You need to be logged in to your Sky Poker account above to post discussions and comments.
You might need to refresh your page afterwards.
Poker variance isn't normally distributed is it?
Doubleme
Member Posts: 2,141
So this is one issue I have had before when looking at my poker variance it just is not statistically normal and I got to thinking about why? people may not have noticed or their may be much weeping but I have not really played much on sky of late I moved over to spin and goes on another site. I use test of proportions to check myself regularly an noted 4.18 standard deviations in the positive direction (that is testing whether I win more then a third of the time which is how much I would win if I was an exactly average player) I then had a sample where I was 3.24 standard deviations in the negative direction this seems utterly bizarre did everyone just adapt to me straight away? did I have extreme bad luck?
Well the thing was luck definitely played a part I had some very nice luck for my 4.18 and some absolute horrific luck for the negative variance. Now if the random bad luck happened to @TheWaddy or @misterpj there would not be a few posts on the forums instead there would be a bollywood blockbuster with lots of singing and dancing.
However I noted that for my positive run I was very alert very focused noticing what everyone was doing remember my ranges and playing A game and I was playing against people that either did not care about money or had help accessing the internet from someone very irresponsible probably both?
where as the negative run was when I got complacent due to the strong positive SDs over such a small sample size and played when tired when I wasnt playing a lot of attention when the tables were dominated by players at the top of the leaderboard day in day out. Hence in fairness I very likely had the edge in the first bit but did not in the other sample but this highlighted to me that poker variance likely wont be normally distributed over small sample sizes because there are multiple confounding factors not just the cards. the players you get allocated to your table can be of different calibre your state or their state could differ. Also you notice a weakness in your opponent eg they under bluff so you make a lot of laydowns but they notice that and shift gears suddenly over bluffing and you may not have clocked this in time and thus a player you had an edge against maybe even a weak player suddenly for a small amount of time has the edge on you until you readapt. over a large sample size the confounding factors would be equally represented and the data set would approach a normal distribution but then over time yourself or others can improve so really you can make conclusions about the edge you had previously not the edge you have going forward. Maybe those that tell me not to analyse my results constantly over small samples have a point but surely its good to acknowledge when you likely were been outplayed.
Anyway I noted a weakness in my game that has hurt my MTT performance on a number of occasions and hurt me against competent players at spin and goes, specifically in the run bad so now I know an area to study quite a bit and change this.
Well the thing was luck definitely played a part I had some very nice luck for my 4.18 and some absolute horrific luck for the negative variance. Now if the random bad luck happened to @TheWaddy or @misterpj there would not be a few posts on the forums instead there would be a bollywood blockbuster with lots of singing and dancing.
However I noted that for my positive run I was very alert very focused noticing what everyone was doing remember my ranges and playing A game and I was playing against people that either did not care about money or had help accessing the internet from someone very irresponsible probably both?
where as the negative run was when I got complacent due to the strong positive SDs over such a small sample size and played when tired when I wasnt playing a lot of attention when the tables were dominated by players at the top of the leaderboard day in day out. Hence in fairness I very likely had the edge in the first bit but did not in the other sample but this highlighted to me that poker variance likely wont be normally distributed over small sample sizes because there are multiple confounding factors not just the cards. the players you get allocated to your table can be of different calibre your state or their state could differ. Also you notice a weakness in your opponent eg they under bluff so you make a lot of laydowns but they notice that and shift gears suddenly over bluffing and you may not have clocked this in time and thus a player you had an edge against maybe even a weak player suddenly for a small amount of time has the edge on you until you readapt. over a large sample size the confounding factors would be equally represented and the data set would approach a normal distribution but then over time yourself or others can improve so really you can make conclusions about the edge you had previously not the edge you have going forward. Maybe those that tell me not to analyse my results constantly over small samples have a point but surely its good to acknowledge when you likely were been outplayed.
Anyway I noted a weakness in my game that has hurt my MTT performance on a number of occasions and hurt me against competent players at spin and goes, specifically in the run bad so now I know an area to study quite a bit and change this.
Comments
We can then expand this to illustrate variance. Variance in simple terms is the difference between individual results in the short term and the average set of results we find in the long term. The truth is we all run the same over any decent sample/period of time, the only thing we cannot choose is when or where we run good/bad, but we can choose how we deal with each.
When we run bad the main difference is how we respond to this and the better/experienced/skillful players do not lose their heads and keep making good decisions regardless of outcomes. As humans what tends to happen is we prolong a bad run as we play worse, make more mistakes, take on too many marginal spots and tilt uncontrollably in periods.
The reverse can also happen when we think we are on a "heater". When we are running well, we tend to be overconfident, take on too many flips as we are winning more than our share and we may call to many draws expecting to hit and again push the marginal edges much further. We also tend to stop studying as much when we are winning, and in our heads we are now playing the best poker of our career.
Many players tend to think the best time to learn is when we are losing. I would argue almost equally or more so we need to learn when we are winning.
Great topic and nice post, This calculator may be useful for looking at field sizes and what we expect @Doubleme https://www.primedope.com/tournament-variance-calculator/
The bollywood blockbusters come because people relate to what you are talking about, even if they dont want to get involved. 4.5k reads my posts get, although only around 9 of you ever comment.
If you want to talk about the maths, maybe consider that for a start when the claims are that no-one agrees with me.
I mean are you really going to talk variance whilst playing spin and go's? If you play this format, you are considering yourself an incapable player, giving up any edge and just gambling it up for fun.
I dont understand why you claim you know what your talking about, but give up playing slow blind formats on Sky where you have an edge, faith in the randomness of the deck.... to play spin and go's on the worst poker site of the lot! Doesnt ring true......
Without getting too off topic my target in poker is to get to the point where I can have an expected win rate of £200 a day on 8 hours of play a day. so £25 an hour I wont go into great detail on my results so far but whilst I may be profitable I am no way near that at the moment. considering that this sort of win rate has only been hit by the very best MTT players on sky it seems unlikely I can achieve that from sky alone my choice is multi table across multiple sites move to another site with more volume and or play higher stakes (assuming I can beat which I dont know) or try another format.
to be absolutely clear my spin and go record so far is promising but in no way implies I am there yet but the signs are there that with hard work and progression it may be obtainable.
with regards to the site I am currently putting my volume in on been the worst poker site by what measure? I dont think I mentioned which site it was but it is pretty easy to work out. And your right in one regard the standard of spin and go players there is likely the highest standard anywhere, which if I want to find games I an beat is making it harder for myself. However they also have a very generous leader board and rakeback program which effectively reduces the rake you are actually paying to around 1-2% and possibly lower depends on circumstances and can vary day to day week to week etc.
I am not going to preech spin and goes as the way to go, for some people it would be the best way to go for others the worst way. I think the issue with MTTs is once I start am pretty much locked in all day with only 5 minute breaks to go to the loo. spin and goes take between 5-20 minutes between a round depending on if playing ultras/flash or standard speed etc. It means I could go for a walk take a few hours break or etc in the poker day which works better for me.