CardSharp

If you can’t evade a tax, pay a little too much to confuse their computers.

-Friday, Robert Heinlein

Here’s a funny little story.  A few years ago I was playing a 1/2 NL home game.  It was my first time there, and we were playing somewhat deep for a small game – I think I had 150BB and several people had me covered.  Anyways, because of the stack depths and general table atmosphere I was using a slightly oversized preflop raise – I was thinking something like 4.5BB+1BB/limper would be about right (3 to 3.5BB +1BB/limper is more standard).  Anyways, just for shits and giggles, I decided to do 4+1/limp half the time and 5+1/limp half the time. The choice of which I did was more or less random, but not really rigorously randomized in any way.

Because there was a fair amount of limping going on, the first 5 or 6 times I entered the pot I did it for a different amount of money each time.  Anyways, at the end of the night as I was racking up one of the other players asked me what the pattern was to my preflop bets and told me he thought he had it figured out, but then it turned out he was wrong.  I more or less dodged the question and left, not wanting to talk strategy with a likely future opponent.

As I was driving home and thinking over the major hands of the night, I realized something – the player who was trying to “figure out” my preflop play made a couple of weird preflop plays against me later in the night.  Now I don’t know for sure, but I’d happily bet at even odds that the reason he did that was that he thought he knew something about my hole cards based on the raise size.  Of course, in reality he didn’t since the raise sizes weren’t based on my hole cards at all.

The point of the story is this:

Whenever you have a decision that just plain doesn’t matter much, randomly pick between your options.  If no one notices, you’ve lost nothing.  But if someone concludes a pattern exists where there is none, they may well make a mistake and then you profit.  Throwing a spanner in the works is a freeroll.