River Boat Epistemology

Monday, January 04, 2010

Last week, my in-laws treated my wife and me to a welcome break from the New England winter in the form of a Caribbean cruise. One of the highlights for me occurred one afternoon when I participated in a Texas Hold'em poker tournament I learned about from my wife's mother, who remembered that I used to play poker regularly during grad school around the time Mrs. Van Horn and I started dating.

Back then, my group didn't play Texas Hold'em that much, and I have since indulged my interest in the game only intermittently, watching games here and there on television and occasionally joining co-workers in informal Texas Hold'em tournaments with very small buy-ins.

Playing for a few hundred dollars and facing complete strangers were both new to me, and the buy-in was just enough so soon after starting to work again that I did have to think it over a little bit. But I wanted to know what it would be like and how I would do enough that I finally decided at the last minute to play.

Although I did not win -- only the top two players won money -- I did far better than I expected to, finishing third after eliminating four of the other ten players directly (two of them in one hand) and leading my closest opponent at one point by a factor of three. Bragging aside (not to mention probably also admitting that I don't know what to do with a huge lead), what really fascinated me were a few thoughts I had about the process of learning as a result of that game.

Poker is simple in some respects and complex in others. It is fairly straightforward to estimate the potential strength of one's hand at any given point and, at least in Texas Hold'em, to come up with reasonable estimations of what an opponent might possess.

Betting patterns can help one refine one's estimate of what an opponent might have, but the ability of an opponent to bluff has to be accounted for. (e.g., There are whole books about "tells" -- minor expressions or actions that can indicate deception -- that have been written with the object of determining whether an opponent might be bluffing.) So the rules of the game are simple, but its strategy is actually quite complex. What interests me here is the fact that both explicit knowledge and intuition are both involved at all levels of the game.

As I just indicated, there are different mixtures of explicit knowledge and gut feel for different aspects of the game and I suspect that this might vary over certain ranges by individual. For example, I am almost certain that it is possible to calculate the mathematical odds for any given hand to win. (I'd normally check on something like this and provide a link, but I'm writing this on a plane and probably won't be home until after midnight... [Actually, it ended up being 2:30 am. I woke up at six, for work beckons. Oh, and I'm less sure that this can be done, but the question is still interesting to consider. --ed]) I don't know how to perform such calculations, but I did notice myself feeling generally better or worse about my pocket cards after the deal as I played different hands even before thinking something like, "Hmmm. An ace and a seven in diamonds. Good high card and a fair chance for a flush." Clearly my subconscious was working from previous experience on some level.

Thinking about what I did right and wrong after the game, I wondered whether I might want to learn how to calculate the odds of my hand winning. Strange as it might seem, decided against doing so -- for now. At the time, I was not able to state explicitly why this seemed like a bad idea to me, but on reconsidering this point, I now think I understand what I was grappling with: The details of such calculations (and maybe the results) would distract me from learning other aspects of the game, among them figuring out how strategies generally evolve over the course of a game and how they can vary among individuals.

Among the reasons I might have run into trouble later in the game was that I realized too late that the risk of staying in a hand changes in multiple ways as the game goes on. For one thing, due to increasing initial bets it's more expensive to see the up-cards and this fact in turn seemed to make players more reluctant to stay in with the weaker hands they might have chanced earlier in the game.

On top of that, bluffing can really pay off at such a point: One of the two players left after me took advantage of this fact to bluff me successfully, one of the three most important plays in the game as far as I was concerned. (This is not to say that I got bluffed easily. I did call a bluff by the other top player early in the game after considering a fold. I decided to stay in when I realized that his bet made no sense. I remember feeling quite comfortable with that decision, although not completely sure about why. This first-hand experience of using intuition was part of what I was after when I decided to play.)

So I would say that eventually, I could improve my game by learning how to calculate odds, but that at this point, I still need to learn too many other things about the parameter space of the game before going into that much detail. Something I observed suggests that I am right: The people who talked before the tournament about their gaming on-line dropped like flies early on. I strongly suspect that this was due to their concentrating too much on the mathematical aspects of the game and too little on learning how to read their opponents: You don't see your opponents in on-line poker and so cannot observe everything they do. (Incidentally, I considered and rejected the idea of playing on-line poker for that reason alone: On-line gaming in general and on-line poker in particular do not appeal to me. That said, people who are weak at estimating how good their hands are might do well to practice on-line.)

I wouldn't call myself a serious student of poker, but if I were to study anything at this point, it would be tells. But still, it's really too early to do that in much detail, either. At this point, I'm a rank beginner and not too sure that the cards weren't a little bit kinder than average to me. What I would really need to see are more examples of play before I can gain an overall grasp of what a game of poker typically entails and where I am strongest and weakest. (The best advice for someone like me would be of a general nature that touches on all the important points of the game without dwelling on any one of them.)

In parallel with gaining experience, I could conceptualize the knowledge by making it more explicit or studying certain issues in more detail as I feel ready for them or in need of them. (This is an example of the spiral theory of learning. To get lost from the outset in details about one issue or another as the on-line players may have would be an example of what I have heard called "over-thinking," although I regard that colloquialism as misleading. Such an approach strikes me as rationalistic, but there could be any number of other reasons someone might make such an error.)

Poker appears to be a good model for certain aspects of how we learn generally. In a certain respect, I am almost glad I didn't win that tournament. I might not have learned as much as I did about how to approach poker in particular and knowledge in general.

-- CAV


: Corrected a typo.


Galileo Blogs said...

Interesting post, Gus. The fact that your intuition can be stunted if you focus too soon on studying mathematical percentages is interesting.

There is a lesson here in avoiding rationalistic thinking and I agree that it has to do with the "spiral" nature of learning.

When I am learning something new, if I am studying strictly at the abstract level, I find myself getting uncomfortable. Then I need to gain a variety of assorted concrete facts. Then I can go back to the abstractions. And so on.

Part of this process is listening to my "intuition." Intuition is important. It is similar to emotion, but different. Intuition is a judgment of facts that occurs in an automatized manner. The basis for that judgment is all of one's experiences and conclusions of prior facts. In the case of poker, those are one's experiences with many prior hands. In each of those hands, a player experiences knowledge about both the probability of that hand and the ways his opponents played their hands in response to his hand.

To ignore such intuition because one cannot explicitly identify all of the reasoning behind it is to lose the benefit of all of those experiences and learning that is held in one's sub-conscious. It may also mean getting cleaned out at the poker table.

In summary, explicit study can enhance and improve one's poker playing, but one must take care not to trump one's intuition in a cognitively reckless manner.


As a slightly separate point, cognitively intuition seems to have a similar status as emotion. Because one cannot necessarily trace out the basis for that "hunch," one has to be careful about how one relies on it. But, to dismiss the hunch simply because one does not know how it arose is to deny oneself a potentially valuable basis for acting when explicit, clear knowledge cannot be obtained.

Examples of those situations are playing poker, where there are too many variables to be analyzed in a short period of time. Another case of dealing with people in time-limiting circumstances is a potentially dangerous situation, for example encountering a stranger in a forlorn location. Following one's hunches in such situations can be life-saving.

Mike said...

Over time, I have come to realize that poker is one pursuit in which there really is no substitute for the experience that comes with playing literally millions of hands over a lifetime.

One can accelerate understanding of those hands and what sort of situational betting or bolting strategy might have served best by playing online, reading up on the poker literature (especially Sklansky & Malmuth), and varying play as much as possible -- going from a neighborhood $10 sit-and-go to a $4-$8 casino Omaha limit game, and so forth.

Essentially one wants "good" exposures to hands, and since the other players can sometimes be uncooperative at playing predictably, one mitigates by adding more conduits for information. (Darn those guys!)

There are 169 possible two-card holdings for normal Hold'em. Setting aside the issue of obvious rags and equivalent hands, a player really does build up knowledge of each hand with each flop type with each bettor type opposing, and so forth, to the Nth degree... internalizing all this so that it's there in your subconscious, ready to guide you at crunch time, as you discovered.

In short, experience makes the player, and with more exposure comes more experience, so maximize exposure. It can often make the game less fun, at least in the social/casual sense, but walking away with more and more money each time is kind of fun too! :)

Gus Van Horn said...


Your explicit analogy to emotion is much appreciated: It's a connection I wish I'd made when posting and it's the nut of the matter.

To imply that intuition alone is dependable is to make the same mistake a mystic makes regarding emotions. To discount it out of hand is also mistaken, though.


I definitely noticed a respect vs sociability divide in that tournament. I chose to be hard to read as a strategy, and did lose out on some friendly banter -- but that's what a monthy game with your coworkers or beer buddies is for...