Good Decision Bad Outcome
Was Theresa May right to call the general election? When sound decision making does not necessarily guarantee the outcome you want.
For more information on Aleph Insights visit our website https://alephinsights.com or to get in touch about our podcast email podcast@alephinsights.com
Transcript
Hello and welcome to the Cognitive Engineering Podcast produced by Tell Me Studios for Aleph Insights. In this series of podcasts we take a look at interesting topics and discuss what we think they tell us about analysis and decision making. I'm Fraser McGruer and I'm here this week with Peter Coghill and Nick Hare of Aleph Insights and we're discussing how wise or not it was of Theresa May to call a snap election. Peter, you have some thoughts on this.
Speaker B:Yes, so it's an interesting case. Obviously she didn't do nearly as well as everyone in the Tory party hoped and everyone's now calling the question was it a sensible thing to do? Was it a bad decision? And on the face of it, yes, it looks like it was a bad call. On the face of it, it looks like a bad call. So the immediate aftermath, everyone goes,
Speaker A:oh what a stupid thing to do, that was so idiotic to do that, what a terrible decision you made.
Speaker B:So she's getting a lot of negative press and her leadership is likely to get challenged and she's not achieved any of the aims that she really set out to achieve during the election. So yes, the outcome is very bad. However, this I think is a good excuse to talk about a phenomenon where actually you can, it is possible to have a bad outcome that results from what is a good decision. And there is a distinction there because the decision you make in the past, you only have a certain set of information, you only know a certain amount about the world and your decision is separate from the world. If you had all the information, yes, you'd probably make a different decision but you only have a subset of information that's either available or just information that's not there because of random events, weirdness and what is often bagged together as luck.
Speaker A:So that being the case, we'll start off with this example that we've begun with and then perhaps we'll broaden it out. But can we build a case in this example of why it might have been a good
Speaker B:decision to take at the time, either Peter or Nick? Yeah, well I think her decision was largely based on polls and information she got about what the likely level, what the level of support that she had from the voting public and what that would likely translate to in the number of seats. And the court with the decision to have the election was six weeks prior to the election, so what was the date, around roughly April time. So at that time it looked like it would have been an easy win, it looked like she would have got a strong majority. But then as things played out, as the world rolled on a few notches, things seemed to change and the popular support that she had seemed to somewhat evaporate and those six weeks give you a lot of time to make
Speaker C: ou know, bear in mind that in: Speaker A:I mean, I'm more or less willing to go with the argument. I think in this particular instance, I've seen elsewhere, someone saying that, and this is not to do with hindsight, saying at the time, actually, you can't just go off a general, you can't just go off the general polling. And it's more the amounts of swing that she would have needed in certain constituencies would have had to have been huge. And so actually, it was more of a gamble than it looked like just by going through from the popularity. Well, no, I mean, anyway, let's run with that. Well, let's put that aside anyway, because I still want to sort of go with this idea that you can make a good decision, even if it results in a bad outcome. Can we broaden that out? I mean,
Speaker C: s a sort of poker, one of the: Speaker B:And in addition to make it being very difficult to predict the likelihood of various, uh, outcomes is also in, in poker, it's relatively easy to do this, but in the real world, it's very difficult to value the outcomes in terms of what, what it is you gain and what it is you lose. So, um, there's no winning a general election. You can't, you can't really put a price on that very easily in terms of what was, what does that give you in terms of influence? And, uh, what is, what is an additional seat in the house worth, uh, in, in some sort of measure that you can then compare with other, without, with other outcomes. In poker, it's easy because you can, you can calculate the winnings on any game. And so you can derive kind of the, what the, uh, the, the value of a particular
Speaker C:outcome is with respect to other outcomes. Um, but I mean, even when we, our urge to blame bad outcomes on people is so strong that even where we can, even when we've got really good data, uh, we still are looking for, we still try and assign outcomes to, to stupid behavior. And I mean, the, the classic example for me is the, because I'm not a big soccer fan, I kind of find this amusing. Uh, but the fact that, you know, managers of football teams are constantly being, they're being, they get sacked if their team does badly, you know, in a way that's completely consistent with chance. There's a guy called, um, uh, Shimansky who's, who's looked at the data, looked at basically trying to work out what predicts football success. Do managers make a difference? And the answer is no. Right. The only thing that makes a difference to your team's chances of winning by and large is how much you pay the players. You buy, you spend more on players, you get better players, you're more likely to win. Managers have almost no influence. And yet we can't, we can't stop sacking managers, you know, especially poor old England managers, the worst job in the world. All they've got to do is lose three matches on the bounce and they're out. Um, you know, I mean, the flip side is probably doesn't matter if you do that because they're not doing much anyway.
Speaker B:Well, maybe, maybe, well, maybe it's difficult to see what effect they're having because they don't stick around long enough. Well, there's that, yeah. Actually, that sort of makes me think there's something
Speaker A:interesting that they do in American football. The irony of having what is in effect a socialist system within a game that is emblematic of a country that's the most dynamic economy in the world, which is the draft system that they have, where I think, how does it work is that, well,
Speaker C:it's, it's about as bonkers as their electoral system. Yeah. I can't, I can't, no, I don't, I don't understand it at all. Well, basically, so I think, basically they get, they get dealt out
Speaker A:like a pack of cards. No, well, I think the way it is, is that I think the American university system feeds the professional system and all the players are rated. And, and let's say, I don't think there are 20 teams, but let's say there are 20 teams. I think it's more like 15. Um, but the, the best rated, um, graduate from college goes to the worst performing team from the year before and, and, and, and so on. And it continues to the second best goes to the second worst. And you cycle through all the players until they all get dealt out, which is, um, yeah, which is unlike, let's say European soccer, where it's just all money. Yeah. But on the other hand, yeah,
Speaker C:no, you're right. But on the other hand, obviously that could well make it more exciting. Oh no, absolutely. It does. One of the things you want to have close competition. That's what makes it
Speaker A:fun. Yeah, absolutely. And that's sort of, yeah. Yeah. And that's what people like. It must lead
Speaker C:to weird incentives. It strikes me. You probably wouldn't want to be the best guy in your college. Yeah. And, and you don't want to be the best guy on paper. Yeah. And you don't really want to be the best team in any given year. You'd rather be in the middle. Except you sort of do in a way. Yeah. But you're, you're then sort of punished for doing well. Yeah. But look, we're getting close. I think that's called rubber banding. Yeah. Yeah. They use it. It's like, it's a way that computer games cheat. So in a driving game, uh, if, if you have rubber banding, it basically gives you a bonus if you're doing badly and slows down the cars in front, you know? Yeah. Yeah. Yeah.
Speaker A:Um, okay. Look, we're close to sort of wrapping this up. I mean, one thing that interests me is looking back at one's own personal decisions and whether they were good or not. I can talk about, we can talk about that or we can go off in a different direction. Well, I think maybe we
Speaker B:should, we should talk about what you can do with your, with your lot. As a decision maker. There we go. What you can, what you can do. Um, so I think there are a few things that when you're making decisions that you should do to really just to, um, save your skin in the future, uh, if nothing else. Um, although, you know, good, good, good decision making is something you can get better at. You can, you can, you can get better at calculating the outcomes. You can get better at working out the probabilities of certain outcomes. Um, it doesn't, all of that effort doesn't guarantee a good result. So at the end, when a bad result happens through some external influence, you want to be able to say, here's my audit trail. Here's why I had that made that decision. So really simple things like recording all, as many assumptions as you can about why, what led you to think the world was going to be a certain way and where possible, employing a robust method that exposes all of the potential outcomes and robust methods for scoring or valuing those outcomes in terms of what, you know, what, what's good and what's bad for you or based around whatever it is you're trying to achieve. So, you know, having an audit trail, um, is essential in some disciplines like law and, uh, insurance and things. They have a, they have a reason to value things in a certain way. Um, but it's often lacking in business
Speaker C:decisions and things like that. And it's weird. I don't know why you would want to not do it that way. It's always surprised me. Same with politics as well. Politicians are always being very obfuscatory about their reasons for doing things. And you just think, look, if you don't have a reason, why are you doing it? You know? Um, it's, it's very strange, but I think that, and also just to add to what Peter said about how to do it better, if you're, um, if you're not the decision maker, but you're trying to appraise someone else's decision, uh, the extent to which you can use the outcome to do that is driven by how much information that person had or could have had, uh, at the time they made that decision. That's really what it comes down to. Um, you know, anything that happened afterwards is, is, you know, no, especially if it was driven by very low probability surprises, uh, should, they will weaken the, the evidence value of the outcome as a means of appraisal, but you should still do it. I mean, I mean, outcomes are evidence in the long run, certainly out, you know, someone keeps getting bad outcomes. They've got to start taking a look at themselves in the mirror and saying, maybe this is maybe, maybe I am the problem.
Speaker B:Yeah. Well, and, and having that audit trail is also very useful as you're, as a decision maker for the future decisions. Cause if something, if you've got a robust method and you've got, uh, you record all your assumptions and you build a record of what data you had and something other happens, something else happens, that's something else then to consider along with all the lists of things you considered last time for the next decision of that kind. Um, so you're, you're, you're, you have, uh, uh, you know, you've got something to build on for next time.
Speaker A:We've got just enough, enough time. This makes me think of a decision I made about two and a half years ago, which was whether to go into business with a certain person or not. Um, and to become partners and this is going to be dirty laundry time. Yeah. And, um, and I just, I decided, um, so I decided to go into business with that person and in many ways it's not worked out well. Um, and looking back at my decision, I think, well, what could I've, what were the, what was the writing on the wall? Was there something I sort of, was there good evidence supporting a different decision? And actually there probably was, which was actually, I, it took me a long time or it took a long time for that person to get back to me and fully engage with my proposal, um, months maybe. And, but I wanted very much to go into business with this person for different reasons. Um, but I think there was a warning sign right there because actually one of the problems with the business turned out to be was a lack of engagement from that person, which is exactly what that person had shown, um, in that bit before committing. And so I guess, you know, the lesson is there for me now. And, but I think that's one of those examples actually was probably not a good decision despite a bad outcome. It was a bad decision with
Speaker C:a bad outcome. Well, I mean, just thinking about setting up Aleph Insights. So when I set this company up a couple of years to two and a half years ago, again, um, I, you know, I looking back, I mean, we've been, uh, I would say probably about as successful as kind of our upper estimate of the kinds of things we were expecting to achieve, which is great. You know, it's done really, but, um, you know, I know, I know that I did not foresee that. And a lot of it has been, uh, driven by, um, you know, some fairly chance events happening to know the right person to collaborate with happening to, you know, get, uh, get onto a particular framework at the right time and a bunch of other things, you know, so I, I, much as I'd love to sit there and take credit for being some kind of business genius, uh, I actually know that that's, uh, you know, it may have been on average a good decision, but it wasn't as good a decision
Speaker A:as the outcome might suggest. Well, we're going to have to wrap up there. And unfortunately, I'm going to have to finish thing on a, on a negative note, which is Nick. I'm going to have to disagree with you because actually you are a business genius. Thanks, Fraser. Um, thank you, um, to Peter Coghill and Nick Hare. I'm Fraser McGruer. You've been listening to the Cognitive Engineering Podcast. Thanks as always. And until next time, bye-bye.
