The Long Term - will you ever get in enough live hands?

[gobbledygeek]

[not a bad beat whine]

So the other week I'm in a gigantic pot (7 ways capped+1 to the flop due to straddles, etc., blah blah blah) and my best hand on the turn is two-outted on the river (OP actually had odds to stay in hand due to incredible size of pot). No big deal, happens, on to next hand. So last night, similar thing happens. Again, no big deal, I'm not too concerned about it, I'll win my fair share of those big pots IN THE LONG RUN.

But then I began thinking about the first two outter pot. The pot was one of the biggest I've played in, it's not the type of pot the comes around all that often; I'm involved in a pot like that maybe only once or twice a month live, if that, let alone leading big time on the turn. So sure, 22 times my hand will hold up to every 1 time it doesn't so IN THE LONG RUN I kill. But am I ever going to see enough of those kind of hands (i.e. huge pots, leading hand on turn with great odds of holding up) in my live lifetime play to outrun the variance? I mean, I'm guessing I'd have to be involved in at least a few hundred of those particular hands (maybe a few thousand?) in order for the numbers to start converging to 22:1; am I really going to encounter that many situations in my lifetime of play?

I guess that basic question I have is this: Do you really think you'll be able to play enough lifetime live hands in order to make up for short term variance? Sounds stupid, but is one man's lifetime live hands ever going to be a large enough sample size?

GusingvarianceasascapegoatwhenindownswingG

[Bob T.]

How many hands do you play in a month?

[gobbledygeek]

Without looking at my records, I'd guesstimate around 3.5 times a week * 4 weeks per month * 4 hours per session * 25 hands per hour = 1400 hands/month.

[Scarmiglio]

The short answer is 'yes', but obviously the more you play and the better you table select, the faster the numbers will converge. Table selection is very key here, as some tables are practically incapable of creating monster pots.

[Bob T.]

[ QUOTE ]
Without looking at my records, I'd guesstimate around 3.5 times a week * 4 weeks per month * 4 hours per session * 25 hands per hour = 1400 hands/month.

[/ QUOTE ]

I could do some arithmatic, but I would guess that it probably takes at least 10000 hands, or six months before things are likely to start to converge. Maybe I will do some arithmatic, after my nap [img]/images/graemlins/grin.gif[/img].

[Frond]

Good question.

25 hands per hr. might be a bit on the conservative side but I hear what you are saying. I play about the same amount of hours per month take or give a few. So to round it off, let's say one is playing only 60 hours a month. In a years time that is 720 hours. 5 years=3,600 hours of play time. Is this enough to tell if we are a winning player? What do some of you consider a decent number of hands or hours played live to tell if one is a winning player? Online obv.one sees way more hands per Hr and the ability to multi table is there as well. We all have heard of well known live pros who have had down years. They are putting in way more hours than the typical player I would assume.

[gobbledygeek]

[ QUOTE ]
Table selection is very key here, as some tables are practically incapable of creating monster pots.

[/ QUOTE ]

Unfortunately, my poker room isn't big enough to table select very well. Usually only 1 or sometimes 2 2/4 tables going, plus 1 or sometimes 2 4/8 tables (yes, I know I should move up, I'll get around to that one day). If I'm there to play (which I am) then I'm basically stuck with what is there.

[gobbledygeek]

[ QUOTE ]

I could do some arithmatic, but I would guess that it probably takes at least 10000 hands, or six months before things are likely to start to converge. Maybe I will do some arithmatic, after my nap [img]/images/graemlins/grin.gif[/img].

[/ QUOTE ]

But I guess I'm more talking about some very particular situations (i.e. monster pot, huge lead on turn), situations that don't arise all that often and yet I'm thinking the results of these types of hands have a big affect on overall winrate (although perhaps I am overestimating the importance of these hands).

[Tugg]

I think about this all the time. I average only 6, 5 hr sessions per month due to 3 & 4 year old kids and a non poker understanding wife. I hope to at least double that when my kids are in school full time, but its crazy when I hear about somebody who's average is based on their last 300,000 hands. I may never get there.
I do think that if you only play live, these situations are so rare that you only have to win 1 or so a year to get slightly ahead, and you wonder if its worth the variance to even get involved.But then again these are the fun pots to play so I say If we have an edge lets go.

[roggles]

I think it is very arguable whether a thing such as a "true winrate" exists. Of course you can define it arbitrarily in various less meaningful way. "Outrunning variance" and other uses of this concept seems to not signal a super-solid understanding of statistics.

The idea is something like this: With each hand you play you can assign a profit value. This gives us a random distribution, input hands and get profit. Now, this distribution is very irregular obviously, most hands you just throw away and in hands like this one you could have won enormous amounts. Now suppose each hand has a similar distribution and that they are moderately independent (note however that this assumption is pretty laughable. Any statistician would probably laugh at this way of making a time series analysis), when we sum up the distributions and divide by the number of hands (i.e. we calculate our BB / hand) we get something that is very close to a normal distribution. The normal distribution is a very nice random distribution that is symmetric around some value, the mean profit. This mean profit is what the nerds in this forum mean by "true winrate". If we make ridiculous assumptions, this can be interpreted as the expected profit on each hand you play (obviously not true with position being important). This model does not say anything about "huge hands" like in your example. Although you could make the same argument and take your average expected profit over "huge hands" and arrive at a positive value. I do believe however that you are correct in that you will not play enough of these huge hands for the sum / number of hands to approach a normal distribution, given how irregular the distributions are in these situations. However, this does not deter from the quality of the original model I presented. Extreme values occur, just not very often.

I would say you shouldn't really worry about it. You just lost a huge pot and you're feeling bad. You should just get over it.

Another thing - you seem to think that guy was the only one drawing live. Was that really so?