Bankroll Theory

[RiverMise]

Ok. The 300BB rule generally states that you need 300 big bets to play the given stakes. What if I want to play 4 tables of those stakes. I assume the 300BB thing is per tables? However, I dont think you need 1200BB to play 4 tables because if you are better than most (or all) of your opponents then your overall variance will be lower with 4 tables. This is because you are playing 4x as many hands. SO, I would tend to argue that since the variance for 4 tables is lower than a single table... maybe even 300BB (or slightly above) for 4 tables would be OK?

Is my logic flawed?

[chezlaw]

I think that's about right.

Your variance per 100 hands stays the same but as your playing more hands per hour your variance per hour goes down. Your bankroll requirement stays the same.

Thats assuming you're play isn't degraded by playing 4 tables at once.

[fnord_too]

I personally think that if 300BB is right for one table, the number that is right for 4 tables should be higher, maybe 350 or 400. Here is why:

When you multi-table, you miss a lot of information. You just cannot take everything in if you are shifting focus that much. Also, your decisions are probably not going to be quite as good independant of the information you are missing. This is due to time pressure and distraction of the other tables. I could even see playing too tight; since you are seeing so many hands you may not play slightly profitable ones that you would play if you were only playing one table.

All of these factors cut down on your per table EV and most add to variance (the too tight lowers variance some). So, I think the risk of hitting a skid is higher by multi tabling, since in addition to the luck factor you have single tabling you have some additional sub optimization factors going on.

I don't think these sub optimizations have to be large however. That is, I think that it is possible to sarafice very little EV multi-tabling, though it takes certain skills to MT effectively. (I personally don't like playing on more than two tables right now. Though often I will be playing vastly different games on two tables at once, say PLO on one table and limit HE on another.)

Certainly you could make some assumptions about EV and SD and compute the 95% or 99% bank roll (i.e. BR size such that chance of going broke is <5% or <1% respectively), but I would guess that it is between 350 and 400.

[fyodor]

This has been discussed several times here. Consensus is that how many tables you play is close to irrelevant. Yor are just getting more games in faster.

However when it comes to size of BR I tend to agree with those who lean towards larger. I think 500BB is much better. I accidentally put my BR at about 280BB and then went off for 180BB over a few days. Although I still had 100BB left I was no longer comfortable playing at the limits I wanted to and had to drop down. Probably my downslide was about over and 100BB was enough to start the recovery but if I went off any more it would be disasterous.

[RiverMise]

This is a good point. I was thinking that overall the bankroll requirements for 4 tables would be slightly more than 300BB, but obviously not (4*300BB). 350-400 seems about right. Thanks for your input.

[ QUOTE ]
I personally think that if 300BB is right for one table, the number that is right for 4 tables should be higher, maybe 350 or 400. Here is why:

When you multi-table, you miss a lot of information. You just cannot take everything in if you are shifting focus that much. Also, your decisions are probably not going to be quite as good independant of the information you are missing. This is due to time pressure and distraction of the other tables. I could even see playing too tight; since you are seeing so many hands you may not play slightly profitable ones that you would play if you were only playing one table.

All of these factors cut down on your per table EV and most add to variance (the too tight lowers variance some). So, I think the risk of hitting a skid is higher by multi tabling, since in addition to the luck factor you have single tabling you have some additional sub optimization factors going on.

I don't think these sub optimizations have to be large however. That is, I think that it is possible to sarafice very little EV multi-tabling, though it takes certain skills to MT effectively. (I personally don't like playing on more than two tables right now. Though often I will be playing vastly different games on two tables at once, say PLO on one table and limit HE on another.)

Certainly you could make some assumptions about EV and SD and compute the 95% or 99% bank roll (i.e. BR size such that chance of going broke is <5% or <1% respectively), but I would guess that it is between 350 and 400.

[/ QUOTE ]

[Peter_rus]

I try to keep bankroll of 1000BB. Maybe it's too rock but now i live from poker and don't want to be warry at all about lack of bank.

[Moozh]

Mathamaticlly, the reason to have a large bankroll is to handle the swings. Of course, your variance is directly related to how many hands you play (not how many hands/hour you are playing). Thus, theoretically, multitabling does not affect your bankroll requirements at all.

Intuitively, you may initially think that you would need a larger bankroll because now that you're playing two tables instead of one (for example), you're liable to lose twice as much money. But, the second table you add will often buffer your swing so instead of losing a bunch at one table, you'll lose at one and win at the other. Thus, your swings are reduced.

Yet another way to intuitively think about it. Do your bankroll requirements change depending on how many spereate sessions you play? Not at all. Yet, playing two tables instead of one is equivalent to playing two seperate single table sessions of poker. When you play a one hour two-table session, it will look exactly like two seperate one-hour sessions after you're done.

That said, the one reason you would want to increase your bankroll when you start to multitable would be because multitabling will reduce your EV. With a lower EV, you'll be more liable to have larger downswings, even if you're still a winning player. Make sense?

[naphand]

Peter is right. The only consideration you need to make is SD and EV. Playing more tables will not affect your variance directly, but it may reduce your EV due to the reasons people have mentioned.

Bankroll, not surprisingly, has been much discussed here already, including the correct formulae for calculating risk level, bankroll, length of losing run etc. A simple search under the following for the last 6 months on the SH/General forum should give you what you need;

BANKROLL
RISK OF RUIN
STANDARD ERROR

[Guy McSucker]

[ QUOTE ]

Bankroll, not surprisingly, has been much discussed here already, including the correct formulae for calculating risk level, bankroll, length of losing run etc. A simple search under the following for the last 6 months on the SH/General forum should give you what you need;


[/ QUOTE ]

For the search-lazy: Sileo's formula is

BR = -0.5 * ln(risk of ruin) * (stdv dev)^2 / win rate.

Thanks to Bozeman for this post (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Number=565873&page=&view=&sb =5&o=&vc=1).


For a 5% risk of ruin (yeah, right, I wanna go broke one time in twenty) this works out at

1.5 * (std dev)^2 / win rate.

Big problem: you will never know your win rate: see this post (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Number=780339&page=&view=&sb =5&o=&vc=1) (by me!) for why.

Guy.

[RiverMise]

wow... thanks for the great posts.

TO summarize: playing 4 tables does not change your bankroll needs much if any.