Hey Guys, Some folks in the 'old school' of 2+2 had some great debates on the subject of how big a sample you need to be confident that your SNG results are not the result of variance.

I used this a while back as the basis of an article at SNG Planet on Confidence intervals in SNG ROI will let you have a link (one of the early articles and in need of a re-write at some point)... the key finding is that the sample required to mathematically even out the chance runs is much higher than you think... for example 1000 SNGs gives you just 66% confidence in the numbers and 500 would be about 50% (not nearly enough for anything even remotely scientific / political or to base a financial decision on right?). 2000 gives you >80% which is considered a baseline for reporting by most SNG Pros over at 2+2 / P5s etc

If anyone has a good ROI at any sample without any Multi-table SNG wins included I am honestly happy for them (do not want to get into any individual debate to be honest).

Do think the debate on confidence intervals is a valid one though... as the Albatross77 put it there is a thin line between your poker fantasy and your poker nightmare!!!

SNG ROI and Confidence Intervals

Cheers, Mark