Poker Science, Religion, and Fish

An old saying advises that if you give a man a fish, you'll feed him for a day, whereas if you teach a man to fish, you'll feed him for a lifetime. There may or may not be some wisdom behind this grandfatherly advice. In frontier days, it probably made a lot of literal sense. Nowadays, most of us don't consider fishing a practical way to feed ourselves (and in the interest of a varied diet, we'd also have to learn hunting, farming, gardening, and so on). As a metaphor, the saying has good points and bad points. Self-sufficiency is often a worthwhile goal. But most of us don't have the time, inclination, or (sad to say) aptitude to learn how to (for example) build a house, even though most of us want to be housed for a lifetime. It's a catchy saying, but the truth is that some problems you need to know how to solve, while for others, you really just need the fish.

Educators make a related distinction between rote learning (or memorization) and "deep" learning, or concept learning. Rote learning is probably what you did when you learned your multiplication tables or state capitals – basically mindless memorization, so that you could spit back the answer without necessarily understanding anything. Sometimes deeper understanding can emerge through rote learning. Most of us eventually figure out that multiplication tables are not completely arbitrary. And perhaps there are advantages to discovering that structure on your own. But often you just end up with a big collection of inflexible facts in your head, names like Pierre and Albany, and numbers like 63. By contrast, deep learning supposedly involves not just learning the answers, but understanding where they come from, and forming an articulated, well-integrated body of knowledge.

Although some educators and scientists like to belittle anything that smacks of rote learning, sometimes you just have to memorize stuff, and learn the principles later. I like to think I avoid rote learning like the plague. But when it comes to, for example, cooking, I really just want the answer. Sure, I'd love to really know how to cook, but there are too many things higher on my list of priorities for me to really want to know what egg whites do to cake batter. Just give me the recipe, or even better, give me the cake.

Cooking is a good example for another reason. Rote-learning approaches to problems are often described as "cookbook" approaches. Do what the recipe says, and don't ask questions. The easiest way to tell the difference between someone who knows something by rote vs. someone who actually understands it is to ask "why" questions, like "Why do you add eggs to the batter?" If the answer is along the lines of, "Because that's what the recipe said," you have cookbook-style (literally) rote learning. If the answer is something like, "To give it a more cakelike texture," you have evidence of a little deeper understanding.

This brings us to poker. Some poker writers want to be cookbook writers. They know the right answers, and their goal is to share those answers with you, not necessarily to teach you where those answers come from. They present assertions, not arguments. You are not supposed to be able to work out the next example yourself, unless you manage to intuit the underlying principles on your own. And there's nothing wrong with that, to a point. Few beginners really want to build up their poker knowledge from abstract principles, without some kind of base of concrete knowledge. But while this may (or may not) be a good pedagogical strategy, it presents at least one difficult problem. How do you know who to trust? There are all sorts of answers to this, and it's probably true that you can get some mileage out of community consensus, book reviews, and other heuristics. But the bottom line is that cookbook-style advice is hard to evaluate.

Here's an example: Suppose the issue is how to play pocket fours from one off the button in hold'em when it's folded around to you. A cookbook writer might tell you, "Raise." Or, perhaps, "Raise against opponents you think you can outplay, or if you think you can steal the blinds at least 62 percent of the time." A more thoughtful writer might write, "Raise, because even if you don't steal the blinds, chances are good that your weak opponents will check and fold anything but a monster, and your strong opponents will call you down with overcards." Note that these are both, at some level, cookbook approaches. Both writers believe they know what is best, and have told you how to play, but the second was kind enough to explain his reasoning – he gave you the "why." This opens up all kinds of terrific possibilities. First, you can actually work through the probabilities in question and see if they add up the way you expect. What assumptions does it take to produce the same answer? Second, you can try out other examples. Does it work as well with A-K or 9-2? Third, you can second-guess assumptions. Maybe you don't think raising is a good idea. With the first writer, you have an unresolvable disagreement. With the second writer, it becomes obvious that the situation he's describing doesn't come up in your game – everyone defends his blinds most of the time, and the players frequently bet when they miss the flop. Neither way is proof against misanalysis. However, the extra detail puts you in a better position both to do the right thing under variable circumstances, and to appreciate where your reasoning might have gone awry if it turns out your strategy was flawed.

Ultimately, almost all decent poker books present a mix of rote advice and detailed explanation. But if you want to be a critical consumer of poker writing, you need to be profoundly suspicious of information that comes by way of assertion rather than argument. Without the reasoning to back it up, a writer says in effect, "Trust me." It's not always a bad idea, and there are many poker writers whose advice you would do quite well to trust, at least provisionally. After all, poker books are not rigorous mathematical proofs, and there's no way to be completely thorough about every topic you might want to cover in a 200-page book – or, for that matter, a 2,000-page book. You have to pick your spots, and you can't explain poker all the way down to subatomic particles. But the best writers don't give up easily. Anyone can write that he knows from experience that J-8 shouldn't be played from up front in a ring game. And no one (as far as I know) can present ironclad proof to that effect. But the best poker writers, even if they can't convey the entirety of their experience in print, can at least convey a bit of the answer to the "why" question.

As a scientist, I tend to hold poker writers to a high standard when it comes to reasoning. One of the essential features of scientific findings is that they can be reproduced. Premises, methods, assumptions, and logic can all be debated, but a conscientious researcher with adequate resources can attempt to replicate any well-reported earlier experiment. In his "Fighting Fuzzy Thinking" articles, David Sklansky explicitly attempted to bring poker writing toward this standard. Whether or not we really agreed that his logic was unassailable, or that his premises were correct, at worst we could usually reproduce his reasoning. That's a reasonable standard to apply to reasoning about poker in general, even if we know that no one can measure up all the time.

Editor's note: The second edition of Daniel Kimberg's book Serious Poker is now available through Card Player. His new and improved poker website is now up at www.seriouspoker.com.