My friend Markov plays until he\'s even, but...
 

Let's say I have a friend, we'll call him Fairly Aggressive Guy ([censored]) and he plays PotLimit poker. He has a funny style tho: He will happily quit if he gets ahead by 1 buyin or more. But if he is stuck, he will not quit unless he hits -10 buyins. I know he is a strong winning player. He has told me over the last 200 days he has played, he has 180 winning days. If it matters, his style is fairly aggressive, pushes small edges to maximize EV.

my question:
How good must his longterm winrate be to make this reasonably within the ballpark (ie. he's not running super hot)?

Or, If his longterm winrate is something like 8 PTBB/100, how likely is it that these results are typical/expected?

I'm basically trying to get a sense of "is this is realistic sustainable result, or is he full of [censored] running good"

-g

Re: My friend Markov plays until he\'s even, but...
 

If the only time he quits is when he's up 1 or down 10, then he's won 1 buyin 180 times and lost 10 buyins 20 times, which means he's a losing player.

A break-even player, using your friend's system, would stand to have a winning session ten times out of eleven.

Re: My friend Markov plays until he\'s even, but...
 

it's all one long game. which means it doesn't matter when you leave or come in. assuming all other variables are constant, of course.

consider this as a coin flipping question and you'll see that it doesn't matter at all.

Re: My friend Markov plays until he\'s even, but...
 

[ QUOTE ]
My friend Markov

[/ QUOTE ]I liked this part. Is he from Monte Carlo?
-Sam

Re: My friend Markov plays until he\'s even, but...
 

If he plays exactly the same all the time and is a winning player then he will be a long term winner.

More likely though, even if he is a winning player, he is costing himself money by playing this way.

When he is winning, then it is more likely than average that he is playing at a weak table and therefore has a higher EV than normal and should play longer. Thus in this case, quitting early is often costing him money.

When he is losing, especially when down multiple buyins, its more likely than normal that he is either tilting or at a tough table and is therefore not earning as much as normal. So this situation also costs your friend money.

So basically your friend is probably hurting his winrate significantly.

Re: My friend Markov plays until he\'s even, but...
 

Note: I'm sure there are better statistical methods of working this out, but since no one else answered this should give a quick and dirty answer.

Since he can't play more than around 2000 hands/day, there will be many days where he finishes a loser but not down 10 buyins. So I'm going to restate the question and assume:

1. He never loses for the day after reaching 1 buyin (50BB)
2. He will play up to 2000 hands if he hasn't reached one buyin.
3. He is a 10BB/100 winner with a 50BB/100 standard deviation.

I don't know how to incorporate (1) into the calculations, but you can work out a lower bound by determining how often he'd be a winner if he played 2000 hands every day.

With a 95% confidence interval, his distribution will look like:

20*10 ± 1.96*20*50/ sqrt(20) = 200BB ± 438BB

So the distribution ranges from +638BB to -238BB, with a mean of 200BB. Thus, roughly 67% of his days will be winning days. If he stops each day as soon as he reaches 50BB, this will push this number significantly higher, as it cuts out the days he reaches 50BB and ends up losing. It's easy to see how this could push his daily win% to over 80%.

There are other factors that could add credence to his figures and bring the total closer to his 90% figure:

1. His winrate includes period of tilt and poor play, and is actually higher than stated when he's playing good. Sheer determination and the goal of getting unstuck may make him a play better poker. Given that he's something of a Fairly Aggressive Guy, this seems very likely.

2. His table image deteriotes after losing for a while, and he is able to exploit this and artificially increase his winrate to get back up.

3. The quality of his reads and/or domination of a table increase with time spent there.

4. He has a higher winrate when playing a deep stack.

So I don't think it's an example of running good at all, seems pretty standard results for a Fairly Aggressive Guy (or even a Very Aggressive Guy, as may be the case here) playing a stop-win method.

Re: My friend Markov plays until he\'s even, but...
 

Excellent, just what I was looking for. It seems that [censored]'s 80-90% of days having won is sustainable then. It just seems in my gut like that is too high, but apparently not".

PS I love you Phil

-g

Re: My friend Markov plays until he\'s even, but...
 

[ QUOTE ]
[ QUOTE ]
My friend Markov

[/ QUOTE ]I liked this part. Is he from Monte Carlo?
-Sam

[/ QUOTE ]

But of course!