Bram Cohen (bramcohen) wrote,
Bram Cohen

A new card game

I came up with a new card game the other day. It's been play-tested a bunch, and is a lot of fun.

The Rules:

There are two players. A standard poker deck is shuffled, and sixteen cards are dealt out face up, eight to each player. Both players get to stare at the cards for a while.

Each player then gathers up their eight cards and sets aside four of the eight cards for the other player. They then exchange the cards they've set aside for each other. Each player then forms the best poker hand they can with any five of the eight cards in their hand, and the one with the better hand wins.

That's the entirety of the rules. What makes play interesting is that if you can guess what cards your opponent will keep there's almost always a set of cards you can set aside which will make you win.

  • Moving

    I've moved my blog over to See you all there!

  • Practical Cryptography Corrected

    The book 'Practical Cryptography' is perfectly good for giving an overview of basic concepts in cryptography, but its immediate practical advice to…

  • Git Can't Be Made Consistent

    This post complains about Git lacking eventual consistency. I have a little secret for you: Git can't be made to have eventual consistency.…

  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded 

Anaconda variant. One of my favorite card games!

Was there specific reasoning for the numbers you chose? Or does it just seem to play well that way?
I think I was inspired by chinese poker (I woke up thinking about the idea, so the beginning of my thought process was kind of lost) but yeah, it somewhat resembles Anaconda for two players.

It just seems to play well with those numbers. Keeping only four makes it possible to keep most of a straight or flush, which makes it possible to reason very specifically about what the opponent might hand over. Passing more cards results in an excess of strong hands, which makes the game less fun to play. In principle one could write a script to check what the probabilities are of one side having an unbeatable set of cards to hold with various numbers of cards held and passed, but I haven't done that, and the question of 'how interesting is it' isn't really well defined.
>> What makes play interesting is that if you can guess what cards your opponent will keep there's almost always a set of cards you can set aside which will make you win.

Doesn't that lead to a Vizzini Delimma? (ie "I can't drink the wine in front of me, but you would know that I know that, so I can't drink the wine in front of you?")
The question, then, is how to poison poker cards with iocane powder....
Yes, that's central to the strategy of the game. I don't know how practical it would be to write a program to actually calculate the nash equilibrium of 'ideal' play.
Well, exactly. But then it's not really different from any particular poker variant in that the skill comes in playing the opponent rather than your cards, and that, ultimately, you're still only playing the percentages (ie there's a lot of luck to guessing what your opponent is going to do), which means that you still need to play a large number of hands to let skill rise in importance.

As a side note, I've often wondered why there are so many variants of poker. The mathematics of, say, Texas Hold'em are essentially the same as Omaha are essentially the same as Stud Poker etc (certainly, the odds of particular hands are different, but the process of calculating those odds is the same). The techniques of playing the people around the table are also the same (as far as I can tell). Yet, different people seem to excel at different games. To be sure, many top-level pros are very good at several varieties. But, everybody seems to have a specialty.
The big difference between poker variants has to do with how much available information there is to do reads with, and how much variance there is. In hold'em reads are very possible and variance is extremely high. In Omaha reads are very possible but variance is low. Obviously the game-specific math is different enough in some cases for the crossover to be low as well, for example hold'em and lowball.
This might make a good Facebook game.
that's remarkable post....