Annie Duke retired from professional poker in 2012 as one of the most successful players of all time. Today, she is taking what she learned at the poker table about decision making and sharing it with the world as an author and public speaker. We sat down to talk about her new book How to Decide: Simple Tools for Making Better Choices and why business is dealing with poker like environments and all the uncertainty that comes with it.
Dave Knox: You retired from poker in 2012 to focus on your next career as an author and public speaker. Your first book, Thinking In Bets, has a concept that life is like poker, not chess. Can you explain that viewpoint?
Annie Duke: As I transitioned from cognitive psychology where you’re certainly thinking about learning and decision making, into poker, which is obviously a really high stakes, fast paced, decision making environment, what I kind of thought intuitively was that this is very feedback rich. You are making decisions and getting feedback right away. And obviously if someone’s losing their chips, they are going to kind of figure out that maybe they need to, over the long run, they should be able to see that feedback and then adjust their decision making accordingly and become better at the game. And while that’s true for a handful of players, it’s not true for most people who play the game. The learning environment actually turns out to be quite unkind in terms of your ability to really kind of take that feedback and use it in order to become better. And I started to really explore why that is and it has to do with this very deep difference between chess and poker
When we’re playing poker, there’s this really strong influence of uncertainty in terms of hidden information and simple luck. I can have a hand where I’m going to win the hand 82% of the time and what that means is 18% of the time I’m not going to win the hand. When you observe that I lose the hand, is it because of bad luck or is it not? And this becomes particularly difficult because at the end of the hand, the cards tend not to be revealed. It’s not like when you’re watching on TV and the camera can see the hands. As a player, mostly I don’t know what your hand is when I’m making those decisions and then you have this really strong influence of luck. And you can see that you don’t have that in chess. I can see where all the pieces are so I don’t have that problem. I know what your position is. And then the other thing is that those pieces are only going to move because of an active skill. In other words, you choose to move the piece or I choose to move my piece. There is no luck. And what that means is that because you’re removing that strong influence of uncertainty, we can take the quality of the outcome and we can say something real about the quality of the decisions that underlie those outcomes.
But we can’t do that in poker. It is such an unkind learning environment because when I lose a hand and I’m trying to figure out why, I now have this sort of escape hatch that allows me to not really confront the possibility that maybe my decision making wasn’t so good. And that escape hatch is luck, I can say,” I got unlucky.” And vice versa, when I win a hand, I can kind of downplay that and I can say that I played really well. And that really unlocked for me, sort of what’s happening is that once you introduce uncertainty into an environment that what you’re giving people the leeway to do is sort of that’s where all these cognitive biases that really frustrate our decision really can get a foothold. They can come in and wedge themselves into that uncertainty, that gap that uncertainty is creating between the relationship between decision quality and outcome quality and really wreak a lot of havoc in our decision making.
That is just what most of life’s decisions look like. We are mostly dealing with poker like environments and all the uncertainty that comes with it.
Knox: In business, leaders are often encourage to be decisive, but you give the advice that the best answer is sometimes “I’m not sure.” Why is that?
Duke: Well because you are not sure! We are afraid to say that we are guessing when we’re making judgments, but anytime that there’s uncertainty, then there is hidden information. The answer is that you cannot possibly be sure how it is going to turn out. And in some sense, you will be guessing. To say anything else to say, “I know for sure. I know this is right or wrong. This is how I’m going to guarantee you, this is how it’s going to turn out. I know this is the right decision.” None of those are accurate representations of the world or the state of your knowledge or your ability to guarantee anything.
The more accurate your representation of the world, the better your decisions will be as a result. That does not mean that I shift all the way to saying, you should not even try because you are not sure. In fact, just the opposite.
We have imperfect information for sure and the information that goes into any decision that we make is not complete, but it is also not zero. Our job as a decision maker should be somewhat singular and focused in terms of top line item, which is how much educated can we get into our guess? Because the more educated we can get into that guess, the more accurate our representations of the world are going to be. The closer we are going to get to what objectively is true of the world and that is going to make our decisions better. We can really do that with anything. When we are deciding, I want to start thinking about two things. One is what are the things that I know that can inform the decision? And what are the things that I could find out? That really unlocks a lot in terms of decision making as we build better models of what is objectively true of the world.
Knox: Your next book is How to Decide: Simple Tools for Making Better Choices. What lead to this being the subject of your next book?
Duke: My first book, Thinking In Bets, is a love letter to uncertainty. It’s sort of saying, “Look, you accept it. It exists and the more that you kind of see the world accurately for what it is, the better off you’re going to be. Let’s acknowledge that the uncertainty is there in these two forms, luck in hidden information and let’s try to kind of think about how that messes our decision making up and how that really frustrates our decision making.”
And then, there was a little bit of, how might you address it? I would say sprinkle of things that you could think about that might help improve the situation since we’re making decisions under these conditions. But what I got from a lot of readers was “You’ve convinced me, but how do I make a good decision? How do I do it?”
I started off just thinking I was going to write a workbook to go along with Thinking in Bets, but then I realized that I have a lot to say about what a good decision process looks like. There are a lot of places to go that when you start to get into the practicalities of making decisions. These were topics that were brand new and wouldn’t make sense in a workbook for Thinking in Bets kind of format. It ended up morphing into How to Decide, which is a book that has some interactive qualities to it. It’s not really a workbook, but it has thought experiments and exercises and decision tools and checklists and wrap ups and things like that, that really kind of codify and help you to instantiate the principles that are in the book. But it goes into it a lot of new material with a real focus on practicality and how can you make better decisions? And what would that process look like?
Knox: One of the topics you dive into is the paradox of experience. What do you mean by that?
Duke: Experience is certainly necessary for learning. If we want to think about how would we become a better decision maker, certainly experiencing the outcomes of your decisions should be helpful. In fact, I would say it’s a somewhat necessary component of becoming a better decision maker. But the paradox is that any individual experience that you have can really interfere with learning. Why? Well really, for two reasons, one has to do with resulting and the other has to do with hindsight bias. The problem with resulting is that we think that the result tells us something more than it actually does. It goes back to the chess versus poker problem. And then with hindsight bias, we think we should have known, or even worse, that we did know something. But if you were to look back, everybody seems to think that they did know something beforehand but there is usually no evidentiary record of this whatsoever.
Think about how much you doing this in your own life. You are getting outcomes and all of a sudden you are doing this resulting. You are deriving lessons about the decision quality from misremembering the past and thinking that you knew things that you did not or that you should have known things that you did not. Or when you are looking at other people’s decision making and their outcomes, you are over indexing on the quality of the outcome. You are not really looking at the decision quality. You are thinking they should have known stuff that they could not have known. All of this stuff happens to create the paradox of experience, which is that these individual experiences really can interfere with our learning. Over the long run it can be very helpful, but that is not the way that we process outcomes. We process them in sequence. We do not wait. We do not wait to aggregate a large enough N to say something significant about it. It is not like we are flipping a coin 10,000 times and then deciding whether the coin is fair. We are deciding that as we go.
Knox: Talking about hindsight bias, after the financial crisis in 2008, Charlie Rose sat down with Warren Buffett and asked, “Should wise people have known better?” And Buffett responded that, “People should always know better, but we have a natural progression of getting caught up in the silliness and not being willing to call it out for what it is.” Why do you think in these moments of hype, good decision making just goes out the window?
Duke: Well, because good decision making goes out the window all the time. Obviously we get this problem because a lot of the times we’re drafting off of other people in order to figure out if a decision is good or bad. And we can think about that just in terms of, do you go through a green light or a red light? I don’t think most people have done a cost benefit analysis of that, but they just assume that that’s the correct choice because that is kind of settled by society. We take a lot of cues by what other people are doing around us and so what happens is that very often, the market’s right. And particularly when there is a lot of information liquidity, we kind of assume, well if the market kind of says this is true, then that is the price. That the quality of the decision is kind of priced in as it has been crowdsourced in a liquid market.
We cannot do cost benefit analysis on every single thing that we decide about. So it is not the worst heuristic, but it is a bad heuristic when it is something that you are really betting on that you actually need to be thinking it through for yourself and really looking at it from different perspectives. And that is where we can get into trouble because we end up relying on these proxies for whether a decision is good or bad, or whether we want to be doing something or not. And when something weird happens so that the market in general is wrong so that the proxies are actually giving us bad signal, we do not spend enough time considering that the signal might be poor because we are not thinking it through necessarily for ourselves.
And this becomes particularly problematic when whatever the signal is telling us is something that we want to be true. Which I think is actually the broader problem here. When the world is telling, when we have some sort of strongly held belief which can occur either because it’s some sort of signal of our identity or because we’ve acted on that belief in the past in particular, we can become very entrenched in those beliefs. If we have some sort of subject matter expertise in that area that we’ve acted upon and particularly if we’ve had success with making decisions, it’s sort of driven by that subject matter expertise that has to do with the very strong models that we might have of the world. What ends up happening is we get pulled down into the trenches of our own models. And the stronger our model of the world, particularly we have had success, and the more identity driven it is, the deeper that trench is. And then if there is information out in the world that might give a signal that maybe we should be altering our model, we should alter it, but that’s not actually what happens. What happens is that we pull the information down into the trench with us and we will massage the information in order to fit with our model, as opposed to massage our model to fit with the information, which is really what we would prefer to be doing. That is kind of where things can really start to go wrong in that way.