Annie Duke on Trading and Poker
Annie Duke is a professional poker player and author. She was at one time the leading female money winner in World Series of Poker history, having also won the 2004 World Series of Poker Tournament of Champions. Her 2005 autobiography is titled, How I Raised, Folded, Bluffed, Flirted, Cursed, and Won Millions at the World Series of Poker..
She may not be a trader, but in this discussion with Michael Covel in Trend Following Radio, the two discuss several of the countless ways in which the psychology of gambling overlaps with that of trading, investment and other aspects of business.
Annie explains the importance of thinking probabilistically for decision-makers. Gamblers, like investors, can sometimes become so focused on their losses that it begins to affect their decision-making process in a negative way. Annie calls this “tilt” and says it occurs when players put too much emphasis on outcome. She points out that as long as you are getting an overall return on your investment via a positive expectation, small losses should be both expected and absorbed.
“The percentage of times that you’re winning, is only relevant in relation to what you’re getting in return” – Annie Duke
On using odds in every day life…
I think it’s incredibly important to think that way, because if you’re going to be a good decision maker, you will have to be thinking probabilistically. Because you have a lot of alternatives in front of you all the time, one branch of that tree is not 100% even though it might even feel that way.
What people are generally saying when they say “thats 100%”, is, it’s highly likely. If you’re not considering the other stuff that could happen, I think a few things happen. Number one, you tend not to consider them, because let’s say you think somethings 80%, you’ll just say thats 100%…I’m pretty positive…it stops you from looking at what the other 20% is, which I think is bad.
So I think what happens is when you think somethings going to happen the majority of the time, you default to “100% thats going to happen”, that actually prevents you from considering other alternatives. And when you actually explore other alternatives, you can come up with some pretty creative solutions. Once you default to 100% you’re not considering any other choices. The other thing that happens, which is really bad and in someways might be worse, is that when it doesn’t work out, by definition it wasn’t 100%.
Now what you do is you just assume you were wrong. Instead of explored in advance, “well this was 80% and then 20% of the time other stuff was going to happen”, so now when it doesn’t work out..it doesn’t necessarily mean my thinking wasn’t correct. So you’re much more likely to change your behaviour in a way that’s actually not productive, when you define something as 100%, because when it doesn’t happen you just assume you were wrong.
There are traders who enjoy betting and there are poker players and sports punters who see spread betting as a natural evolution of their activities. Poker players are probably more suited to crossing the divide to trading, because of the mathematical skills required to work out the betting odds on each and every hand that they play. Managing your bank-roll in a poker game has very similar principles for adjusting the size of your trades to the size of the investment. Understanding the probabilities of the cards you have is not that much different from analyzing a candle stick chart and assessing the expectancy of what a trade can yield.
On thinking about expectancy…
My PhD work was in science, Psychology. We were thinking about expectancy all the time and you are thinking about P values, which had to do with confidence intervals, different sciences have different P values that they allow. And in Psychology, generally 5% is ok. If there’s a 5% margin of error, then you’re alright.
In Physics you want it to be much less than a percentage before you have any confidence. So that idea of confidence intervals that different sciences were demanding different levels of confidence before you would accept a result, that was already in my thinking. I went into poker with that mindset, but it is something that absolutely separates out the top thinkers from everybody else. That idea of expectancy and really thinking about things in terms of what is your expected value over time.
On luck…
It’s interesting, I’ve actually thought about this quite a bit, what does good luck and bad luck mean? I feel like for people who are traders, good and bad luck means something very specific, which is that I had positive expectancy, and I had enough iterations that I really should have realised it, and I didn’t. As an example, if you are laying me 2:1 on a coin flip, and I lose the first time, I wouldn’t call that bad luck. I had positive expectancy there, but I haven’t run enough iterations in order to realise the gain yet.
But if you’re laying me 2:1 on a coin flip and we flip 50 times, and I somehow manage to lose, I would consider that unlucky. Because now I’ve really gone against expectancy and I certainly ran enough iterations that I’m way out at the tail. But I think the way most people use it in every day life, bad luck is kind of like, a bad thing happened, and good luck is a good thing happened. And in someways I think that they think of it as 50/50. So if you’re above 50%, and it doesn’t work out then that’s bad luck. And if you’re below 50% and it does work out, it’s good luck. I don’t think most people get beyond that step.
On process vs outcome…
When you focus on outcome instead of process, you actually reduce innovation in your company. And the reason for that is because innovation requires failure, it requires a lot of bad outcome. Because innovation is very high volatility. You know that you’re going for a big win but in going for a big win, you’re going to have a lot of losses along the way.
If you as a manager are very outcome focussed, you will create for the people working for you a motivation to create a lot of wins, a lot of good outcomes. But that reduces your expectancy. So you have to tease those two things out. If what you’re trying to do is book the biggest number of wins that you can, that doesn’t mean in any way, shape or form that you’re maximising expectancy and in some cases you’re minimizing it.
So you can think of a poker analogy where what would happen if I was to just win a lot? Well what I would do is I would go in and play and anytime I got to above zero I would just quit the game, so I could go and mark the win in my books. I could go in and win the first hand, I would just quit, because I would risk losing that win. You can see I assume how that isn’t a winning strategy, because what happens if I lose the first hand? My losses are going to be bigger than my wins because I’m taking a lot of small wins.
I could look over the month and say I played 20 days and I could win 16 of them and still end up either losing for the month or making a very small amount, where I wasn’t actually maximizing the amount of time I was playing at the table in some sort of positive expectancy situation…
On poker and in business…
What I would say in poker, people think far too hard when they’re playing poker and when they have the best hand. This is something they think about all the time, “do I have the best hand, do I have the best hand?” And I try to say to them, that 0% of the time you should be thinking whether you have the best hand. What you should be thinking about rather is when you can win the hand enough of the time. Now that’s a really important distinction.
I’m not really going to think very hard whether I can get to 100% as to whether I have the best hand. What I’m going to think of instead is can I win the hand enough of the time for the return that I’m getting?
I think thats how venture capitalists are thinking as well. When tech companies win, they really win, like Uber or Angry Birds, or something like that. I don’t have to think too hard that this particular company I’m investing in is going to win, I just need to think about how often I need it to win in order for it to be profitable, to the kind of decisions I’m making.
So they can make a decision that the company is going to succeed 25% of the time and they’re totally fine in investing in that. That means 75% of the time the company is going to fail, and they’re totally fine with that.
On the concept of Tilt…
I think it’s something that’s under appreciated. I haven’t played poker in over 3 years, what I do instead is I go and talk to people, in all sorts of different areas, a lot of finance, HR, sales, energy etc. I’ve done a lot of work in the energy sector recently, which actually tilt applies to right now.
So I talk to all these different industries because when you are talking about behavioral economics, these are generally things about the way people think. You’re talking about very robust processing, in the way we think, the way we process information, the way we remember things. It doesn’t really matter what industry you’re in, the concepts are going to apply across. It’s actually specific to the way the human brain works.
Now the reason why poker is a really good place to be looking at this stuff, is because when we think of the definition of game theory, which is decision making under certain conditions over time, it happens to be a very good definition of the game of poker, which is not accidental. John Neumann, the father of game theory, modelled his original thinking on that topic of the game of poker, he was actually an avid gambler. And he understood the difference between poker and chance, which had to do with this uncertainty piece, the answer can come from two places, hidden information and volatility.
The question is what happens to people when they are making decisions under those conditions, you have hidden information, it’s volatile, those two things combined create a problem, which is you don’t really know what’s in your control and what’s out of your control. Because you don’t have perfect information. If I lose to you at chess I know exactly why I lost, you can go back and analyse the situation, I can’t shift if off to luck. It’s interesting what happens to people in poker when they lose a hand, where they are very sure that they had a positive expectancy in a hand or a couple of hands in a row, the ball goes against them, they would perceive that they were unlucky, they would do on tilt.
Tilt is a word that essentially means “I’m so emotionally distressed by something that has happened to me recently, most likely something I perceived to be most likely not to be in my control, that I will now make terrible decisions going forward because I’m in an incredibly hot state, and I’m going to do a lot of ridiculous things going forward because of that.”
On a trader’s mindset…
You can set up structure in your company that will actually encourage both kinds of tilt. What you will generally see in the trading world is that, if somebody is in the bottom 25% in terms of performance for that year, again you’re not thinking about process, you’re only thinking about outcome here.
And you know we’re coming up to the end of the year and you will tend to see them making much more high volatility choices than someone who is in the top 25%. Who probably will be making choices which are too low volatility, in order to maintain that position because they’re going to feel like the bonus is going to be dependent on outcome and not on process.
If someone’s in the bottom quartile they’re going to be displaying this tilt behaviour or seeking out volatility. You see this in traders all the time, who haven’t got out of a position because they were losing and now you see people who all of a sudden used to be a good trader, they’re now on a losing streak and they don’t understand because they were so good before. You will generally see that they are making choices that are too low vol, just trying to avoid that state of pain.
Whether you’re a trader, business leader, or poker player, the principles of focusing on expectancy, controlling emotional biases, and prioritizing process over outcomes are indispensable for sustained success.
This insightful discussion between Annie Duke and Michael Covel effectively bridges the gap between poker, trading, and decision-making under uncertainty. It highlights how concepts from poker—probabilistic thinking, expectancy, and emotional control—translate seamlessly into the realms of trading, investing, and broader business strategy.
