Too much of a good thing

[HeavilyArmed]

I play $11 SNGs. I win at a reasonable rate and since I'm generally untrustful of online poker I won't move up because I don't trust anyone with much of my money.

6 months ago I tried a new site. My results were good at the start and then I flat lined after a week or two, my exact memory is not 100% here. I abandoned them.

New rules, new needs, I go back to the above mentioned site. This time I keep strict and accurate records of every all-in situation with cards to come. I run each situation through twodimes.com to generate my true EV and this I compare to my actual results. So I'm looking at the first ~230 hands where I'm all-in since my return and my EV is to win 110 and my actual is to win 133.

SOunds like no BFD, you say. Well, maybe not. My ROI for this period is 74%. Yes, it's a short run but...74%, wow.

I have some statistical education but I can not tease a SD out of this data set, too convoluted, I'm too uneducated. I can simulate it to get some ideas. This is a 30 minute coding chore, I do it. The result is that my outlying result can be expected once every 500 runs. That's, er, maybe 3-4 SDs out of the mean. That will always get my attention.

Sending good cards to new and returning players strikes me as perhaps the most likely cheat a dishonest site might try. It's nearly invisible, but only nearly. Obviously you don't deal them AA, you bugger the turn or river to lend a helping hand.

For you kneejerk skeptics, design an experiment to catch this cheat and you'll see just how small the chance of detection will be.

And for the site's name, well, that's unlikely to make a difference for my fellow Americans in a couple weeks.