![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Wall Street
Aug. 22nd, 2006 11:46 amI received Aaron Brown's book "The Poker Face of Wall Street" this morning. Looks interesting. More later.
Mauboussin, meanwhile, makes an interesting point on the difference between "risk" and "uncertainty". It's a line which has been covered before, but it's interesting how Sklansky and Miller blur the line between the two.
Briefly, "risk" is where an outcome is unknown, but the chance is not. When the house spins the roulette wheel, it does not know whether it will win or lose on that particular spin, but it does know that as the number of spins increases, the closer will its win percentage approximate to the theoretical win percentage.
In poker, an analogy is when you have one card to come, you are facing an all-in bet, and you are getting odds of 6-to-1. You need a club to win the hand. here your bet is risk. You know that it has positive EV, even though you are more likely than not to lose an individual hand.
However, Sklansky throws "uncertainty" into the mix as if it is risk. (Taleb, it must be said, makes a similar error when advising against checking short-term movements in prices).
For example, Sklansky will combine a 22% chance of you hitting your hand with a 30% chance that your opponent will pay you off if you hit. One of these is risk, while the other is uncertainty. If there is really a 0% chance that your opponent will pay you off if you hit, then all your carefull calculations (you have to see the Miller Sklansky NL book to appreciate some of these) are so much dogshit. Garbage in, garbage out.
For this reason, any line that begins "suppose you know that your win rate is 1.5 big bets an hour..." should be treated with the contempt that it deserves, because you never know this. Even if I have played 300,000 hands and have an average of 1.5 big bets an hour, there is a chance that the game will have got tougher when I sit down today. it's a small chance, I know, but it is still an area of uncertainty rather than risk. Short-term movements in profit rates (or stock prices) cannot be "ignored" as random noise, because there is a finite chance that the short-term movement is not random noise at all.
One of the big differences between NL and limit is that many more of your calculations in No Limit are based on uncertainty than on risk. Suppose I have $200 in front of me and the blinds are $1 - $2. Another guy with $200 raises to $10. I am on the button and look down and see a pair of 10s.
Now, we have two areas of risk - the chance of me hitting a set and the chance that either the small blind or big blind has a monster.
But we have two more (at least) areas of uncertainty. The first is the chance that the initial raiser has a pair bigger than mine. If I know his normal range of raising hands, I can assign a probability, but it is no more than that, an assigned probability.
And, bigger in uncertainty land, is the chance of my opponent calling me for various sums of money. Miller and Sklansky produce a useful graph on this, whereby the chance of an opponent calling decreases fairly drastically at a certain level. But this graph is player dependent, as all experienced NL players will tell you. Some players WILL call you down with TPTK (this 100% assignement of a chance is another error of most poker players), whereas others will ALWAYS fold if you represent a flopped set (once again, the 0% assigned chance is usually inaccurate, and poker players will, if challenged, often say "well, as good as").
Poker players have to assign "uncertainty" chances. But what they have to be careful about is giving those chances as strong a weight as "risk" chances. You have to add to the "uncertainty" chance my favourite percentage - "What is the chance that my assessment here is totally wrong?"
Mauboussin, meanwhile, makes an interesting point on the difference between "risk" and "uncertainty". It's a line which has been covered before, but it's interesting how Sklansky and Miller blur the line between the two.
Briefly, "risk" is where an outcome is unknown, but the chance is not. When the house spins the roulette wheel, it does not know whether it will win or lose on that particular spin, but it does know that as the number of spins increases, the closer will its win percentage approximate to the theoretical win percentage.
In poker, an analogy is when you have one card to come, you are facing an all-in bet, and you are getting odds of 6-to-1. You need a club to win the hand. here your bet is risk. You know that it has positive EV, even though you are more likely than not to lose an individual hand.
However, Sklansky throws "uncertainty" into the mix as if it is risk. (Taleb, it must be said, makes a similar error when advising against checking short-term movements in prices).
For example, Sklansky will combine a 22% chance of you hitting your hand with a 30% chance that your opponent will pay you off if you hit. One of these is risk, while the other is uncertainty. If there is really a 0% chance that your opponent will pay you off if you hit, then all your carefull calculations (you have to see the Miller Sklansky NL book to appreciate some of these) are so much dogshit. Garbage in, garbage out.
For this reason, any line that begins "suppose you know that your win rate is 1.5 big bets an hour..." should be treated with the contempt that it deserves, because you never know this. Even if I have played 300,000 hands and have an average of 1.5 big bets an hour, there is a chance that the game will have got tougher when I sit down today. it's a small chance, I know, but it is still an area of uncertainty rather than risk. Short-term movements in profit rates (or stock prices) cannot be "ignored" as random noise, because there is a finite chance that the short-term movement is not random noise at all.
One of the big differences between NL and limit is that many more of your calculations in No Limit are based on uncertainty than on risk. Suppose I have $200 in front of me and the blinds are $1 - $2. Another guy with $200 raises to $10. I am on the button and look down and see a pair of 10s.
Now, we have two areas of risk - the chance of me hitting a set and the chance that either the small blind or big blind has a monster.
But we have two more (at least) areas of uncertainty. The first is the chance that the initial raiser has a pair bigger than mine. If I know his normal range of raising hands, I can assign a probability, but it is no more than that, an assigned probability.
And, bigger in uncertainty land, is the chance of my opponent calling me for various sums of money. Miller and Sklansky produce a useful graph on this, whereby the chance of an opponent calling decreases fairly drastically at a certain level. But this graph is player dependent, as all experienced NL players will tell you. Some players WILL call you down with TPTK (this 100% assignement of a chance is another error of most poker players), whereas others will ALWAYS fold if you represent a flopped set (once again, the 0% assigned chance is usually inaccurate, and poker players will, if challenged, often say "well, as good as").
Poker players have to assign "uncertainty" chances. But what they have to be careful about is giving those chances as strong a weight as "risk" chances. You have to add to the "uncertainty" chance my favourite percentage - "What is the chance that my assessment here is totally wrong?"
