Multitabling and variance

[Vice]

I have seen posts that some of the more experienced players here only 3 or 4 table. Why don`t you 8-9 or 10 table?

I used to 7,8 table but hit a downswing, got upset and decided to 4 tables in an effort to reduce the variance. BUT I was recently struck by the idea that variance may have nothing to do with playing 3 or 10 tables at the same time, if you pay the same attention to the tables while doing either. I mean most of the decision I make are pretty automatic, and I get reads from a program called Tournament Indicator (I don`t really pay strong attention to betting patterns, etc.).

I would also be very thankful if you share your thoughts on this post: http://forumserver.twoplustwo.com/sh...ad.php?t=73786

[karnouk]

If you ROI is huge, you have a big edge and you are less likely to get huge downswings.
If your ROI becomes marginally positive, then I believe the swings and variance can hit very hard.

Multitabling usually decreases the ROI you would have if only playing on one table. So if your know your ROI is big, then keep multitabling. If you are not sure, it might be worth playing less tables and see if something changes. It never hurts to pay attention and see betting patterns

the 2 cents of a novice

[Vice]

Your logic is incorrect or at least not practical. When you say, "if you know your ROI is big", you must consider the fact that in order to know my ROI, I must have played at least 5k sngs. Does that mean that I shouldn`t multitable if I have played less then 5k sngs? No. I think one should multitable if he or she thinks he or she can still make the right (+$EV) decisions even while 8tabling, for example.

I also want to discuss this one:
If you ROI is huge, you have a big edge and you are less likely to get huge downswings.
If your ROI becomes marginally positive, then I believe the swings and variance can hit very hard.

Do we assume that in this case ROI directly relates to the skills of the player, e.g. the better he is, the bigger his ROI is? Hence, the better you are, the less swings you are likely to have?

[karnouk]

Does that mean that I shouldn`t multitable if I have played less then 5k sngs?

No, but I think you should be ready to DECREASE your number of tables.

[bjsaust]

if you pay the same attention to the tables while doing either

Thats pretty much the rub right there. IF you can pay the same attention to 8 tables as you can to 4, then you should be playing 8 tables. Plain and simple. For the vast majority of people though, they cant. It may only be a slight dropoff in attention for some, or a much higher dropoff in attention for others, but usually there is a drop.

Now, if you can make an 8% ROI playing 4 tables, and a 6% ROI playing 8 tables, then you should play 8 tables. Your ROI is lower, but your cash profit over time is higher. If you make 8% ROI playing 4 tables and 3% ROI playing 8 tables you should stick to 4, because whilst ROI is the measurement we use, its cash over time that affects our profitability.

Now your point about nbr of hands/etc. I think 5k sngs is a silly mark to have, but fair enough go with that. No, you shouldnt wait till you've played 5k sngs. Anyone whos honest with themselves and their playing should at least have an idea of how its affecting them. If people step up their nbr of tables and notice they're making mistakes they normally wouldnt, then its having an affect on their ROI, and they need to assess whether they believe its worth it or not.

[badgers]

I have seen posts that some of the more experienced players here only 3 or 4 table. Why don`t you 15-20 table?

Fixed

[Vice]

Ok, badgers, excuse my ignorance, but why don`t you multitable with >6 tables (I`m making some assupmtions here, so again, excuse my ignorance)?

[badgers]

Ok, badgers, excuse my ignorance, but why don`t you multitable with >6 tables (I`m making some assupmtions here, so again, excuse my ignorance)?

I do, I play 15 and have played 20 in the past, hence the fix. I also don't get why some of the winning players here don't add waayyy more tables, unless winning money isn't their primary objective.

[FlyingSaucy]

The question of multi-tabling has two facets. First is figuring out the right balance of quantity / quality that will maximize your earning rate for the session. Second is how multitabling affects your ability to grow as a player in the long run and move up in stakes where skill is higher and fish are fewer.

Note neither of them have anything to do with variance, because each game is an independent event.

A lot of good players are more concerned with the second facet than the first because becoming a better player gives leverage to your earning rate down the road. It is an investment in your future earning rate, whereas the first facet of just looking at "what's the most I can make right now?" is a bit short sighted in my opinion.

I like to find a balance between spending some of my sessions just playing to maximize profit and build roll (when I play 5-8 tables), and other sessions really focusing on quality, learning, reads, playing the players. You can't do that multitabling more then 3 in my opinion, ideally 2 or 1.

[bjsaust]

I thought badgers was jokingly making a point I was going to.

If you can play 8 tables as well as you can play 4, then why not 16? Why not 24? Eventually people reach a point where they cant maintain concentration enough.

I'm a winning player, although not in the league of some of these other folks, but I dont use any kind of HU display yet. My success depends a LOT on my individual reads from watching players and reviewing HHs during games.

I can play 2 as well as 1. I can play 3 basically the same as 2, but once I get to 4 or above my game tends to suffer. I probably should work on making the move to using PA HUD and relying on it, and making my moves a lot more automatic. Even still I carefully think through a lot of my moves and a lot of my hands.

[JeffreyGB]

I think FS hit it on the nose. Player development is a very big thing.

Plus, I don't care who you are, you can't watch 15 tables and pick up decent reads. That means, in the end, you're leaving money on the table. The information you can get from a HUD is not a read. It's a hint of one, but it's not nearly enough to base decisions on the same way you can a full read. Just yesterday I was going over a game with one of my students in which he clearly could have finished at least 2nd (and likely won), if only he had noticed some specific patterns that his opponents were falling into.

[Trikflow77]

I think FS hit it on the nose. Player development is a very big thing.

Plus, I don't care who you are, you can't watch 15 tables and pick up decent reads. That means, in the end, you're leaving money on the table. The information you can get from a HUD is not a read. It's a hint of one, but it's not nearly enough to base decisions on the same way you can a full read. Just yesterday I was going over a game with one of my students in which he clearly could have finished at least 2nd (and likely won), if only he had noticed some specific patterns that his opponents were falling into.

you really dont need reads AT ALL in sngs to churn out a 5% roi. With a hud you have enough info to put players on ranges and thats about all you need. Hence the reason most good midstakes sng players play 12+ tables.

[JeffreyGB]

you really dont need reads AT ALL in sngs to churn out a 5% roi. With a hud you have enough info to put players on ranges and thats about all you need. Hence the reason most good midstakes sng players play 12+ tables.

True, but 5*12 = 60, while 25*3 = 75. Higher ROI FTW!

[taipan168]

True, but 5*12 = 60, while 25*3 = 75. Higher ROI FTW!

25% ROI just isn't sustainable over the long run anymore, except for possibly at the $5.50 and maybe $11 normals.

Multitabling is all about trading off ROI for $/hour. For those players who are pro or semi-pro, $/hour is what matters not absolute ROI. That's not to say that if you play few tables (like me) you can't exploit the 10+ tablers' weaknesses, inattention and aversion to playing postflop.

[JeffreyGB]

True, but 5*12 = 60, while 25*3 = 75. Higher ROI FTW!

25% ROI just isn't sustainable over the long run anymore, except for possibly at the $5.50 and maybe $11 normals.

We keep going back and forth on this. I even adjusted my number because of our last conversation

Still, the games I've played and the ones I've seen from my students haven't impressed me as being that much tougher than they were when I played more. Yes, all of those together is still only a sample size of ~50, but it just seems weird to me for me to see only the beatable games...

[FlyingSaucy]

I am with JGB on this. 20-30 ROI is sustainable in the long run on single table sngs.

[FlyingSaucy]

I should say, at $5- $30 levels

[taipan168]

I am with JGB on this. 20-30 ROI is sustainable in the long run on single table sngs.

Depends on the buyin. I agree that at $5.50 and $11 normals it is probably sustainable over the long run, at any buyin above that or at the turbos (maybe other than the $3.40s) it most likely isn't. By long run I mean a large sample size like 2000+ SNGs at any single level.

I've played almost 1000 $27 tourneys and my ROI varies between 8% and 12%. Now I am not for a second suggesting that I play every hand and tourney perfectly, like everybody else I have my leaks, but I can't imagine that the gap between my current level of play and "perfect" play (which is in any case unattainable) would more than double my ROI.

[JeffreyGB]

I am with JGB on this. 20-30 ROI is sustainable in the long run on single table sngs.

Depends on the buyin. I agree that at $5.50 and $11 normals it is probably sustainable over the long run, at any buyin above that or at the turbos (maybe other than the $3.40s) it most likely isn't. By long run I mean a large sample size like 2000+ SNGs at any single level.

I've played almost 1000 $27 tourneys and my ROI varies between 8% and 12%. Now I am not for a second suggesting that I play every hand and tourney perfectly, like everybody else I have my leaks, but I can't imagine that the gap between my current level of play and "perfect" play (which is in any case unattainable) would more than double my ROI.

To me, a 12% ROI at turbos is just as impressive as a 25% ROI at normal speed games. Quite frankly, many people just suck at normal speed games, and the slower levels give you plenty of time to find out how to exploit them and to do so. I don't think for a second I could maintain even 15% at turbos. I guess we're just in different worlds of the same game, huh?

[taipan168]

Quite frankly, many people just suck at normal speed games, and the slower levels give you plenty of time to find out how to exploit them and to do so. I don't think for a second I could maintain even 15% at turbos. I guess we're just in different worlds of the same game, huh?

Yeah one good thing about the normals is that my guess is that there are way less multitabling, perfect push/fold, 2+2 type players on them than in the turbos. Worth thinking about when considering game selection I guess...

Anyway, a discussion on long term sustainable ROI is most probably pointless because it's so difficult to accumulate a meaningful sample size over a short enough timespan to minimise the effects of changes in the game (eg. UIGEA) to prove it.

[bjsaust]

Whats a number of SnGs to have in a row for a fair sample for ROI purposes? I know people like huge numbers for 'true' ROI, but whats acceptable for a running measurement?

[taipan168]

Whats a number of SnGs to have in a row for a fair sample for ROI purposes? I know people like huge numbers for 'true' ROI, but whats acceptable for a running measurement?

The 2+2 crowd say you need 500 SNGs to show that you're a winning player, and 1000+ to have an idea of what your ROI is. Of course, the more tourneys you have as a sample size, the smaller the deviation between your "true" ROI and your observed ROI.

[JeffreyGB]

Whats a number of SnGs to have in a row for a fair sample for ROI purposes? I know people like huge numbers for 'true' ROI, but whats acceptable for a running measurement?

The 2+2 crowd say you need 500 SNGs to show that you're a winning player, and 1000+ to have an idea of what your ROI is. Of course, the more tourneys you have as a sample size, the smaller the deviation between your "true" ROI and your observed ROI.

Format has an impact too - you need more turbos than you do normal speed games, to have the same level of accuracy.

[Trikflow77]

Whats a number of SnGs to have in a row for a fair sample for ROI purposes? I know people like huge numbers for 'true' ROI, but whats acceptable for a running measurement?

The 2+2 crowd say you need 500 SNGs to show that you're a winning player, and 1000+ to have an idea of what your ROI is. Of course, the more tourneys you have as a sample size, the smaller the deviation between your "true" ROI and your observed ROI.

with reg sngs this is prob correct. With turbos i would have to disagree, i have had +/- 9% differences in my r0i (in blocks of 1000)in the 27s when i played them. I had 1k games with a 22% roi and then the next 1k was 13%. This was a few years ago and the few turbos i have played since there is NO way to have an roi over 15% over the long run. honestly, it is prob 10-12% with the level of play now. the 27s+ play like the 114s and 225s did 2 years ago.

On another note, i would prop bet any amount of money that a 25%+ roi is sustainable over 1-2k 9 man games at a level over 50 dollars. The level of play is just 2 good, and you cant beat the math if your playing with 6-7 other players that understand icm.

[drmcboy]

45 mans ftw