how long before you see an accurate winrate?

[zombies kill]

i searched for old posts about this but nothing fitting came up.

about how long does it take to get a reasonably accurate winrate at the lower limits online... say 2/4 or 3/6. im pretty sure its around 10k hands?

thanks

[SparkyDog]

Win Rate +/- (confidence factor)(SD/100)/(sqrt of hands played in 100's)

Factors give the following confidences:
1 - 68%
1.96 - 95%
2 - 95.4%
3 - 99.7%

For example, if you had a winrate of 2BB/100 and SD/100 of 15 for 10000 hands you would do the following for 99.7% confidence:

2 +/- 3*(15)/(sqrt100)
2 +/- 4.5 BB/100

[SinCityGuy]

[ QUOTE ]
about how long does it take to get a reasonably accurate winrate at the lower limits online... say 2/4 or 3/6. im pretty sure its around 10k hands?

[/ QUOTE ]

I've had 10K stretches where I've made 7bb/100, and I've also had 10K stretches where I've lost money.

You need about 100,000 hands to get a fairly accurate rate.

[SomethingClever]

Eleventy billion hands. No joke!

[Snoogins47]

[ QUOTE ]
Eleventy billion hands. No joke!

[/ QUOTE ]

That's more like it. I think six digits is really when you can have some sort of an idea. *cue streaks spreadsheet*

STREAKS! (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Number=1350059&page=&view=&s b=5&o=)

[Richard Berg]

10K hands may be enough to determine that you're a winning player, but not even in the right order of magnitude to name a winrate.

[zombies kill]

for the love of Jehovah... im very aware of the fact that you need 100,000s of hands to smoke out variance... but thats really not my question (i shouldve been clearer in my original post i suppose). im referring to moving up a level or trying out a new game. provided that youre an experienced player and can tell youve gotten a relatively normal run of cards, then at what point can you assume that your winrate has become a decent forecast of future results?

there... thats a little clearer i think
thanks

[bisonbison]

Uh, I think they answered that question.

It depends on you: what do you consider a decent forecast?

[zombies kill]

ok then, nevermind. i feel as though i have enough experience to be able to differentiate between a good/bad/average run of cards/results. i just wanted to hear some other opinions. but for the record i dont believe the answer to be anywhere near 6 digits, as long as an experienced and knowledgeable player can discern the cards/results are relatively average.

[chesspain]

[ QUOTE ]
ok then, nevermind. i feel as though i have enough experience to be able to differentiate between a good/bad/average run of cards/results.

[/ QUOTE ]

So why are you wasting our time?


[ QUOTE ]
But for the record i dont believe the answer to be anywhere near 6 digits, as long as an experienced and knowledgeable player can discern the cards/results are relatively average.

[/ QUOTE ]

Again...why are you wasting our time?

[afk]

Check out this (http://forumserver.twoplustwo.com/showflat.php?Cat=&Number=1342415&page=2&view=colla psed&sb=5&o=&fpart=1).

[Hoi Polloi]

Aristotle said you cannot know if a man had a happy life until he is dead.

One could certainly make the argument that you cannot tell whether a man is a winning player until he is dead. /images/graemlins/wink.gif

Perhaps this applies to winrate as well.

[zombies kill]

because like i said in the first sentence after the one you highlighted is that i wanted to hear other 2+2ers opinions.

and you dont know me well enough to speak to me like a dickhead so act like an adult and watch your tone when you speak to me. ill leave it at that

[bisonbison]

Please, his name is Dr. Dick.

[zombies kill]

thank you afk, i missed that

[chesspain]

[ QUOTE ]
and you dont know me well enough to speak to me like a dickhead so act like an adult and watch your tone when you speak to me. ill leave it at that

[/ QUOTE ]

You're not gonna cry, are you Peaches?