Even the most calculated decisions don’t always yield good outcomes. There’s always information hidden from view — always an element of luck. That’s what makes business and poker so alike, and why former professional poker player Annie Duke is an expert on the topic of decision-making. For two decades, Duke was one of the top poker players in the world. She’s written numerous instructional books for poker players and also authored two books on decision-making when you don’t have all the facts. Duke recently joined Jawwad Rasheed, senior principal in the office of finance at Alteryx, to share lessons and strategies on decision-making cultivated throughout her academic studies in cognitive psychology and experience at the poker table.
How businesses can use data analytics to make decisions when they don’t have all the facts
At the professional level, a hand of poker takes about two minutes, and during that time, a player is making upwards of 20 decisions, constantly evaluating minute variables like how often their opponent folds or bluffs or how likely certain cards are to appear.
In essence, players are striving for optimal decisions with incomplete information. “Whether you’re an amateur or an elite player, if you decide to raise, you’re making some sort of forecast or guess about the range of possible outcomes given that you’re choosing that particular path,” says Duke.
“The difference between what an amateur is doing and a professional is doing is that they’re making that explicit. They’re actually breaking the decision down to its components and then thinking as objectively and explicitly as possible about the data that they might input into the way that they’re modeling that particular player.”
Duke explains that it’s the same with data analytics and business decisions. Decision makers can use data analytics to break down choices by their related variables, understand the base rates of those variables — or how often they occur under normal circumstances — and analyze how those different factors may affect an outcome. Armed with that information, decision-makers can make educated decisions while removing their cognitive biases.
The problem with “resulting”
Resulting is the tendency to judge a decision based on its outcome. As people, we’re skewed to think decisions were good or bad depending on what happened, regardless of the inputs for the decision or how well we accounted for unknown variables.
“When you’re making decisions under uncertainty — meaning there’s a whole bunch of stuff you don’t know — luck can interfere in the way the decision turns out,” explains Duke. “When you think about the relationship between decision quality and outcome quality, there are four possibilities. We can think about this in a two-by-two table.
“You could make a good decision and get a good outcome. You could make a good decision and get a bad outcome — that would be because you had bad luck — and you can make a bad decision and get a good outcome. That could be because you had good luck. And you can make a bad decision and get a bad outcome.” Duke says that any of these four are true when resulting is a problem.
|Good decision||Bad decision|
|No luck||Good luck||Good outcome|
|Bad outcome||Bad luck||No luck|
“So we can think about where that’s a problem. Let’s imagine that you make a terrible decision that’s gonna work against you, kind of like the hand where I made this big bluff because I was just so sure that the person was gonna fold, and then they called, and I won the hand anyway. Resulting would be making the mistake of saying, “Oh I played that hand well because I won. …
“This is what we do all the time, like when someone goes for a two-pointer in football instead of kicking the field goal. If they miss, we think that the coach was an idiot, and if they actually get the extra two points, we think that they were a genius.”
Duke explains that it’s the same when we miss business targets or fail in our strategic initiatives. “The fact is that it could have been a perfectly reasonable choice that just got a bad outcome, but that’s not really the way that we think about it, and that’s the problem with resulting.”
How to use prospective hindsight to overcome resulting
Resulting is a difficult bias to overcome, especially when we don’t understand why our decisions may or may not have been good except for the outcome. Luckily, Duke has a countermeasure.
“One of the best ways to do this [overcome resulting] is through something called prospective hindsight, and there are two ways to do it,” explains Duke. “One is called backcasting where you imagine that things have gone well and then you say what happened along the way, what the early signals were that things were going in our favor. And the other, which is actually more useful, is called a premortem where you imagine failure, and then you think about what happened along the way.
“What did we not execute on very well? What were the early signals that failure might occur? The reason why premortems are more useful is because we’re naturally overly optimistic. We overestimate the chances of a good outcome. When we do backcasting, we’ll tend to reinforce that particular bias and become more optimistic in imagining good outcomes.