![]() |
#21
|
|||
|
|||
![]()
[ QUOTE ]
[ QUOTE ] Yup, winrate takes a really long time to converge properly. You can have a high theoretical winrate and still break even for close to 30k hands, which is why you need really large samples. [/ QUOTE ] Just a hunch, but anyone who breaks even for 30k hands probably isn't a very good player, and therefore wouldn't have a high theoretical winrate. [/ QUOTE ] I think I read a post were CTS said that he lost money or broke even over like 100K so this is obv correct. |
#22
|
|||
|
|||
![]()
[ QUOTE ]
[ QUOTE ] Yup, winrate takes a really long time to converge properly. You can have a high theoretical winrate and still break even for close to 30k hands, which is why you need really large samples. [/ QUOTE ] Just a hunch, but anyone who breaks even for 30k hands probably isn't a very good player, and therefore wouldn't have a high theoretical winrate. This is a good thread, here's hoping some more regs post. [/ QUOTE ] It would be pretty amazing to see any 5ptBB+ winners who have 300k+ hand sample to NOT have at least one 30k hand breakeven periods. Anyway, lets not hijack this thread, this topic usually hijacks this type of thread every single time and every single time an old time poster comes and chimes in how they have had a great run over 100k and a bad run over 100k, posters such as fsu, HEK, cts have all said it. Plus there's a few sites in which you can plug your win rate, variance and hand sample in and it'll tell you. I'm 5.8ptbb over 5/10 10/20 for +250k hands. I'll merge all my databases at the end of the year. |
#23
|
|||
|
|||
![]()
My sample is teh tiny as I am not playing full-time, yet. I'll contribute anyway.
For MSNL (I mix limits as I play): 50K hands total across 2/4, 3/6, and 5/10 (20K, 15k, 15k): 6.03 PTBB/100 FWIW, I am very aware that the sample is small, but I like to believe that it is enough for me to know that I am at least a winner at these stakes. |
#24
|
|||
|
|||
![]()
infinite
|
#25
|
|||
|
|||
![]()
[ QUOTE ]
Anyway, lets not hijack this thread, this topic usually hijacks this type of thread every single time [/ QUOTE ] Seems like break even stretches are a fitting part of a winrate thread, right? [ QUOTE ] every single time an old time poster comes and chimes in how they have had a great run over 100k and a bad run over 100k, posters such as fsu, HEK, cts have all said it. [/ QUOTE ] Who are these guys to be an authority on anything? True math is not subjective. [ QUOTE ] Plus there's a few sites in which you can plug your win rate, variance and hand sample in and it'll tell you. [/ QUOTE ] Where's the website that shows you what break even stretch you'll have? [ QUOTE ] I'm 5.8ptbb over 5/10 10/20 for +250k hands. I'll merge all my databases at the end of the year. [/ QUOTE ] That's a sexy win rate. I bet you haven't had 30k breakeven. Do you coach? |
#26
|
|||
|
|||
![]()
[ QUOTE ]
plz come up with an exact number where they converge with reasoning, otherwise i will just continue to think you have no statistical background at all. why is 150k the magic number? how do you take in account the level that a player is playing at? why can't players be playing like [censored] for 30k hands? yadadada [/ QUOTE ] 150K is not a "magic number". I just threw out an arbitrarily large number that is not quite the long run but not quite meaningless. [ QUOTE ] To calculate this you'd need to know a typical standard deviation for aggressive, short handed games. You can get this in PokerTracker by clicking "more detail" on the session notes tab. The value will be in BB/100. Once you have this, you can calculate the 95% confidence interval around your winrate like this: true winrate = reported winrate ± 1.96 * standard_deviation/sqrt(hands/100) [/ QUOTE ] and 1 in 20 times, a persons true winrate will be outside the range you get from this calculation |
#27
|
|||
|
|||
![]()
if you're playing abc, low variance poker, you dont need anywhere near 150k hands to get your winrate to converage to a real number.
unless you're grimstar... in which case 150k hands still gives you an error term of like 6. statistics joke... hah [ QUOTE ] [ QUOTE ] [ QUOTE ] [ QUOTE ] [ QUOTE ] [ QUOTE ] can everyone stop posting sample sizes of under 150K hands? k thx [/ QUOTE ] sry we don't all run at 2ptbb/100 but play a gagillion hands per day [/ QUOTE ] then the accuracy of your results are almost worthless [/ QUOTE ] o rly? [/ QUOTE ] Yup, winrate takes a really long time to converge properly. You can have a high theoretical winrate and still break even for close to 30k hands, which is why you need really large samples. [/ QUOTE ] plz come up with an exact number where they converge with reasoning, otherwise i will just continue to think you have no statistical background at all. why is 150k the magic number? how do you take in account the level that a player is playing at? why can't players be playing like [censored] for 30k hands? yadadada [/ QUOTE ] |
#28
|
|||
|
|||
![]()
anyone know how much longer it should take to converge for HU, just a correlation to your variance which is higher than 6m?
|
#29
|
|||
|
|||
![]()
since Oct 1st 2006 (when i started playing solid):
6.16ptbb/100 over 243K hands. roughly: 50% at 2-4nl (at about 6ptbb/100) 30% at 1-2nl (at about 7 ptb/100) 20% at 3-6nl and 5-10nl (at about 5ptbb/100) MT ratio 6.26 hourly rate $218 |
#30
|
|||
|
|||
![]()
Played around 100k hands of 5/10 at 5.5ptbb/100 despite sucking to begin with. Have around 60k hands of 10/20 at the moment winning a little less after two 20+ buyin downswings.
30k break even is nothing btw. |
![]() |
|
|