ranked players, and the match was also flawless, last piece left decided result.

Please note: I don't claim anybody cheats! But of course my findings makes me wonder... are some guys just using this game and this site to test if their AI is perfect... well I am very suspicious now... only me?

between high ranked players

I see your concern and I agree that it's a possibility
but isn't it kinda expected that matches between the highest ranked players, who we assume are the best of the best, end on the very last move or even in a draw and are played flawlessly until the very end?

Well if the level of the best players here are "no mistakes" in Quarto... then the result is decided by who starts. To play a totally "solved" game must be fun? Of course that there is only ONE match between the top players, and this is a perfect match... it's something not "healthy" about this.... sorry but I feel that way... I believe one or both use AI....

Daggerheart wrote:Well if the level of the best players here are "no mistakes" in Quarto... then the result is decided by who starts. To play a totally "solved" game must be fun?

So, I guess nobody has played chess since Deep Blue defeated Kasparov in 1997, as it isn't fun anymore?
Also, I am wondering about your definition of a "perfect" match. Is the match perfect because it went to the last play, or is it perfect because the plays were what an AI would do in that situation?

I've got to agree with Kex on this one: while it is a possibility that AI is involved in any "perfect" game (and that goes for every game that has no randomness and no hidden information involved), it is irresponsible to state that there must be AI involved. Except for AlphaGo, I know of no AI that wasn't modeled after good player's behavior (and AlphaGo was modeled after the big human behavior of "I win when I make this move in this situation.")

Well after what I have "discovered". I don't know how hard it is to play a "perfect" game of Quarto on advanced settings.

So my current "questions" are these:
How hard is it to play a perfect match in Quarto on advanced settings? and the "best" plays perfect each time?
How many of the players here at BGA are capable of playing "every" match perfect (that is, the result is decided by the last remaining piece)?
Why are the top 3 players not playing vs each other? Is it due to they know it will be decided by who starts? And if the answer to this last question is yes.... it's not good for a game to be "solved" like that... = the complexity of the game is not high enough.

Let's face it, the chance of one or more of the top 3 are using AI is over 50%.... Sorry, but for me the "proof" is that they don't play vs each other, they are not "active" seeking to "test their skills" vs the other best players... and the answer to why this... is not pleasant.

To a degree, I sympathise with Daggerheart's suspicions. The suspicion of AI hangs over the best players. Meaning the use of computer programs.
Knowing that your opponent might always or even sometimes resort to AI is irksome.
But I see it as having advantages. Having a few bots around to knock you for six is a useful corrective to vanity. After all, why are you playing the game? To cut a figure, and boast of your prowess? In that case the bots will really annoy you!
But if you play the game for the pleasure of thinking, judging situations, trying out new strategies, developing methods to reduce your careless errors (such efforts might save you from a serious road accident!) then you can happily ignore the bots. If they are playing as a contest between rival programs, what harm are they doing us? It might even be interesting to kibbutz their games, they might give you some interesting ideas.
Personally I confess I like winning, so I limit my games to exclude the masters. It's fun playing against weaker players who are not making too many blunders. And good for them too, to see how I win. And I do lose games! I even lost three in a row, which did hurt, against much lower-rated players. I find the slightest let-up in concentration is fatal against such players. So on a bad day I am going to lose.
Daggerheart: another suspicious thing in Quarto is the presence of masters in the 700s, then a huge gap of over 100 rating points before the next player (who currently is me!)

There are only 16 pieces and 16 spots, and a lot of symmetry. The reason why the game is 'solved'. When you start talking about 'forced moves' then the logic can boil down to just a few likely choices.

It is possible that they are just THAT good - they have figured it out.

Neutreeko and Kalah could be the same too.

Like previously suggested, it should NOT stop you from enjoying the game. The only problem is that of ELO top ranks.

Thanks for pointing out without saying this that Shummi is a cheater.