Nick, Peter and Fraser discuss what the Olympics show us about the the human pursuit to reach for the limits of performance.

For more information on Aleph Insights visit our website https://alephinsights.com or to get in touch about our podcast email [email protected]

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 with Nick Hare and Peter Coghill of Aleph Insights and this week we're discussing the Olympics. So I woke up this morning and was really pleased to see my notification on my phone that Great Britain has just in the early hours of the morning had won its first medal and it was a gold at the Rio Olympics. Now rather handily for our purposes the medal was won by a chap called Adam Peaty in a world record time of 57.13 seconds. Now this time was four-tenths of a second better than his own record which he set earlier in these games. So tell me Nick, in classic Cognitive Engineering podcast we're talking about the Olympics but actually we're possibly really talking about something else. Well I think the question here which we're

interested in is what are the limits to performance? Are we going to get to a point where world records are all in the past and we should hardly ever see world records being broken or are there reasons to believe that things will continue to improve, that we will have ever faster times and higher jumps and you know longer throws and so on. Now the question here is how do we model what is happening to generate the results that we've got? Once we start to think about how we model what generates those times we can start to see well what things do we think might be influencing it and my first thought was well maybe it's just a statistical effect right. So let's imagine you've got a hundred dice and you roll them and you add them all up and then you roll them all again and you add them up and you roll them again and each time you take note of what the highest number was. So you know the first time you roll it you might end up with you know you might get somewhere around 300 then you might have 250 and then you might have 350 and that's your new record. So you keep rolling and then maybe you get 360 then you keep rolling you keep rolling and then a bit later on you know you throw a record 385 and the point is that in a process like that the maximum will early on will change quite frequently and go up by quite a long way. Whereas later on as you repeat the process over and over and over again the maximum will only change very infrequently because you know most of the time your rolls are going to be below whatever your previous maximum were and also the difference each time will be less and less. So you know eventually your maximum might be you know 420 and you might get a 422. So are the Olympics a process a bit like that? Are we simply sampling from the same distribution over and over again and occasionally we get you know a freak outcome and that then becomes the time. Now this is probably not the case right because if you look at what I mean okay how big is the sample you might say the sample is actually everyone in the world. So if we sample if we've got the same sample if we're sampling for the same population a hundred years ago the world record was 10.8 seconds for the 100 meter sprint now it's 9.58 seconds and in fact Bolt's improvement was the largest margin of improvement in between world records when I think he lopped 0.11 seconds off. Now that's not what you'd expect to see when you have such a large sample you would expect to see a real slowing off of growth and it should be very infrequent and very very small but actually the improvement has been kind of linear. Now what that means because we have such a large sample we have been sampling from millions and millions of people every year to try and effectively find the fastest person on earth. With processes like this if it's you know if it's 100 billion or if it's 1 billion you actually sort of affect expect to see more or less the same maximum. So what I'm saying is we've been getting faster than we should be if it was purely a statistical process so that means something else is in play. Okay that's a good starting point. Peter. I think well the the dice rolling

so if we if we depart from there for the moment but I want to talk about marginal gains because although it connects to what we're talking about the Olympics and I think for example what's happening in the British team is is interesting in certain areas where for example in cycling where I don't know the name of the chap but the coach or strategy director whatever he's called for for cycling has made a big point about marginal gains and this is how we're going to be successful and sure enough that's precisely what has happened but I want to sort of take it outside the realm of Olympics and sport and Peter can you sort of tell us about marginal gains and what's a better way to phrase that I mean what is the usefulness of marginal gains what does this tell us about the impact of marginal gains? So marginal gains just to explain

but you kind of said that at the beginning well the the the reason that that approach can work in those contexts and in sport is that we have a very easily defined objective you can't use this sort of approach where we don't have that so so um you know things like uh well an example i've mentioned before but trying to design you know trying to trying to build a perfect blockbuster film for example we don't know what we're trying to achieve so we can't we really only have sort of sales and that's driven by tons of other things which aren't just about your film um whereas whereas uh things like you know whether or not a nozzle gets blocked are reasonably easy to measure and and um the outcome of a sports training regime is also easy to measure you can you know because we know that it is entirely described by the time at least in terms of what we're trying to achieve with the olympics um so yeah i mean i think that points to uh you know a useful way of thinking about whether the evolutionary approach is going to work for your problem is you know do you know exactly what it is you're trying to achieve and if you do actually this kind of scientific approach is quite a good one um okay um but but i think it's interesting is there also the question is whether there were going to be non-marginal improvements so classic example might be the fosbury flop which is the style of uh doing high jumps that uh was invented it was an innovation and um you know previously there were kind of the flopping over forwards and there was the scissors sort of approach and uh dick fosbury invented a way of doing it where he jumps backwards and um actually you know there when you when you look at the the physics of that uh it sort of makes perfect sense because you're keeping you can keep your center of gravity below the bar so you effectively don't have to propel your body up as high as the bar because your body is kind of work almost coiling around the bar so you know it makes it makes sense but but no one came up with it until he tried it or that you know would have said well that's never going to work i'm not going to not going to build my strategy around that um now the question is are they are they out there do we have any more of these gains to be made somewhere in sport um can you bank on them other is there is there a process you can use to try and find them

okay so marginal gains yes but innovation and that's a that's a big criticism of the marginal gains approach because it focuses you down onto improving on what you're already doing you're potentially missing out you're you're you're you're you're saying you begin to maximize towards a local maxima rather than making bigger jumps and hopefully finding a a bigger max the maxima of the

the thing you're trying to optimize okay so um so where do we go from here nick well one other thing we could think about is the is the theoretical limits right now uh you know how fast could a thing the size of a human do the 100 meters in and um when you when you ask that question almost for every category of uh activity the answer is we're absolutely nowhere near those sorts of the physical really well in the sense that you could sort of imagine build now it depends what your constraints are you know if it has it still got to have two legs you know um can it have is it allowed to have wheels you know so the the the question of how much if you could genetically engineer a human you know to to be as efficient as possible while still being in some sense conforming to what we understand to be a human uh you know if that was possible would we then start to want to regulate that and say actually you know you're not allowed to enter the 100 meter sprint unless you've you know you've got certain uh you conform to certain limitations you've got you can only have two legs those legs have to be a certain percentage of your body height um i mean one of the reasons they're saying bolt is so fast because he's very tall so you know could could you as peter touched on it but could you imagine breeding for height um really optimizing a person towards um 100 meter sprints and and do it that way peter yeah this and this this touches on a

bit of sports science which is only beginning this is it's you see it being used in big institutional sport like football and things where athletes at the very earliest age when they are taking an interest in sport and looking like they have potential are pre-selected based on their body type and their and their stature there are still you still see the odd um long distance runners and the odd cyclists who are just a slightly strange shape compared to many other cyclists and um that may give them an edge in certain things like sprinting you want big powerful people but they are not going to be a general category general category rider um but there are certain optimizations so you could take the margaret games approach and optimize even further and if somebody's showing an interest in archery suggests actually well actually you might be a bit wasted why not try 100 meters sprint 100 sprinting instead um and there's lots of optimization still there and that that would probably get you a lot closer to the the theoretical maximum that uh the human body can out yeah i mean there's a this is the um

you know it's a question about mozart of of whether or not he you know everyone says he's a prodigy and wrote you know symphonies at the age of five and so on but how many other mozarts are out there who just simply didn't have a piano in the house or a dad who kind of pushed them in that direction that technology which enables you to identify good people and to promote them is a perfectly good way of getting better performance so effectively we're widening the pool now you know if if if we're so um you know just even some of the things people have pointed to is the greater participation of women in society um means that if you have you know if you've got relative gender equality you're you're going to do better in the olympics because women are more likely to to um take part in sport uh you know things like that social technologies which which sort of enable people to realize their potential are just as good as coming up with a new design of shoe or even better have a really autocratic government yeah which pushes people

yeah you have an army of people yeah yeah or a chinese approach i'm thinking more but yeah russian as well but just have an army of people scouts who go out there looking for someone who's four foot tall when they're eight years old or whatever go great you're gonna be a gymnast um and off you go that's probably not quite the right parameters but anyway look we're pretty much there um anything burning that anyone wants to add or anything you want to say to wrap this

up picking picking up a point that nick was alluding to about the need for greater controls and legislation around sort of enhanced performance and things is there's a good sort of test debate going on at the moment in paralympics a lot of the olympians are uh under a course of gene therapy for their particular disabilities um and there's a very fine line between gene therapy and gene doping and it's it's a very difficult area to to unpick if you're getting therapy which is um fixing a neurological disorder and you are in the and you are competing as a uh neurologically disabled person at what point does your your therapy become doping and and and should you should therefore be excluded because you are having your your your uh your you're getting therapy which is fixing the thing which is which is not working correctly

yeah it's possible to imagine a kind of disability in inverted commas which makes you a really good sprinter well no this is already good swimmer so this is what was happening with

Oscar Pistorius where he wanted to race in the normal olympics um but they were saying that you can't because your prosthetics will give you an advantage and i know i know that's taking a step further than what you're saying with the natural um well although you know is it any

different to just shoes how do we know you know why do we it's still only his own power it's not like they're powered legs i mean you know it's still his own power so uh it is a very tricky call i mean i think you know why why isn't it just perhaps the best way to be a good sprinter is to um you know to to be an amputee with uh with you know a different sort of shoe what if that is

you know why not oh i look forward to seeing your submission to the british olympic committee about this so it's a great idea um okay we're gonna stop there i mean what i take away from this and what i really liked was uh peter's advocacy of being able to breed with whomever we want um i i like that so that's that's a good takeaway um so if they'll have you for each other damn that's the yeah that's not good um okay so we'll stop there um thank you very much guys um you've been listening to the cognitive engineering podcast i'm fraser mcgrew we've been here with nick hair and peter coghill thank you for listening and until next time goodbye