MicroLimit Personal Progress...All Posters Post

[helpmeout]

Here is my current progress at pacific for anyone who is interested.

31-Jul deposited $50+$13 bonus

9-Aug $81.25 (0v) 47hrs +$18.25 3.88BB/HR 5c/10c

21-Aug $157.17 (41.26v) 28hrs +$34.66 6.19BB/HR 10c/20c

6-Sep $267.13 (19.14v) 58hrs +$132.08 4.55BB/HR 25c/50c

v=variance (SNGs+MTT)

[PokerBob]

I don't think you need 500BB to play at Party 1/2. I keep mine around 300. That said, you haven't played nearly long enough to be sure that you are a winning player. If you can get your bb/100 at 0.5/1 to +2-3bb/100 after 50,000 hands, then I think you can safely move up to 1/2 with $600. The games really don't play that much differently.
Go Badgers. I'm class of '93.

[uw_madtown]

[ QUOTE ]
I don't think you need 500BB to play at Party 1/2. I keep mine around 300. That said, you haven't played nearly long enough to be sure that you are a winning player. If you can get your bb/100 at 0.5/1 to +2-3bb/100 after 50,000 hands, then I think you can safely move up to 1/2 with $600. The games really don't play that much differently.
Go Badgers. I'm class of '93.

[/ QUOTE ]

You don't think that 20-25k hands is enough of a sample? I guess I just keep hearing 10k being thrown about a lot and figured 20-25k hands and 500 BB would be playing it safe on all counts. If I have 4 BB/100 when I hit 20k hands, then I'll have 500 BB for 1/2 anyway. So, I mean, if I'm a winning player and if I wait until 20-25k hands, I'll have that big a bankroll anyway. Not planning on taking any out, and not planning to hop to 2/4 either, so...

But yeah, is 20-25k hands still not a big enough sample? 50k just sounds like a LOT of hands to me, especially at the .5/1 level. I could see doing that at 2/4 before moving to 3/6. I dunno, I could be wrong.

Also keep in mind that I have played and lurked around here a lot, so I'm not starting from scratch. I've read SSH (re-reading it constantly now) and have done well at 2-4 B&M the handful of times I've gone. I dunno. Maybe I should wait until 50k hands.

Oh, and go Badgers. [img]/images/graemlins/cool.gif[/img]

- UW

[brettbrettr]

Class of 2000- Go Badgers.

[sluttysteve]

7600 hands at Party 1/2 since I've started keeping track.

VPIP 14.59%.
PFR 8.40%.
BB/100 = 3.67.

Am I being irrational in saying that I feel like my BB/100 should be higher, and is only at 3.67 because of an inordinate amount of bad luck?

[dfscott]

[ QUOTE ]
But yeah, is 20-25k hands still not a big enough sample? 50k just sounds like a LOT of hands to me, especially at the .5/1 level. I could see doing that at 2/4 before moving to 3/6. I dunno, I could be wrong.

[/ QUOTE ]

This is probably more info than you want, but if you want to "prove" statistically that you are a winning player, someone posted a formula a while back. If you've got PokerTracker, it's trivial to calculate.

Just take your Std Deviation/100 hands, and divide that by the square root of the number of 100 hand samples you have. The gives you your standard error (StdErr). If your BB/100 > (4 * StdErr), then you are (from a statistical point of view) a winning player.

For example, if I look at my totals:
Hands: 27522
BB/100: 4.78
StdDev/100: 19.99

StdErr = 19.99 / SQRT(275.22) = 19.99 / 16.59 = 1.20

Since my BB/100 (4.78) is less than 4 times my StdErr (4 * 1.20 = 4.80), I can't say for certain that I'm a winner. However, I am within 3 StdErr's, and at that point, you can be 99% certain, which is pretty good.

The original thread had a lot of other info in it, but I can't find it right now. A search for "Standard Error" might turn something up.

[PokerBob]

Maybe I'm just paranoid, but in my experience it is possible to simply "run hot" for 20K+ hands. By the same token, one can "run cold" for that many hands (I seem to be in the middle of that now.) That said, you seem intent on moving up, so go for it. You can always move back down to build your bankroll if you take a big hit.

One other thing to consider when moving up: Make sure you think of things in terms of bets, as opposed to actual money. I know when I tried to move up to 2/4, it was hard not to think "damn, I'm down $50 already. If this keeps up.....". That is only 12.5 BB, which doesn't phase me at 1/2 but drove me nuts at 2/4. Make sure you can focus on proper play, not your bankroll.

[RED_RAIN]

Sounds excellent. Your numbers look great for now.

[DrBob]

I don't have my exact stats here at work. I started Paradise in late May with their $2 free, and have managed to work it up to $140, playing initially $0.02/$0.04, mostly $0.05/0.10, and recently $0.10/0.20.

My stats improved considerably after reading SSH. At $0.05/0.10 I ran 9.??BB/100 over about 8000 hands, and am also 9-something at $0.10/0.20 but with less than 1000 hands. I think my $0.05/0.10 pre-SSH was around 5BB/100 and around 13BB/100 for maybe 3000 hands post-SSH, but I'm not certain my recollection of this is right.

[DrBob]

The std error stuff is pretty good, but assumes a normal distribution. Unfortunately the distribution of poker winnings per hand is grossly non-normal. The Law of Large Numbers says that doesn't matter with a large enough sample, but it's hard to say what's large enough.

I've developed some non-parametric tests that can be applied to your poker stats. You start by developing a table (do it in a COPY of you poker tracker database) where there's an entry for each outcome (in units of BB) that gives the number of hands with that monetary outcome. This is real easy if you have and can use MS Access. Just plotting this data is interesting.

The test is pretty intuitive -- use your table and "pretend" that it represents the underlying "true" distribution. Draw 1000 random samples (with replacement) from this distribution, where each sample size is equal to the number of hands you've actually played so far. Sort the 1000 samples by each sample's overall BB/100 results. You can be pretty confident (about 90%) that your actual BB/100 is better than the 100th lowest ranking BB/100 sample, and also (about) 90% confident that it isn't better than the 100th (out of 1000) highest ranking sample.

I've got some software (visual basic 6) that works this, that I'll share. However, ask me next week (Sep 13) as I won't be home until then.

Regards