Does being up 300 BB prove you are a winning player?

[ctv1116]

People always suggest that 300 BB is a good bankroll for limit hold'em. They say that it would be extremely unlikely for a winning player to have a 300 BB downswing. So is the opposite statement also true: it is extremely unlikely for a losing player to have a 300 BB upswing. And if a losing player cannot have a 300 BB upswing, then if he/she does have a 300 BB upswing, then the player must be a winning one. Is this analysis correct?

[Eihli]

No.

[j0n_blayze]

why not Mr. Kool?

[NLSoldier]

I believe there are a lot of winning players who have had 300BB downswings, and this would lead me to believe that being up 300BB does not mean you are definately a winning player.

[Irieguy]

Remember that the "300BB bankroll" is based on the concept that a 300BB loss would be 3 Standard Deviations from a winning player's mean win rate over some long period of time.

But that mean is a number greater than zero if the player is winning. So a 300BB win is not statistically equivalent in magnitude to a 300BB loss... it is less.

That's the stupidest, most nitpicking reason why the answer to your question is no. But your concept is correct in a way. The more money you win, the more likely it is to be due to skill. The converse is also true.

Irieguy

[pzhon]

[ QUOTE ]
They say that it would be extremely unlikely for a winning player to have a 300 BB downswing. So is the opposite statement also true: it is extremely unlikely for a losing player to have a 300 BB upswing. And if a losing player cannot have a 300 BB upswing, then if he/she does have a 300 BB upswing, then the player must be a winning one. Is this analysis correct?

[/ QUOTE ]
To show that a 300 BB swing is rare, you usually assume that a player not only wins, but wins a substantial amount, say 1 BB/hour. A player who loses at least 1 BB/hour will rarely have a 300BB winning streak. A priori, someone who loses 0.1 BB/hour might often see 300BB winning streaks.

Neither 300 BB streak would be conclusive, just evidence.

[Ed Miller]

A 300 BB losing streak is very unlikely for someone who wins 1 BB/hr with a standard deviation of 10 BB/hr. In the same way, a 300 BB winning streak is very unlikely for someone who loses 1 BB/hr with a standard deviation of 10 BB/hr.

But almost no one loses 1 BB/hr with a standard deviation of 10 BB/hr. To get a standard deviation that low, you have to play tightly. If you play tightly, you won't lose 1 BB/hr.

A player who loses at a rate of 1 BB/hr will very likely have a much higher standard deviation. With a high standard deviation, you could easily have a winning streak larger than 300 BB and still be a losing player.

So the answer to your question is a definite NO. Winning 300 BB is in no way conclusive evidence that you are a winning player.

[InfernoLL]

I guess the next question is, given the rate of the average losing player (who has a lose rate and standard deviation I don't know), how many big bets do you have to win to prove you are better than this player, to within the same margin that losing 300 BB proves you are worse than the typical winning player?

[Grisgra]

Answer is, I think you have to just sit and do the math -- BB/100 hands (or per hour), SD, and # of hands/hours, and calculate a confidence interval.

I've always been interested, though, in exactly what the distribution would look like -- it's not necessarily normal given the fact that different sessions take place at different times, different tables, with different levels of concentration. Could for some people be skewed or even bimodal.

[TonyBlair]

I like Irieguy's answer, the part I understand anyway. What you are up or down initially can only ever be an indication as to what your probable skill level is. Theoretically, you could be playing the best poker on the planet every day for 10 years solid and still be down. You wouldn't be God-loving though.
Play for long enough and you'll just know the answer unless you have no analytical skills (in which case you're losing). No, not because of what you can or can't see in your account but because you can always see if your decisions have carried +ve expectation. (Side thought: do they make Braille chips or cards?).

[Nottom]

From a post I made in the famous Homer Streaks (http://forumserver.twoplustwo.com/favlinker.php?Cat=&Entry=6741&F_Board=smallholdem& Thread=445039&partnumber=&postmarker=) thread.

[ QUOTE ]
All the playing with numbers seemed like fun so I made a spreadshhet to represent a LAG with SD of 20 and a EV of -.25/Hour

The first time I ran it, he was up 1800 after 2500 hours after peaking at almost 3K after 2000 hours.


[/ QUOTE ]

Its actually pretty common for an aggressive yet losing player to post some very big wins. If he happens to string a couple of those wins together he could easily win 300BBs or more.

In general, it is usually much easier for a losing player to win than for a winning player to lose. (Basically for all the reasons Ed mentioned in his post)

[blendedsuit]

are BB's "big bets" or "Big Blinds"???
thanks

[purnell]

Typically on these boards BB refers to big bets.

[mikimaus]

1 BB, 2 BB, winning or losing expectation per 60-100 hands has practically no effect to swings on the beginning, only in the long run (6000 hands is not a long run).

It's better to think about hands; I consider that 18,000 hands of limit holdem (FR, SH, HU) is enough to prove anyone a loser.

[mikimaus]

At limit form it means big bets. At big bet poker it means big blinds as there's no big bets, but currently because of Poker Tracker (and possible limit reasons) it might rate to mean 2x the big blind there too (PT gives its win amount like that).

[pzhon]

[ QUOTE ]
1 BB, 2 BB, winning or losing expectation per 60-100 hands has practically no effect to swings on the beginning, only in the long run (6000 hands is not a long run).

[/ QUOTE ]
Do you think all of the mathematicians are just blowing smoke up your as[/i]s?

300 BB downswings don't happen in one session (at least, not to sane limit players). The danger is that you experience a series of bad and break-even sessions totalling thousands of hands.

What does winning 2 BB/100 mean after 6000 hands? It makes a difference of 120 BB, roughly 1 standard deviation. Being down 300 BB after 6000 hands is

/images/graemlins/diamond.gif 3.5 standard deviations below the mean for a player who wins 2 BB/100. That happens about 1 time in 4300.
/images/graemlins/diamond.gif 2.5 standard deviations below the mean for a break-even player. That happens about 1 time in 160.
/images/graemlins/diamond.gif 1.5 standard deviations below the mean for a player who loses 2 BB/100. That happens about 1 time in 15.

To put it another way, a break-even player might be down 300 BB after 6000 hands. A winning player hit by the same bad luck would only be down 180 BB. The winning player would still have a long way to fall before dropping 300 BB. A losing player facing the same would be down 420 BB, more than twice as much as the winning player lost.

These assume the same standard deviation of about 15 BB after 100 hands. As Ed Miller pointed out, winning and losing players may tend to have different standard deviations.

[ QUOTE ]
It's better to think about hands; I consider that 18,000 hands of limit holdem (FR, SH, HU) is enough to prove anyone a loser.

[/ QUOTE ]
What, precisely, do you mean by that? A player who loses a small amount on average, say 1 BB/100, can easily be ahead after 18,000 hands.

[Blarg]

[ QUOTE ]
I consider that 18,000 hands of limit holdem (FR, SH, HU) is enough to prove anyone a loser.



[/ QUOTE ]

I may yet prove you right, but I don't think I have yet, unless you mean all at one limit. If so, I hope the screen door doesn't hit my brain on the way out.

Anyway, most of my hands are at $1/2. I have about 40,000. I'm about a 1 BB/hr winner. Because my win rate is so low, and even 40,000 hands is such a small sample, I've had an incredible number of times where a day or two could take me from winner to loser. My standard deviation wasn't particularly high, like 11 BB, but I play at Party, and I have to admit that crazy swings seem like more and more the norm to me.

So anyway, I've been a loser once I had 20,000 hands too many times to count according to Pokertracker, yet I'm actually quite a bit ahead from when I got it and averaging roughly 1 BB/100. Not a sterling win rate, but my suckage isn't 100% total. I don't think 18,000 is a big enough number to determine loser or winner.

I keep in mind too Puggy Pearson's legendary 2-year losing streak. He's one of the greats, and if he can do it, terrible streaks can and will happen to anybody and everybody, good players, bad, or indifferent(like me!).