Aaron Brown

Six Degrees of Idiocy

ne of the classic works of poker, and risk management, is Herbert Yardleys 1957 best-seller, The Education of a Poker Player, Including Where and How One Learns to Win. Yardley is an important transitional figure. 19th century poker was forged in selforganized frontier societies such as mining camps, farm settlements and transshipment points. During the first half of the 20th century, it evolved into a game of extraordinary mathematical and psychological subtlety. Yardley learned his poker from a genuine old west gambler with deep 19th century roots. He later applied his talents to codebreaking, running the US efforts during World War I and after. That experience imbues his poker analysis with 20th century applied mathematics and proto-game theory. David Kahn wrote an excellent biography last year, The Reader of Gentlemens Mail: Herbert O. Yardley and the Birth of American Codebreaking. The title refers to an incident in 1929. Secretary of State Henry Stimson shut down Yardleys codebreaking operation with comment, Gentlemen do not read each others mail. Yardley responded by writing a 1931 bestseller The Black Chamber about his career, which led to accusations of treason for revealing secrets. Kahn concludes, however, that Yardley was a rotter, not a traitor. Yardley spent his remaining 28 years working as a spy, freelance cryptographer, real estate developer, restaurant owner and anything else that

caught his fancy in between writing bestsellers and playing poker.

Complete idiots, degrees and parity

Someone one degree removed from a complete idiot will do the opposite. She will think how other people will interpret her actions if they assume she is a complete idiot. When she gets a good hand she will grimace and pretend to be pained while putting money in the pot. With a bad hand, shell laugh and ostentatiously count her chips as if determining how much money she can make. There arent many complete idiots, at least not ones with enough money left to buy in to a poker game. Most players are one degree

14

Wilmott magazine

removed. If someone is acting strong, theyre The simplest way for an investor to be two other people are thinking at the same level you most likely weak, and vice versa. This rule is pretdegrees removed from a complete idiot is to buy a are. Second-degree solutions are equilibrium ty reliable outside of poker as well. low cost, well-diversified index fund. There are solutions. Moving to the third or higher degree, Making one more reversal, so you act strong more sophisticated second-degree strategies like in poker as well as finance, requires considerawhen youre strong hoping other people will Black-Litterman or Markowitz optimization that tion of non-equilibrium factors in other people, think youre weak but acting strong to fool them attempt to integrate a variety of factors in a consuch as learning. into thinking youre strong, is not two degrees sistent theoretical framework. This is the lowest You gotta know when to hold em removed from a complete idiot. That just degree at which we are practicing finance. Lets see if we can apply these concepts to the changes the parity of your idiocy, not its degree. A complete idiot does not consider other peopoker game of Texas hold em in a way that However many flip-flops you do, you are still givple, someone one degree removed from a commakes sense in finance. In this poker game, playing away complete information about your hand. plete idiot considers other people, but assumes Someone two degrees removed from a complete idiot understands that other people are trying to decode his reactions. This is the degree at which real poker begins. Zero degree (honesty) and first degree (reflexive deception) are natural human instincts, common to many games and non-game situations. The second degree requires thought. One way to be two degrees removed from a complete idiot is to maintain a poker face, to act the same with strong hands and weak. Another way favored by game theorists is to randomize your play. In either case, you give away nothing ers are each dealt two cards face down. These are by your actions. those other people are all complete idiots. To be known as hole or pocket cards. There is a round This idea translates naturally to finance. The two degrees removed from a complete idiot you of betting, after which five community cards are complete idiot always goes for the money: the not only have to consider the reactions of comdealt face up (with some intervening betting bond with the highest yield, the mutual fund plete idiots and one-degree thinkers, you have to which we will ignore). Of the players remaining with the best historical return, the stock of the make the conceptual leap that at least some the pot until showdown, the one who best company. Of course, Im using can make the best five-card poker complete idiot in the technical FIGURE 1: WHICH HANDS ARE WORTH PLAYING? hand out of the seven available to her sense. These people are not necessarily (her two pocket cards plus the five stupid. Grabbing money as fast as you $3.00 community cards) wins the pot. You can is sometimes a good strategy. In do not need a detailed understanding other cases, these people are victims $2.50 of poker betting or hand rankings for of unscrupulous financial marketers. this article. The point is they look only at the most $2.00 The most important decision in obvious criterion and dont ask themlimit hold em is which starting selves why the investment looks so $1.50 hands to play. A pair of Aces is the good. strongest possible starting hand, it A person one degree removed from $1.00 will be the strongest hand 31 per cent a complete idiot does just the reverse. of the time when playing against nine He looks for the bond that trades flat, $0.50 other players. Seven/Two of different the mutual fund with a one-star suits is the weakest possible starting Morningstar rating, the stock in the $0.00 hand against nine other players, it company in headlines for incompewill win less than 5 per cent of the tence and scandal. He reasons that ($0.50) ($0.50) $0.00 $0.50 $1.00 $1.50 $2.00 $2.50 time (against only one other player, everyone else shuns these investHand Equity if Everyone Stays Until Showdown Three/Two of different suits is a weakments, thus they are underpriced.

This idea translates naturally to finance. The complete idiot always goes for the money: the bond with the highest yield, the mutual fund with the best historical return, the stock of the best company

Actual Hand Equity

Wilmott magazine

15

AARON BROWN

The quickest way to improve the game of a bad or average player is to have him play fewer hands. Moreover, the disciplined complete idiot strategy is quite effective against more optimistic complete idiots
er hand). Clearly there is an advantage to starting with stronger cards, but the question is exactly which hands are worth playing and which should be folded at the start? The graph in Figure 1 offers some helpful information for analyzing this question. Each of the points are one of the 169 possible starting hands (13 pairs, 78 non-pair combinations of the same suit and 78 non-pair combinations of different suits). The X-axis shows the value of the hand if every player bets $1 and stays in until showdown. For example, the hand at the upper right is Ace/Ace. Under the assumptions, it will win $9 31 per cent of the time and lose $1 69 per cent of the time for an expected value of $2.10. The value on the Y-axis is the actual average winning over a very large sample of hands played at the online poker site [Link] (scaled to a constant $1 betting limit). The solid line is a quadratic least squares fit to the data. The first point is that actual poker winnings go up with the square of hand strength. At the weak end, this is because you lose no more folding the weakest possible hand as an average hand. Notice that average hands, with $0 equity if everyone stays until showdown, have negative actual expectation. At the strong end, this is because you lose the most when you have the second strongest hand at the table, so slightly better hands are worth considerably more money. The complete idiot will look at this and decide to play only the hands with positive actual equity. Those hands are any pair 8 or higher, any two cards Ten or above of the same suit except Queen/Ten and Jack/Ten; and if the suits do not match, Ace/King, Ace/Queen, Ace/Jack and King/Queen. These represent 122 of the 1,326 possible sets of pocket cards, meaning you will play in 9 per cent of the hands and fold immediately the other 91 per cent of the time. This is the advice you will find in most beginners poker books, although it might differ slightly in the precise hand recommendations. Also, the advice typically allows you to play more hands if you act later in the hand, because the set of other players is effectively reduced. Why would a poker book recommend playing like a complete idiot? The advice is only strategically idiotic; it is based on sound statistical principles. An informed idiot, who has a terrible strategy but calibrates and executes it properly, is much better than the average player. The quickest way to improve the game of a bad or average player is to have him play fewer hands. Moreover, the disciplined complete idiot strategy is quite effective against more optimistic complete idiots. Someone one degree removed from a complete idiot assumes everyone else acts like a complete idiot. That makes most of the hands above unplayable. Unsuited Ace/Jack, for example, is the weakest of the hands. If you play it, you know everyone else in the pot has better cards. Moreover, suppose three players enter the pot with Ace/King, Ace/Queen and King/Queen, while you have a humble hand like Five/Four of different suits. You are the favorite to win. You have the same chance of pairing your cards as the other players do, but they will compete with each other for the top cards while you will enjoy the low cards by yourself. Therefore, the player one degree removed from a complete idiot will look for hands that play well against the top hands. These are pairs and suited connectors (two cards of the same suit and adjacent ranks) of Jack or below. Of course, shell play higher pairs and suited connectors as well, because these are so strong theyre good even against similar hands, but shell throw away unsuited and unpaired hands regardless of how high the cards are. This can work well against disciplined complete idiots, especially if combined with good betting after the flop (the first three community cards, dealt all at once). However, its not a good strategy in general, since you will rarely encounter a table full of disciplined complete idiots. If that were common, the hands you play would have positive equity. Most poker games contain either undisciplined complete idiots, who play both kinds of hands and more, or good players, or both. In any case, its not poker. At a minimum, in my opinion, true poker requires considering what happens if other people reason the same way you do. That opens the door to many complex and subtle strategies. In this article, I will consider only how a financial person could use the hand equity information to come up with a workable second-degree strategy.

Poker in the second degree

16

Wilmott magazine

FIGURE 2: CLOSE UP OF FIGURE 1

($0.06) ($0.08) ($0.10) ($0.12) ($0.14) ($0.16) ($0.18) ($0.60) ($0.40) ($0.20) $0.00 $0.20 $0.40 Hand Equity if Everyone Stays Until Showdown $0.60

cards are high enough to beat a lot of other players smaller card. But why do these onegap hands play better than connectors (Seven/Six, Eight/Seven and Nine/Eight)? Because the chance of getting a straight is not that much less, but the straight will be far more unexpected, and thus will make you more money. Other hands that play better than their strength are medium two gapers (like Nine/Six), small pairs (below Eights)

time and would be near break-even if everyone stayed in until showdown. It doesnt make sense that an average hand could have less equity than a weak hand, after all you can fold either one with the same loss. Those average hands with the most negative equity are all either Ace combined with a small card of another suit, or two cards of the same suit too far apart to form a straight and without an Ace or King. Why are these hands so bad? They win about 10 per cent of the time if everyone stays until showdown but they play very badly against the hands other people stay in with. Either no one else will stay in, and the player will not win much, or a player will stay in with a superior hand. Unsuited Ace/Five, for example, is an aboveaverage hand. But if someone has an Ace and any card higher than a Five, Ace/Five is a strong underdog. It wins only 25 per cent of the time against Ace/King, for example, or 30 per cent of the time against a pair of Eights. On the other hand, we see some hands that are play much better than their strength would indicate. Among below-average hands, the three biggest overachievers are unsuited Seven/Five, Eight/Six and Nine/Seven. One reason these hands play well is that if you get a pair, you are not likely to share it with anyone else. Also, these

and King/small. As a second-degree player, we know we have to play the strongest hands, but also mix them up with some weaker hands. Otherwise we give away

against good players, but its enough insight to be well above average. But you cant plan your play based on large population averages, you have to figure out the individuals you are playing. If they are undisciplined complete idiots, its enough to play only your strongest hands. If you are playing disciplined complete idiots, its enough to play the hands that do well against the strongest hands. You have to recognize these players and adapt to them. If anyone at the table one degree removed from a complete idiot, you should note the type of weaker hands she favors, and play accordingly. That means favoring hands that play well against her favorites when youre playing her, and playing hands that will surprise people when youre playing others. Even pretty good players have habits, some like small pairs, some like suited connectors, some like suited high card/low card combinations. Remember, youre not looking for the hands that play best against the table as a whole, but the hands that play best against the players who happen to be in this hand. There are other second-degree poker strate-

This is the one with the strongest parallel to finance We can be greedy, and play the strongest hands, but we avoid the hands just a little bit worse. These hands are overplayed and have negative expectation
too much information, the other players will know exactly what to do against us. Since other players are also playing some weaker hands, we dont select only the weaker hands that play well against the strongest hands. Instead we pick the weaker hands that are most surprising, but still give us a good chance of winning. We dont pick the absolute weakest hands, like unsuited Seven/Two, but we also avoid the obvious weaker hands, like small suited connectors and unsuited Ace/small. This is not enough to play winning poker gies, but this is the one with the strongest parallel to finance. We can be greedy, and play the strongest hands, but we avoid the hands just a little bit worse. These hands are overplayed and have negative expectation, even though they win more than 10 per cent of the time in a table of ten players. Unless were playing undisciplined complete idiots, well mix in some weaker hands. Well choose those hands carefully to balance some degree of strength (two middle cards, one high card or a pair) with good playing qualities against the likely other hands, and surprise value.

Wilmott magazine

17