Short-Handed Flucuations

[QTip]

Things probably belongs in the HUSH Forum; however, this is the forum in which I've been involved in debates on this topic, so I'm posting this here. Some of us have gone back and forth and far as whether variance is great in SS play or not.

This is a quote from Feeney from "Inside the Poker Mind" that I thought shed some light on that topic. (By the way, if you haven't read this book yet, YOU MUST GET IT!)

"Be Prepared for Some Fluctuations"

"Short-handed play does tend to magnify your fluctuations. Playing hands with small edges aggressively to the river creates unavoidable swings. You should be ready for this. The swings are partially tamed, however, by the greater overall edge you can have short-handed if you are a bit game selective, and by the larger number of hands played over a given time. As you improve in your sort-handed play, your swings should decrease as long as you remain selective of your opponents. In fact, in my experience, players fears of the fluctuations are often exaggerated. Yes, they are great, but if you play against the right opponents they can be smaller than you might guess. When your edge is very pronounced you may be surprised at how modest they are."

[chief444]

I think the variance is exagerrated in short handed play as I've said before. I do think it's greater than full games. But I think the difference seems greater than it is in reality for a couple of reasons. For one, you're playing a lot more hands/hr per table than at a full game. So obviously your variance per hour is much greater. But your variance per 100 is not that much greater, necessarily. Another reason I think it gets exaggerated is that a lot of people here play party skins and therefore go from 3/6 to 5/10 short. So they jump monetary levels and go to short handed at the same time. I know for me at least this made the swings seem greater than they were. So again, I don't disagree that the variance in terms of BB/100 is greater. But I do think the difference is exaggerated.

As far as the quote from the book...A greater edge doesn't mean your variance will change. It just mean your winrate will be higher.

[JoshuaD]

[ QUOTE ]

As far as the quote from the book...A greater edge doesn't mean your variance will change. It just mean your winrate will be higher.

[/ QUOTE ]

That's true, but intuitively it seems to change. The higher your winrate, the less likely you are to have a losing session, and when you do, they'll tend to be more shallow.

The reason people think variance changes as EV goes up is because people tend to think of variance as "variance from even", when it's actually variance from winrate.

[DMBFan23]

higher winrate = lower standard error after N hands. your mathematical variance can still be 15BB/100 hands, but IMO it's the standard error that makes us "feel" swings

[Octopus]

[ QUOTE ]
higher winrate = lower standard error after N hands.

[/ QUOTE ]

Huh? What do you mean by 'standard error'?

[CallMeIshmael]

[ QUOTE ]
Huh? What do you mean by 'standard error'?

[/ QUOTE ]

Stanard error is the standard deviation of a sample. If you take like 40 poker hands, you tend to get a smaller deviation of that data than of the true population.



For example... a population (that is ALL poker hands) may have a standard deviation of 15 BB/100.

BUT... a 20,000 hand sample doesnt have a deviation, because it is a sample. It has a standard error.

Standard error = Standard Deviation / (square root of number of hands/100)

So.. for this, standard error =

15 / 14.14 = 1.06


What DMB was saying, is that the value of:

Winrate / Standard Error is bigger, with a bigger winrate. The larger this number, the less likely it is that you have a losing session (assuming you are a winniner, and winrate is therfore positive)

[QTip]

[ QUOTE ]
The reason people think variance changes as EV goes up is because people tend to think of variance as "variance from even", when it's actually variance from winrate.

[/ QUOTE ]

I didn't grasp that. Would you hand feed me, please? Break it down a bit more...Smash it into baby food....

[chief444]

Accuracy and consistency are two different things. You can be consistently inaccurate. Make sense now? [img]/images/graemlins/grin.gif[/img]

Basically, if your win rate is somehow 10BB/100, a "downswing" for you may be a period of winning just 5 BB/100 and you really are less likely to "feel" it as DMB puts it. People notice the downswings more when they're actually in the negative direction. But it's possible to have a very high variance but never lose money even in any given session or period of time.

[QTip]

I've never been real clear on some of these terms.

I was thinking that a winrate was different than a BB/100 number...is this wrong?

[JoshuaD]

[ QUOTE ]
I was thinking that a winrate was different than a BB/100 number...is this wrong?

[/ QUOTE ]

Winrate can be anything. The important part is that it's <something> per <another something>. The word "rate" implies that structure. Miles per Hour, Big Bets per Hundred hands, all of these are examples of rates.

Big bets per hundred hand is the easiest/most common rate used, but it's not the only one. You could easily use BB/hour, BB/hand, SB/20 minutes, etc. As long as you do all your calculations in the same units, you'll get the same answers.

(It's like doing distance calculations in feet or miles. You can always convert back and forth, and it doesn't change the answer)


[ QUOTE ]
I didn't grasp that. Would you hand feed me, please? Break it down a bit more...Smash it into baby food....

[/ QUOTE ]

Let's say you're a 10BB/100 winner. If you go 1000 hands and only win 50 Big Bets, you're in a "downswing". You expected to win 100 Big Bets, but you only won 50. It doesn't "feel" like it though, because you've still won money.

Now instead, let's say you're a 2BB/100 winner. If you go 1000 hands and lose 30 Big Bets, you've had a downswing of the exact same size, except this time it "feels" worse. You expected to win 20, and you lost 30. This time however, the effect is much worse. It "feels" like a downswing this time, because you lost money.

Do you see?