Analysing SNG results - how?

[keilan303]

Played a few of the 55's and 100's recently (outwith my bankroll) and faired pretty well. I have someone willing to help out with the stakeage if I can prove my results could be consistent and not merely one offs. Short of asking him to read Moshman's book etc to give him a better understanding of SNG's, is there an easier way to analyse such results? I have SNG Wiz but can't really figure it out. I presume if I run a few results into it and can show that there were more +EV moves than -EV moves this would be the best way. (Does anyone want a look?)

[Moskvich]

If your staker was himself a good SNGer then he could watch you play for a bit and make his own mind up about whether you were any good or not.

Otherwise there's really no way of proving, to him or to yourself, that you can beat the games consistently - other than playing shedloads of them in a fairly short space of time. Like a couple of thousand over a couple of months or something. The variance in SNGs is pretty frightening, so you need a really big sample size to allow things to even out. (Much cleverer people than I have done the maths to show how spectacular the variance can be - ie you can be a 10% winner and still be expected to have a 50 buy-in downswing at some point, etc). Playing a few and faring ok demonstrates next to nothing, I'm afraid.

As for SNG Wiz - you probably should work out how to use it, in order to improve your own game. But you can't just stick hand histories into it and wait for it to tell you whether you're great or not. You have to make your own judgments about other people's styles - for example, if you want to know whether to shove Q9 on the button you have to tell it what hand ranges the blinds are going to call with. It can't know that. So as people on forums are fond of saying, you put rubbish information into it and you get rubbish information out of it.

My advice would be to move up the stakes more gradually, and play a decent number of games at each level to establish some sort of credible positive ROI, before jumping into the higher-stakes games - then you can give your staker your stats at the 20s and 30s and 50s, or whatever, and you can make an informed decision about whether to give it a go at the 100s.

[Longy]

Post hands you are struggling with on here and will become pretty obvious quite quickly roughly where your game is at.

As Mosk says learn how to use sngwiz, it is the single quickest way to improve your short stacked play which is vital in sngs. Also to echo Mosk big samples are needed to prove your ability completely, for example I am sure i could dig out 2k break even stretches and other 2k samples where I ran at twice my expected roi.

Moshman is quite restrictive fwiw and learning how to use sngwiz properly is far more useful. Also reading STTF on 2p2 regularly is far better if you want to read strategy.

[thetank]

Would agree with the above, except for the part about playing intermediatary stakes for long periods of time when you're getting staked.

That's the whole point of getting staked, to skip all that stuff.

[thetank]

Don't use Wiz to review tournaments ie to tell you where you went wrong. It's very shit at that because the hand ranges it typically assigns to people by default are way off. Not as bad as it was, but still not to be relied on.

Far better thing to do is just to look at hands that interest you. Maybe flag them while you are playing, if unsure of a desicion then look it up later sort of thing.
If you're not sure where to start looking, then likelyhood is that anyplace is a good place. (don't take that the wrong way)

When you do look at them at Wiz, be sure to play around and ask yourself what if questions. What if I have 5 big blinds instead of 8, What if the other player has x amount, what if your hand is such and such instead, what if the big blind is calling a little tighter, a little looser etc etc.