16 Aug 2004 Bram   » (Master)

Poker

It turns out that my simplified poker was a bit too simplified. Chris Chang pointed out that since all-in bluffs are guaranteed to fail half the time, raising one is strictly superior to going all in. He suggests adding more cards to get rid of that problem. With the way people play these days that game would be just about as interesting as no hold 'em. (See that link for a good explanation of why no limit isn't really 'the cadillac of poker' and pot limit is better.)

Here's an interesting question: if you have a positive expectation bet, how much should you bet each successive round to make your money grow as quickly as possible? If you bet nothing you obviously get nothing, but if you bet everything you wind up with zero and lose out on the ability to make any money in later rounds. It turns out this is known as the Kelly Criterion.

There's computer security snake oil even in poker. This 'audit' just about made my head explode. An audit is when you go over the internals of a system and analyze that it works correctly. Running statistical tests on the output isn't an audit, it's just bullshit.

While trying to grok the rules to Razz I came up with the following game, which I think has most of the strategic elements of Razz but has a lot of the weirdness removed. I say 'I think' because Razz has a whole lot of weirdness and not having played it I have a hard time figuring out what that weirdness causes in practice.

Each player gets two pocket cards to start and community cards are dealt out as in hold 'em. The difference is in how hands are scored. Unlike in other forms of poker, the hands really are seven-carded. The hand with the highest card wins, with a tie-break of the second-highest card, then third, etc. Duplicate cards are completely ignored. For example, if player X has AQ and player Y has K3 and on the board is A5538 then X's full hand is AQ853 and Y's full hand is AK853, so Y wins. If two hands are identical except that one has an extra low card, then the one with more cards wins. For example, if player X has AQ and Y has A2 and on the board is Q9884 then X's hand is AQ984 and Y's hand is AQ9842, and since Y has the extra 2 Y wins. Having two extra cards of course also wins, and having a triplicate of a card causes both duplicates to be thrown out. As in Razz, flushes and straights don't count for anything.

I think there are a few things which make this an interesting game. A hand which contains two lower cards than another one pre-flop has only a low chance, I think about 7%, of winning, so trying to intimidate people by playing every hand will rapidly get you soaked. Also, knowledge gained after each face-up cards are dealt is directly related to the previous knowledge gained, increasing chances for reasoning what cards a player has based on their behavior. The chances of any particular card turning up on the board are reasonably high, so even if you start with a pocket ace there's almost a one in three chance that you'll find yourself playing garbage and having to figure out whether to bluff or fold.

Alarm Clock Display

I have a digital alarm clock with a seven-pin display which I've long noticed 'warbles' in funny ways when I brush my teeth with an electric toothbrush, due to the vibration it imparts on my head. I always attributed to the weird way different pins warble to odd cognitive phenomena, but the other night I noticed that it's always the pins on the top and right edges which warble as one group and the others warble as another, so now I wonder if there's something funny about the timing of it, for example maybe those three pins are offset by 1/120 of a second from the other four as they cycle on and off.

Monitors also warble when you bite into crunchy potato chips. It's highly amusing.

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!