# SSS, bb/100 and Allin EV line

• Bronze
Joined: 01.02.2009
If I look at my Allin EV line instead of my actually winnings to calculate my bb/100 with SSS, what sample size do you think I'd need in order to get a close to the real result?

10k hands enough?
• 8 replies
• Bronze
Joined: 27.05.2005
It's always hard to determine the real winrate, and i think 10k hands are certainly not enough.
Imo 40k-50k is the minimum you should look at, especially when you're playing something as swingy as the sss.
• Bronze
Joined: 01.02.2009
Yes I know that 50k is the minimum to have some confidence into how real your winrate is with SSS.

What I was wondering is, what if we look at our EV winnings instead of the money we actually did win, which lowers the variance in winrates in smaller sample sizes SIGNIFICANTLY, how many hands would we need then ?

I do realize that luck is still a factor in that how many good hands you get and what hands your opponents get, cause every time you get it in with kings vs aces you lose 60% of your stack in EV without making an actual mistake, but even then, the EV line is nowhere near as incosistant and swingy as the actual winnings line.

So, yes, for winrate, based on results you need 50k hands MINIMUM, what about if we look not on actual results, but at the results IN EV ?
• Bronze
Joined: 09.06.2007
Originally posted by whateverdude
Yes I know that 50k is the minimum to have some confidence into how real your winrate is with SSS.

What I was wondering is, what if we look at our EV winnings instead of the money we actually did win, which lowers the variance in winrates in smaller sample sizes SIGNIFICANTLY, how many hands would we need then ?

I do realize that luck is still a factor in that how many good hands you get and what hands your opponents get, cause every time you get it in with kings vs aces you lose 60% of your stack in EV without making an actual mistake, but even then, the EV line is nowhere near as incosistant and swingy as the actual winnings line.

So, yes, for winrate, based on results you need 50k hands MINIMUM, what about if we look not on actual results, but at the results IN EV ?
Who cares? Do you want to play poker or do you want to be a winrate label? If your winrate during the last 200k hands was 15bb/100, do you think that is your 'true' winrate? There is no such thing as a true winrate. bb/100 shows the past, not the future.
• Black
Joined: 20.02.2008
Originally posted by whateverdude
Yes I know that 50k is the minimum to have some confidence into how real your winrate is with SSS.

What I was wondering is, what if we look at our EV winnings instead of the money we actually did win, which lowers the variance in winrates in smaller sample sizes SIGNIFICANTLY, how many hands would we need then ?

I do realize that luck is still a factor in that how many good hands you get and what hands your opponents get, cause every time you get it in with kings vs aces you lose 60% of your stack in EV without making an actual mistake, but even then, the EV line is nowhere near as incosistant and swingy as the actual winnings line.

So, yes, for winrate, based on results you need 50k hands MINIMUM, what about if we look not on actual results, but at the results IN EV ?
you won't find a graph displaying your EV anywhere. the all-in EV lines etc are still subject to huge variance, but yes, less than the actual winnings, so a smaller sample can give you the same information.i guess if you play 30k hands of SSS and your all-in EV line is pretty much straight, then you can get some idea of your expected winrate by looking at it. SSS regs pleas correct me if i'm wrong
• Bronze
Joined: 01.02.2009
Yeah, I meant "Allin EV".

Anyways, I tought I won't be the first one who has asked this question, and I will get a quick answer, not really a discussion.

Guess I am just gonna wait till the line looks straight, cause thats when you know you've eliminated variance(sort of).
• Bronze
Joined: 27.05.2005
The problem is there's a lot more than how you do in all in situations that can deceive your actual winrate.

I.e. if you flop a set on 50NL three times in a row and always get it in vs an overpair you'll be \$30 above EV.
Then you get AK, the opponent has QQ and you're all in preflop.
Now you are running below EV about \$20!

Now say these situations are scattered over 10000 hands - let's assume somehow the other 9996 hands exactly reflect your winrate in themselves (let's assume you are a breakeven player, so the winrate would be 0bb/100), you have an EV of exactly \$0.00.

Since you won an EV of 600bb*0.9+200bb*0.5=640BB you're playing with 6.4bb/100 of EV.
So these 4 hands out of 10000 make you believe you are a winning player with a winrate of 6.4bb/100 while you actually just ran hot.

That is why you need large samplesizes to determine your winrate.
• Bronze
Joined: 21.09.2009
DAMN! 50k hands! I've only played 8k. Should I play more before moving up the limits?

Thanks
• Bronze
Joined: 27.05.2005
It depends.
If you feel comfortable with the games on the upper limit and your bankroll is big enough i'd try to take a shot, too.