Math People- Explain things to me

I understand variance in the short term- over 1k, 2k, 5k, 10k hands maybe. At the same time, how long is it possible (as in reasonable likely) to run significantly worse than normal. For example, over my poker career (which at the time was like 100k hands) I had a win sd % of about 51. For the last 45k hands (april) it has been 48%, and needless to say, my results have suffered. There have been no grand, sweeping changes in my game.

Essentially, one of the reasons we play a lot of hands is so we dont have losing months. The thing is, I feel like i am crushing the games I play in, but judging from winrate I feel like I should be making adjustments. Does anyone have a link to a thread, or can do math to give what maybe a std deviation in win SD % would be over 50k hands, 100k hands etc and how this correlates to winrate. ...?

sry, heres some hard numbers:
premise 1: I feel like I am playing better than I ever have, but then there is a big old difference between my results for april and for the previous 2 months I have been playing poker.

Feb, March: 63k hands 22.43/17.81 3.32 AF, 24.41 WTSD, 51.93 W$SD, 32.74 ATT STL BLNDS

April- 45k hands 23.79/18.66, 3.2 AF, 23.93 WTSD, 48.55 W$SD, 37.01 ATT STL BLNDS

Ill post more stats later maybe...

but honestly, maybe its jsut because i havent been playing for too long, but when things like APRIL happen I begin to wonder if either the world is ending or i am missing something fundamental about poker. If someone can show some math about long term variance it would be greatly appreciated. Thanks.

I've run significantly bad for the past 50k hands or so. But my WTSD is 26% and my W@SD is 56%. Problem is though the pots that count, the huge ones, I am getting beat badly. 2-3 outters or set over sets all that bad stuff.

Basically my point is you can run bad for a good chunk of hands. And I really don't think your W@SD is necessarily a good indicator for whether you are or not.

What would be cool is if there were a program out there that could go through your PT database and analyse if your hands are holding up as often as they should once the money's in. That way you could know if you're running "unlucky"

Variance can go way longer than you think.
I forget the 2p2 thread that showed standard deviations of winrates but let me try to recount it.
100k = 4ptBB Standard deviation.
This means that over 100k hands 64% of players are expected to swing 4ptBB/100 along their winrate. So pretty much if you're running 4ptBB/100 over 100k hands you could conciveably have a true winrate of 8ptbb/100 or a 0 winrate.
Now that's one hell of a swing. Not only that but 36% of players are going to even have a swingier deviation.

It's not till 500k hands where there is very little deviation among your stats. AND even then, there are going to be a few players who just run so horrid they could be a huge losing player and actually be a winning one.

So as far as statistics is concerned, you have a long way to go before you can really see flaws in your game through stats from raw $WSD%. If you feel like your playing well than you probably are.

Over my last 130k hands I feel like I am running bad. After reviewing my hands, it is hard to fault my play. My WTSD is 25, and my W@SD is 52.35%. I have reviewed every hand in which I won or lost $150 at 200nl, postflop only. I came up with 176 hands, 111 won and 65 lost postflop. Out of the 111 hands won I was way ahead (70/30+) 105 out of 111 when the money went in. Out of the 65 hands I lost I was way ahead 24 times and 5 were pure flips for a total of 29. Out of the 36 times I was behind and lost I only miss played 7 hands. The other 29 hands are under sets, under flushes, big draws that didn't hit, and 2 AA hands against fish. I also miss played 1 hand that I won.

In conclusion:
I sucked out 6 times
I was sucked out on 24 times
I played 168/176 hands correctly for 95.4%

Originally Posted by Galapogos

What would be cool is if there were a program out there that could go through your PT database and analyse if your hands are holding up as often as they should once the money's in. That way you could know if you're running "unlucky"

They're working on these in the 2p2 software forum. Look for the Sklansky Bucks Calculator thread. I'll save you wading through all 10+ pages and summarize... one sounds like it's a huge pain to get running and may have bugs (wildzer0's), another is java-based, easier to run but also may have bugs (advis0r's), and a third is in beta-testing by request only and should eventually blow the other two away (Phil153's @ http://www.pokerevsoftware.com/). I'm gonna wait for Phil's to be released to the public I think.

Lyric posted something a while ago on 2p2 on how ridiculous varience can be even over 100k hands... i'll try to find it.

I feel like i've been running bad ever since i got to 100NL, personally. That's like 130k hands to.

Oh, here it is.

http://forumserver.twoplustwo.com/sh...0&fpart=1&vc=1

Lyric said that if he ran 2.2BB/100 that would = about a 1.78% edge. These graphs just kind of scare me.

Here's something i was thinking about today with varience. I'm pretty sure we all agree that Jamie Gold is a "losing" Tournament player, (or at best a break-even). But he won the WSOP. This means that he'd have to play thousands and thousands of high buy in MTT's before his varience "evened out", and even then he'd probably still be a winner total, even though he's a losing player.

That's pretty ridiculous.

sauce your play is very high varience. if your playing the best poker of your life this means when your running bad you will break even i.e april, and when your running good youll make loot i.e febuary march. we both know bad luck can seem like its going on forever i.e ISF 25 buy in 100nl, any roll ive ever had, you trying to rebuild this year. just chill and play lots of hands your obviously playing well if your breaking even having seen the beats youve been taking

ISF- sticky me every thread pertaining to this so i feel better, as in the 2p2 ones about 100k break even stretches and shit

Originally Posted by Massimo

Oh, here it is.

http://forumserver.twoplustwo.com/sh...0&fpart=1&vc=1

Lyric said that if he ran 2.2BB/100 that would = about a 1.78% edge. These graphs just kind of scare me.

Here's something i was thinking about today with varience. I'm pretty sure we all agree that Jamie Gold is a "losing" Tournament player, (or at best a break-even). But he won the WSOP. This means that he'd have to play thousands and thousands of high buy in MTT's before his varience "evened out", and even then he'd probably still be a winner total, even though he's a losing player.

That's pretty ridiculous.

awesome thread... amazing the differences in outcomes that can occur

Originally Posted by sauce123

ISF- sticky me every thread pertaining to this so i feel better, as in the 2p2 ones about 100k break even stretches and shit

Somewhere in the party regs thread, CTS says that he's had +12ptBB 100k stretches followed by totally breakeven stretches.

Meh there's better graphs.
In 100k hands btw with your winrate you have an expected dowswing of 15 buy ins, and it's not that improbably you have a 20-30 buy in downswing also.

is this "edge" comparable to the ROI for sngs/mtts? i'm assuming it's not....but if you're calculating money in vs money out...

i'm confused

Originally Posted by Irisheyes

Originally Posted by sauce123

ISF- sticky me every thread pertaining to this so i feel better, as in the 2p2 ones about 100k break even stretches and shit

Somewhere in the party regs thread, CTS says that he's had +12ptBB 100k stretches followed by totally breakeven stretches.

this sounds like what i seem to be finding or have seen over my datdabase.

My upswings are roughly 10-20bbs/100 over 30-70k hand samples then i run -3-8bbs/100 when i run bad for extended periods.
This break even stuff is a total loss to me, i dont seem to break even for long periods, maybe 5k hands or so, im either up and running really good or down and running really quite bad (although my downswing levels off because i play tighter, try to pull less moves etc as a natural istinct against running bad)

btw, i thought the graphs in lyrics thread were cool for actually picturing what it looks like to run good/have an edge etc.

this may already have been included in the links (didn't read) but I recall a dataminer posting results where EM2 ran at nearly 11ptbb/100 over 100k hands then ran slightly better than break even for the next 100k hands.

its worse than pointless to quantify variance.

When i heard of a sklansky bux calculator I actually threw up a little in my mouth. I thought "oh this is nice, something that someone who's running bad can prop up against and waste valuable time with, time that should be spent playing more hands."

lol, Renton are you in a bad mood this morning?

no, its just as someone who used to similarly whine about variance, I am certain that thats all this line of thinking is good for.

I've played a ridic large amount of hands, probably more than 95% of the players in my game combined (stars 1/2 fr), and I have seen every aspect of variance. I can tell you variance knows no boundaries, and it can't be reasoned with. Its incredibly easy to lose 10, 15, 30 buyins regardless of how tight/loose/passive/aggressively you play. The bottom line is unless you are a 8ptbb poker savant, you will have sick sick swings in your poker career. It doesn't matter if you play 4 or 16 tables, it doesn't matter if you play 20k hands or 100k hands every month, you will have losing months, even if you solidly beat the game (by solidly i mean 2+ptbb/100).

Poker variance is so insane that I wouldn't recommend professionally playing to anyone. I'm certain I could make 50-150k a year playing poker if I played enough, but Im probably still eventually going to give it up and work as an architect, making likely less money.

At least as an architect I won't fall deeper into stygian despair whilst developing truer and more pronounced hatred for the human heart.

PS: I had a +500 session at 1/2 today and am up 20 something buyins this month in not a whole lot of hands. This isn't a tilt rant.

i feel like after a year of playing poker variance wont bother me anymore.... wishful thinking??

Originally Posted by Renton

When i heard of a sklansky bux calculator I actually threw up a little in my mouth. I thought "oh this is nice, something that someone who's running bad can prop up against and waste valuable time with, time that should be spent playing more hands."

I want a sklansky bux calculator for the excact opposite reason; I want to assure that I'm not just running hot. I know that there are (and always will be) leaks that I need to work on, but it's also nice to know that I'm at least doing something right.

HalvSame - sobers up in an hour!

Originally Posted by Renton

its worse than pointless to quantify negative variance.

FYP.

Originally Posted by sauce123

i feel like after a year of playing poker variance wont bother me anymore.... wishful thinking??

depends if you move up or not.

no, it will always bother you

i thought after i had 100k breakeven stretch that it would never both me, but when you keep moving up levels the downswings get bigger :\

Damn, I really don't want to move up. My recent 8x buy-in downswing is making me hate poker.

gabe- im almost there, now at a 60k breakeven stretch and going doooowwwwnnn

Today:

Originally Posted by sauce123

i feel like after a year of playing poker variance wont bother me anymore.... wishful thinking??

*very* wishful. Although in my experience as soon as someone recovers from that first -$1,000 day, everything gets a LOT easier to take.

Of course, I shrank my bankroll from $10,000 to $100, and now back up to 3,000 but a -10% drop in a day makes me want to puke again. Starting the process alllll over.