For allins as I know all the cards, the pot size and the mean is just the expectation. I know the actual outcome so the variance of this hand would be (mean - outcome)^2

If I did this for many allins and summed all the variances together to give the group variance would it be reasonable to expect the actual total to be coming from a Normal type distribution with a standard deviation of Sqrt( sum of these variances) and mean = all expectations summed ?

Does it matter that the sizes of the pots may all be different?

Can I treat allin results from many tournaments each as independent with variance summable (even though the pot size would tend to increase if the previous allin was won), or do I have to do some co-variance calculations for groups of results.

It seems to me that if I repeated/replayed these tournament exact hand match-ups over and over I would get a nice gaussian bell curve if I plotted the sum of the actual outcomes and also it would be reasonable to expect 68% of the time to be within +/- 1 sd of the this mean, sd calculated as above. But am I mistaken?

Thinking about it a little more one possible 'improvement' for analysis purposes could be to calculate the expected variance rather than use the variance outcome of the exact hand outcome.
I could calculate the weighted variance for wins, losses, and ties for each hand.

Using the variance to estimate the probability of Tournament AllinEv difffs.

I now can go through my hand histories and calculate the variance of each pre-river allin hand (no future actions possible) and I was wondering if using this I can get an estimate of how likely a difference is to occur. Here is a graph of my last 4973 180 seat tournaments, filtered for allin situation,s with 5 or more players at the table:


Adding up the variance for each individual hand gives me a variance of 413726014220 and from this a total sd of 26860093, mean 6657266.

For the next 4973 tournaments I play is it reasonable to expect my mean to be about the same with roughly a 70% chance of the actual being between this mean +/- 1sd, ie. +/-27,000,000?

In these hands the ones from the later stages of tournaments will be dominating the other earlier ones variance/sd wise but I don't think this matters. There will still be well over 100 final table big pot hands in this group, enough I think to get toward a Normal distribution.

screwed post - will redo it in a bit

Using the variance to estimate the probability of Tournament AllinEv difffs.

I now can go through my hand histories and calculate the variance of each pre-river allin hand (no future actions possible) and I was wondering if using this I can get an estimate of how likely a difference is to occur. Here is a graph of my last 4973 180 seat tournaments, filtered for allin situation,s with 5 or more players at the table:


Adding up the variance for each individual hand gives me a variance of 413726014220 and from this a total sd of 643215, mean 6657266.

For the next 4973 tournaments I play is it reasonable to expect my mean to be about the same with roughly a 70% chance of the actual being between this mean +/- 1sd, ie. +/-645,000?

In these hands the ones from the later stages of tournaments will be dominating the other earlier ones variance/sd wise but I don't think this matters. There will still be well over 100 final table big pot hands included in this group, enough I think to get toward a Normal distribution.

Sorry people I messed this up. The sd in the above post should be 643215 not the 26860093 shown. I also meant to just update this thread but accidentally posted the edit as a another thread.
I think it is better to reply to the other thread here:
http://forumserver.twoplustwo.com/25...ifffs-1553509/

Lol I'm not having a good posting day 'cos I even managed to miss the edit time on that one too, sry again.