How many hands to know sth?

At what sample of played hands should I be looking for to know my bb/100 is in range of +/- 1bb/100 of my EVbb/100, with let's say around 75% certainty?

Is it more like 50k or 200k hands? I don't need precise answer.

it's impossible to ever know. you need too many hands and the games/conditions are changing all the time way too quickly compared to how long it takes to reach such epic samples to begin to start getting winrate close. don't worry about it

edit: but if you insist on indulging an exercise in pointlessness, maybe playing around with the stuff on http://www.evplusplus.com/ will bring you closer to the answer you're looking for.. it is an interesting website in general

i'd say anything larger than a 250k hand sample size will give u a "feeling" of your winrate etc.

of course variance can play a huge role in a sample this small tho.

once you hit like 1 million hands + you'll get a more accurate picture but as ket said games are consistently changing. Although I think that matters less since our winrates are a reflection of our overall edge on the games and we won't have edges on a game that's consistently changing without doing adjustments ourselves.

On May 03 2012 13:33 YoMeR wrote: i'd say anything larger than a 250k hand sample size will give u a "feeling" of your winrate etc. of course variance can play a huge role in a sample this small tho. once you hit like 1 million hands + you'll get a more accurate picture but as ket said games are consistently changing. Although I think that matters less since our winrates are a reflection of our overall edge on the games and we won't have edges on a game that's consistently changing without doing adjustments ourselves.

it often do but sometimes it doesnt.

On May 03 2012 13:10 Ket wrote: it's impossible to ever know. you need too many hands and the games/conditions are changing all the time way too quickly compared to how long it takes to reach such epic samples to begin to start getting winrate close. don't worry about it edit: but if you insist on indulging an exercise in pointlessness, maybe playing around with the stuff on http://www.evplusplus.com/ will bring you closer to the answer you're looking for.. it is an interesting website in general

Thanks.

So the quick answer is that with for example 3evbb/100, with 75% confidence and std dev. 40, we stand to be in a range of

1bb-5bb/100 aftter 50k hands,
1,5-4,5bb/100 after 100k hands,
1,8-4,2bb/100 after 150k hands,
2-4bb/100 after 200k hands,
2,1-3,9/100 after 250k hands.

So seems that after around 200k hands, your winrate is probably (75%) within +/-1bb/100 range.
But again, you change std dev. from 40 to 60, and you need close to 500k hands to achieve same level of certainty.

Hm, let's say Phil Ivey on average played 30hands/hour live, was playing for 20yrs, on average 220/365 days, and when played poker played for 8hrs.

Calculating...
.
.
.
so that's around 1mln hands, so yep he's prob a winner. Yet, again, how many "Phil Ivey's" had an idea to play poker for life, and we see just one lucky bastard. Same thing with universe, but that's a different story for a different topic I guess.

I used to think 250k would give you an idea way back when.

250k isn't even close.

You see all kinds of shit like people going 100 buyins below or above ev in 200k hands. That's measurable, but there is similar unmeasurable variance that occur as well.

I agree with Ket, there really isn't any way to have a real idea about your winrate. Variance in poker is so huge and games are always changing.

Best to not worry about it tbh and just focus on improving/playing your best/putting in hours.

On May 03 2012 17:20 Stim_Abuser wrote: I used to think 250k would give you an idea way back when. 250k isn't even close. You see all kinds of shit like people going 100 buyins below or above ev in 200k hands. That's measurable, but there is similar unmeasurable variance that occur as well. I agree with Ket, there really isn't any way to have a real idea about your winrate. Variance in poker is so huge and games are always changing. Best to not worry about it tbh and just focus on improving/playing your best/putting in hours.

This. The real kicker is that despite how many hands you play the big variance hands aren't played all that often. The number of times you go all-in before the river in a 100k hand sample really isn't all that many. We are talking well below the mathematical threshold for your results to reflect your true long-term win-rate. All you can do is play and focus on learning. Some of us run great some of us don't. Not worth thinking about.

depends, if your struggling at micros, it's likely not due to variance... I always hated EV, it's an excuse for people to not take responsibility for their results. Theres also crap like constantly getting set vs higher set or overpair vs higher overpair which by any calculator counts you are getting your money in super bad (its possible to run under ev in all in pots, but over hidden ev in pots where your coolering people, and complain that you run bad, when in reality you are exactly at 0ev)

Yea I definitely agree that effort should be spent in improving your game instead of finding out stuff like "true winrates" although it's a cool idea what purpose does it have other than padding someone's ego?

look at the cashier, the best winrate calculator ever.

ev line has always been very far from my true winrate. Some of my biggest winning periods i have run below ev. just recently i lost10k bb's, over 200k hands and ran on ev the whole time.

last 500k hands i've lost the biggest amount during the period i had improved a ton, and was playing my best.

also your standard deviation is important, find that in hm/pt and put it into the variance calculator. mine is 130 but probably because i play a lot of hands

where is 40bb/100 std dev estimate coming from? you a FR player Try2Be?

what does a lot of hands how to do with your std dev stroggoz?

nobody pays attention to EV anymore Tom (i don't think...) i remember when people use to bitch about it, always made me lol.

Well assuming your getting better the more hands your playing isn't your winrate pretty variable, as the first 50000 hands you play may be a lot worse than the 2nd 50000?

On May 03 2012 18:13 TalentedTom wrote: depends, if your struggling at micros, it's likely not due to variance... I always hated EV, it's an excuse for people to not take responsibility for their results. Theres also crap like constantly getting set vs higher set or overpair vs higher overpair which by any calculator counts you are getting your money in super bad (its possible to run under ev in all in pots, but over hidden ev in pots where your coolering people, and complain that you run bad, when in reality you are exactly at 0ev)

Never thought about that! Sounds reasonable, maybe the focus on the EV-graph is to big?

True pro's don't need hh's, just cross fingers and call. :-)

On May 04 2012 19:33 VanillaDutch wrote: where is 40bb/100 std dev estimate coming from? you a FR player Try2Be? what does a lot of hands how to do with your std dev stroggoz? nobody pays attention to EV anymore Tom (i don't think...) i remember when people use to bitch about it, always made me lol.

i open a lot of hands, and play a high vpip/pfr since im playing at the ante tables. i'm not sure but i think since i have higher vpip than most regs it means i get into a ton more spots, and alot more marginal spots which increases my variance. It also increases my winrate tho, which helps cut down on variance.

Having a higher winrate has nothing to do with variance lol.

It just means that if your winrate is higher, that when you're running bad that you're only winning at 1bb/100 or losing just slightly. If your winrate is lower, obviously when the lower spectrum of variance sets in you're gonna lose more because your winrate (or in that case loserate) is smaller (loserate bigger).

2500.

Well this is embarrassing, it turns out 2500 hands ISN'T enough. Boy is my face red.