16 replies to this topic

[Actuary]

I searched on "utility theory" across all forums and got zilch when looking for both terms. This concept is basic to playing poker. For those that do not know, utility theory deals with "rational expectation principle" vs "human psychology."In simplist terms it says that normal behavior is risk adverse. Thus, losing $100 is "more bad" than winning $100 "is good." Most people would not flip coins for $100 bets. In fact many people would not flip a coin for a $100 to $90 bet. I believe this governs many poor EV decisions in cash games. In order to be a top player, do you in fact have to shut-off your personal utility curve and base decisions solely on mathematics and reads?This dovetails with the idea that you have to play within your means so that you can make the correct play w/o regard to feeling the pain of losing X > thrill of winning X.

[TheIceman05]

Actuary said:

This dovetails with the idea that you have to play within your means so that you can make the correct play w/o regard to feeling the pain of losing X > thrill of winning X.

I have friends with "The Sickness" that throw thousands of dollars around on tiny tiny edges. And it doesn't bother them at all. Remember that scene of Phil Ivey and Lederer and Williamson on the golf course at the World Series last year? Tossing thousands of dollars around? My great gambling friends are just like thatIce

[econ_tim]

Maybe you should search for "risk aversion" or "expected utility." I haven't seen this come up on this forum, but I'm very familiar with it since I'm an economics graduate student.To get risk aversion, you need to have a concave utility function, e.g. U(x)=ln(x)This gives you your utility for a given level of the good x.Calculating expected utility is similar to calculating expected value, except you put the utility function in the equation.So if the expected value of fliping a coin is .5A + .5B, where A is what you get for heads and B is what you get for tails, the expected utility of the same coin flip is .5U(A) + .5U(B)For a concave utility function, the expected utility of such a gamble will be less than the utility of the expected value, i.e.U(.5A + .5B) > .5U(A) + .5U(B)This means that you would prefer a certain outcome of value X to an unceratain outcome with expected value X. This is called risk aversion.This is a standard concept in economics, put people behave in non risk averse ways all the time, such as when they buy lottery tickets. There are several theories to try and explain this kind of behavior, but I've probably gone on long enough.

[garamond10pt]

Actuary said:

I searched on "utility theory" across all forums and got zilch when looking for both terms. This concept is basic to playing poker. For those that do not know, utility theory deals with "rational expectation principle" vs "human psychology."In simplist terms it says that normal behavior is risk adverse. Thus, losing $100 is "more bad" than winning $100 "is good." Most people would not flip coins for $100 bets. In fact many people would not flip a coin for a $100 to $90 bet. I believe this governs many poor EV decisions in cash games. In order to be a top player, do you in fact have to shut-off your personal utility curve and base decisions solely on mathematics and reads?

I think you do. It's hard enough being a +EV player at the higher levels, but I think sacrificing EV to reduce variance is leak in a lot of player's games, and one that the better players routinely take advantage of.

[econ_tim]

Actuary said:

In simplist terms it says that normal behavior is risk adverse. Thus, losing $100 is "more bad" than winning $100 "is good." Most people would not flip coins for $100 bets. In fact many people would not flip a coin for a $100 to $90 bet.

Actualy, it doesn't say that losing $100 is more bad that winning $100 is good. What is does say are that sure outcomes are prefered to uncertain outcomes with the same expectation.There is a theory call "prospect theory" that does make the prediction you mention.

[Actuary]

as an Actuary, we had to study this for an exam. Thanks for refreshing the equation. Generally any Utility Eqation that has a positive first derivative (more $'s is better) and a negative second derivative (each $ is less valuable than the prior one) fits the risk adverse form, e.g. Ln(x).replying to your last post:Given a risk adverse utility curve, a person would not risk losing $100 in order to possibly gain $100 in a 50/50 proposition. This is what I meant by losing $100 feels worse than winning $100 feels good for "normal folk"

[Smasharoo]

In order to be a top player, do you in fact have to shut-off your personal utility curve and base decisions solely on mathematics and reads?Top players ignore dollar values completely. 10 cents a bet or $10,000 it's the same game.

[cdddc75]

Smasharoo said:

In order to be a top player, do you in fact have to shut-off your personal utility curve and base decisions solely on mathematics and reads?Top players ignore dollar values completely. 10 cents a bet or $10,000 it's the same game.

Tell that to the Donks freaking out on General over Daniel's June results to date.They don't even have a clue.

[DKE_XP120]

cdddc75 said:

Smasharoo said:

In order to be a top player, do you in fact have to shut-off your personal utility curve and base decisions solely on mathematics and reads?Top players ignore dollar values completely. 10 cents a bet or $10,000 it's the same game.

Tell that to the Donks freaking out on General over Daniel's June results to date.They don't even have a clue.

Thats why they're in general.They arent there for EV decisions, learning concepts, game theory, or poker for that matter.

[cdddc75]

DKE_XP120 said:

cdddc75 said:

Smasharoo said:

In order to be a top player, do you in fact have to shut-off your personal utility curve and base decisions solely on mathematics and reads?Top players ignore dollar values completely. 10 cents a bet or $10,000 it's the same game.

Tell that to the Donks freaking out on General over Daniel's June results to date.They don't even have a clue.

Thats why they're in general.They arent there for EV decisions, learning concepts, game theory, or poker for that matter.

I really need to stop looking over there. Thanks.

[Actuary]

newbies..to avoid the flames..bring it like this!

baybee!

[LongLiveYorke]

Utility functions that are useful are concave but non-decreasing, meaning in the long run they will approach a constant. Therefore, the "preference" for non risk that comes out of the application of Jensen's inequality becomes minimal. In other words, play within your bankroll and don't worry about it.

[yergan]

Doesn't proper bankroll management contribute to "shutting off your personal utility curve"? i.e. When you have a BR of $500, losing $100 means a lot more than winning $100, however losing $20 SHOULD be the same as winning $20.

[Sluggo]

I disagree with the idea that it's a good thing to ignore the decreasing marginal utility of money. It's a fact of life and shouldn't be ignored.

I personally would not flip a coin for a $99,000 bet against a $100,000 bet. The $500 expecation isn't worth it. For a single flip, it would be a poor choice for me to make the bet.

However, in poker, players play many thousands of hands. By repeating many, many bets (or flips), you squeeze the probability distribution thinly enough such that the utility of the expectation outweighs the differences in the marginal utility of money.

If I could make a $99,000 bet against a $100,000 bet where a coin was flipped 100 times and the money was paid out proportionally, then the $500 expectation might outweigh negative utility caused by risk.

It gets interesting when players, because of their risk averse feelings and financial situation, have different valuations of money. To play the winningnest poker possible, in the long run, one must ignore decreasing marginal utility of money. And this is only possible because the giant sample size of hands reduces the standard deviation (measured in percent) of the outcome distribution.

[GABMAD]

Actuary, where did you go to university? My brother is also an actuary, I told u that b4, but I think you never read that post, because u didn't respond. Anyway, you actuaries sound alike...

[Actuary]

Actuary, where did you go to university? My brother is also an actuary, I told u that b4, but I think you never read that post, because u didn't respond. Anyway, you actuaries sound alike...

what..u think I sit around and read all the threads and post replies right away?
Man, you must have me marked as a loser.

oh, and I did respond.


check pm.

[aim786]

what..u think I sit around and read all the threads and post replies right away?
Man, you must have me marked as a loser.

oh, and I did respond.
check pm.

I thought posting at FCP is what you get paid to do...
(sw)