AI transcript
0:00:05 The Medal of Honor is the highest military decoration in the United States.
0:00:12 Since it was established in 1861, there have been 3,517 people awarded with the medal.
0:00:17 I’m Malcolm Gladwell, and our new podcast from Bushkin Industries and I Heart Media is
0:00:19 about those heroes.
0:00:25 What they did, what it meant, and what their stories tell us about the nature of courage
0:00:26 and sacrifice.
0:00:32 Listen to Medal of Honor Stories of Courage on the iHeart Radio App, Apple Podcast, or
0:00:44 wherever you listen to podcasts.
0:00:57 Around 2010, Sarah Rudd was rooting for this British soccer team.
0:00:59 It was a team called Blackpool.
0:01:01 It had been her great grandfather’s favorite team.
0:01:03 She was really into it.
0:01:06 But she started getting annoyed by this one player.
0:01:08 His name was Charlie Adam.
0:01:14 She just loved to take shots from really long distance, and it drove me crazy.
0:01:19 Every possession just kind of ended with him taking these random shots, and it’s like,
0:01:20 “What a waste.
0:01:21 What are you doing?”
0:01:25 Sarah was living in Seattle, working for Microsoft as a software engineer.
0:01:28 But she really wanted to get into sports analytics.
0:01:32 Basically, she wanted to do moneyball but for soccer.
0:01:38 And then she discovered that this sports data company called StatDNA was hosting a contest.
0:01:44 A contest where people could use the company’s soccer tracking data to generate analytical
0:01:45 insights.
0:01:48 And she thought, “This could be my chance.
0:01:51 This could be my chance to break into the industry.”
0:01:57 And also, at the same time, that guy, Charlie Adam, really should not be taking those long
0:01:58 shots.
0:02:04 I’m a big fan of searching for inspiration elsewhere, so I started just looking around
0:02:07 and seeing, “Well, what have other people done in this space?”
0:02:10 And it turns out for soccer, not much.
0:02:14 But there was some really interesting papers in other sports, and the one that caught my
0:02:17 eye was by a gentleman named Keith Goldner.
0:02:24 He had done something using NFL data, where he looked at kind of this, “If the ball is
0:02:29 on the 35-yard line, it’s second and 10 with X amount of time on the clock.
0:02:32 How likely are you to score?”
0:02:36 And I thought that was really interesting, and I thought, “Well, isn’t that kind of
0:02:40 similar to Charlie Adam having the ball 40 yards out?”
0:02:42 So Sarah took this methodology.
0:02:47 It’s called a Markov chain, just FYI, and she applied it to soccer.
0:02:52 She built a model that let you look at a moment when a player had the ball in a given location
0:02:58 on the field, and the model could evaluate what the player did at that moment.
0:03:03 Specifically, the model let you ask, “Did the choice the player made?”
0:03:08 Passing forward, passing to the side, shooting, whatever.
0:03:15 Did this choice the player made increase or decrease the probability of the team scoring
0:03:17 a goal on that possession?
0:03:25 And, yeah, it really hadn’t been done in soccer before, which was kind of why it was kind of
0:03:26 a revolutionary paper.
0:03:27 I hate to use that term, but —
0:03:28 I’ve seen it listed.
0:03:29 I’ve seen it listed.
0:03:34 538 had a list of the 10 big papers in figuring out this kind of thing, and yours was the
0:03:36 first one on the list.
0:03:40 So it was chronological, to be fair, but still, big first is a big deal.
0:03:44 Yeah, and it was cool, because I was kind of like, “Oh, well, I’m just doing what this
0:03:50 guy in the other football is doing,” but yeah, it was pretty impactful.
0:03:53 And by the way, what did you find out about Charlie Adam?
0:04:01 Yeah, not good to shoot from four yards out was what I found out, which was, yeah, really
0:04:09 soul soothing for me, or it was like, “I was right.”
0:04:12 I’m Jacob Goldstein, and this is What’s Your Problem.
0:04:17 Today we have the third and final episode in our series of interviews with people who
0:04:22 are working at the frontiers of technology to help elite athletes perform better.
0:04:25 My guest today is Sarah Rudd.
0:04:30 She’s the co-founder and CEO at Source Football, a soccer analytics company that works with
0:04:34 professional soccer teams in the US, the UK, and Europe.
0:04:39 There are different ways to frame the core problem that Sarah is trying to solve.
0:04:44 For today’s show, I’m going to go with a problem that generalizes way beyond sports.
0:04:46 The problem is this.
0:04:53 How do you translate analytical insights into meaningful changes in the real world?
0:04:57 In our conversation, Sarah and I talked about analytics and the challenge of getting people
0:05:02 to change their behavior, change the way they make decisions based on analytics.
0:05:06 But before we got to all that, we talked a little bit more about that first paper that
0:05:11 Sarah wrote to prove that Charlie Adam should stop shooting from 40 yards out.
0:05:15 Sarah told me that paper also led to some less obvious insights.
0:05:16 Yeah.
0:05:23 I think one of the first ones was that crossing from wide situations, which was really still
0:05:24 very popular.
0:05:28 So if you think of a player coming down the wings, and then they’re going to kick the
0:05:33 ball high in the air, and then hopefully a teammate is going to head it into the goal,
0:05:36 that is also pretty low value.
0:05:37 Interesting.
0:05:42 Because you have a really high chance of turning the ball over from that.
0:05:46 And was that finding contrary to the conventional wisdom of the time?
0:05:47 Yeah.
0:05:48 At that time, it was.
0:05:53 At that time, there were still a lot of teams using that tactic.
0:06:00 And I think going back, people were starting to just intuitively be like, “Maybe there’s
0:06:01 a better way.”
0:06:08 So it’s not like it kind of broke soccer or football, but it just kind of I think reinforced
0:06:11 some of the intuitions that people had where it’s like, “Maybe there’s a better way to
0:06:12 do this.”
0:06:17 And just to be clear, in the time since then, in the whatever, 15 years or so since then,
0:06:22 have analytics shown clearly that that’s a bad call and has that style of play decreased
0:06:23 as a result?
0:06:24 Yeah.
0:06:29 And now it’s actually seen as a sign of the offense is struggling or we’re resorting
0:06:32 to this really low probability tactic.
0:06:37 So this paper you write gets you hired by this analytics company.
0:06:43 And then the analytics company gets acquired by Arsenal, the famous London soccer team.
0:06:48 And you wind up working at Arsenal for many years.
0:06:54 And one of the things I’ve heard you talk about, about that time, and that clearly is
0:06:59 a big important theme that goes beyond soccer, is trying to figure out how to get the people
0:07:06 who make real world decisions to actually listen to you, to you and the analytics people.
0:07:07 Yeah.
0:07:13 It’s really hard because if you remove sports from it, just getting anybody to change their
0:07:18 opinion based on facts or evidence is really, really difficult.
0:07:24 And now you’re talking about people who are in incredibly stressful jobs where they can
0:07:29 lose their job really based on some random occurrence that happens over the weekend.
0:07:32 Like result doesn’t go their way.
0:07:33 You’re fired.
0:07:34 Good luck to you.
0:07:36 Which is not the way analytics works, right?
0:07:39 Like the way analytics works is you need a large end.
0:07:43 You need to do the thing a hundred times and then 60 times it’ll go the way the analytics
0:07:44 says it.
0:07:45 40 times it won’t.
0:07:48 So that’s just the nature of the probabilistic world we live in.
0:07:49 Yeah.
0:07:54 And analytics is also all about separating process from outcomes and yet these decisions
0:07:56 are still being made on outcomes.
0:08:00 So I have a lot of empathy for people who are in this situation.
0:08:05 And a coach will get fired more likely if they do the thing that the analytics says that’s
0:08:08 contrary to the sort of conventional wisdom in the sport, right?
0:08:13 It’s like why in American football coaches for a long time didn’t go for it enough on
0:08:14 fourth down, right?
0:08:18 Yeah, analytics clearly said they should, but everybody would get pissed at them if they
0:08:19 went for it and didn’t get it.
0:08:21 Now that’s changed, right?
0:08:24 Because of analytics, interestingly, although was it this last Super Bowl where the coach
0:08:28 kept going for it on fourth down and not getting it and people were like, “He did it too much.”
0:08:30 I was like, “No, he didn’t!
0:08:33 Just because it didn’t work out doesn’t mean it was the wrong decision.”
0:08:34 Yeah.
0:08:38 And so, you know, this just kind of goes back to like you need to have kind of top-top
0:08:42 decision makers saying like, “Yeah, it’s okay if you’re going to do something that looks
0:08:44 a little bit unconventional.
0:08:45 We believe in it.
0:08:46 We trust it.”
0:08:48 So, okay, so we’ve talked about why it’s hard.
0:08:52 Did you figure anything out about how to get people to change their decision?
0:08:57 I would say the best thing would be building trust through a common language.
0:09:00 And so for us, the common language was video.
0:09:05 So we could build a model, but like until we could show them the model outputs on video
0:09:10 and walk them through it and say, “This is what the model sees.
0:09:12 This is what the model predicts.
0:09:13 What do you think?
0:09:14 Like walk me through it.”
0:09:19 And so then it became kind of like these collaborative, iterative, model-building processes.
0:09:24 It’s sort of a version of having an idea and then convincing your boss that that idea is
0:09:26 actually their idea.
0:09:28 And once they think it’s their idea, then they’ll do it.
0:09:29 It sounds like that.
0:09:30 It is.
0:09:32 And like maybe a little bit less cynical, but yeah.
0:09:33 Yeah.
0:09:34 Yeah.
0:09:37 So eventually you left Arsenal and you started Source Football, this company that you’re
0:09:39 running now.
0:09:41 What led you to make that leap?
0:09:47 So part of starting this company was that I wanted to learn and to experience as much
0:09:48 as possible.
0:09:55 But the other part is that we felt like there’s a huge need for clubs to get help in terms
0:10:00 of getting started on this analytics journey where a lot of them just don’t even know where
0:10:01 to begin.
0:10:04 They don’t know what’s good, what’s bad.
0:10:10 So our kind of main value-add is helping them get set up, helping them get started.
0:10:16 And then once that happens, doing all of the hard work, hard thinking, that’s impossible
0:10:19 to get done within a football club.
0:10:24 If anyone has ever been in a training ground, I’m sure it’s the same across all sports,
0:10:31 but they’re chaotic, they’re noisy, they’re loud, they’re not good places to do deep,
0:10:32 hard things.
0:10:33 A lot of testosterone.
0:10:39 Yeah, a lot of really loud, bad music, like humping.
0:10:42 If your office is anywhere near the gym, forget it.
0:10:48 You’ve got to wait until everybody goes home before you can actually have a clear thought.
0:10:51 That was one of the things that we realized is you can’t work on solving these really
0:10:53 hard problems.
0:10:57 Football is really hard to analyze in terms of analytics.
0:10:59 You can’t do it within a club.
0:11:01 So that’s why we wanted to be a little bit outside.
0:11:07 And then kind of our long-term vision is actually we’re using this consulting business to kind
0:11:13 of fund the development of our intelligence platform and then the idea is eventually to
0:11:19 look for a set of investors or maybe an ownership group and take control of the club and run
0:11:24 it in kind of like the modern progressive way that we think they should be run.
0:11:29 So what you really want to do is basically buy a soccer team and run it smarter than
0:11:30 everybody else.
0:11:31 Is that what you’re telling me?
0:11:32 Yeah, of course.
0:11:34 Tell me more about the big dream.
0:11:38 We’ll do the pieces, but I’m curious, like that’s a big audacious dream and it’s fun.
0:11:39 Tell me about it.
0:11:42 Yeah, I mean, I think, you know, one of the things that probably everybody in analytics
0:11:47 has experienced is that unless you’re kind of like the key decision maker, it’s always
0:11:53 going to be kind of difficult to influence decisions because there’s always going to
0:11:58 be somebody that has, you know, kind of their own perspective and everything.
0:12:02 And, you know, at Arsenal, we were really lucky because we had like a lot of resources.
0:12:07 But what I see at a lot of clubs is that, you know, they only hire one or two people.
0:12:13 The work that they do is good, but like there’s a limit to what two human beings can do.
0:12:17 And so like a lot of what they produce doesn’t necessarily like match the gut.
0:12:21 And so it’s hard to get this buy-in.
0:12:24 And so then they just say like, well, okay, thanks for that.
0:12:26 I’m only going to listen to you if I agree with it.
0:12:28 If I disagree with it, I’m going to go with my gut.
0:12:31 And so it just becomes difficult.
0:12:36 But I think also like there’s so many issues within football clubs that go beyond just
0:12:39 analytics or, you know, making the right decision on players.
0:12:46 I mean, there’s so much in terms of like building good, good cultures, you know, kind of making
0:12:50 sure that everything is kind of set up in like a professional way.
0:12:53 Because if you think about it, clubs are kind of run by people who’ve never worked outside
0:12:59 of football, but I think there’s a lot of lessons that you can take from working outside
0:13:02 of sports, bring them into into football.
0:13:08 And we’ve seen it in every other major American sport, not to be like super ruthless and say
0:13:12 like, we got to maximize profit, maximize profit, because I think like they are these
0:13:17 weird social institutions that have a lot of meaning to a lot of people.
0:13:19 And so you obviously want to respect that.
0:13:23 And you also have these amazing platforms to bring positive change into the world.
0:13:25 And so you want to take advantage of that as well.
0:13:31 But like there certainly can be a lot more kind of professionalism in them than what we
0:13:33 experience at a lot of places.
0:13:38 So now you have this company and you want to use it eventually to take over the world.
0:13:39 Correct.
0:13:44 Before you use your company to take over the world, like what are the what are the services
0:13:46 you’re selling to teams, to clubs?
0:13:52 Yeah, I mean, so clubs are in really different situations.
0:13:56 And then also like, you know, it’s not like the U.S. where like every single major league
0:14:00 baseball team is like loaded with cash.
0:14:04 Typically in European leagues, like you’re going to have several divisions.
0:14:10 So like you’re talking about anywhere from like a single A baseball team to like a major
0:14:14 league baseball team, like there’s huge difference in resources.
0:14:20 So like orders of magnitude in terms of revenue, how much presumably they want to spend on
0:14:21 your services, etc.
0:14:23 Yeah, exactly.
0:14:27 And so like the main area where we can help them is really recruitment.
0:14:30 So helping them kind of find the best players.
0:14:35 So unlike American sports where you kind of like trade players and draft them, everything
0:14:39 here is kind of like an open market where you can buy and sell the contracts of these players.
0:14:43 And so if you’re a small team and you have a really, really good player, like you can
0:14:49 just sell him, keep that cash, or you can invest it back into your club.
0:14:52 This is this is really echoes of of Moneyball.
0:14:54 I mean, I’m sure I don’t know if you’re tired of hearing about Moneyball.
0:14:58 That book came out 20 years ago, but like that was the basic idea there, right?
0:15:02 It was better, better scouting, essentially, right?
0:15:05 Like you had these scouts who were sort of using their guts and these kind of conventional
0:15:07 wisdom, heuristics.
0:15:12 And then you had what I’m sure now seem like primitive analytics coming in and basically
0:15:16 doing a better job of predicting player success, right?
0:15:17 At some level.
0:15:18 Yeah, exactly.
0:15:22 And so, you know, I hate to say that like soccer is 20 years behind baseball, but like
0:15:24 in a lot of ways we are.
0:15:29 And so like this is still like the main area of value add for a lot of clubs because you’re
0:15:34 competing in this open market with people with varying amounts of knowledge.
0:15:39 So sometimes that knowledge is just, hey, like we had scouts at that game, we’ve seen
0:15:43 this player once, like this is our opinion of him.
0:15:46 And then you have like the more sophisticated teams that are like, okay, well, we have a
0:15:50 database of, you know, 600,000 matches in the world.
0:15:55 We know the 50 best players in every league at every position.
0:15:57 We know how much they’re worth.
0:16:02 So we help teams kind of go from like the former to the latter and just be a little bit more
0:16:05 sophisticated in terms of the amount of information that they have.
0:16:09 I mean, it’s sort of crass to put it this way, but it’s really like pricing assets.
0:16:10 Right.
0:16:11 Like the players.
0:16:13 And again, I realize this is kind of dehumanizing.
0:16:14 I’m sorry.
0:16:22 But it is analogous to people valuing a stock or valuing anything they might buy, right.
0:16:28 And the better you can model the value of the asset, the better you’re going to be at
0:16:30 finding mispriced assets, right.
0:16:33 You want to, you want to find bargains.
0:16:37 You want to, you want to get the most you can for your, for your dollar.
0:16:38 Yeah.
0:16:39 Absolutely.
0:16:44 And so, you know, there’s a lot of techniques that we can take from other industries because
0:16:47 it’s not that different from it where it, where it gets hard is that you have to have
0:16:50 those assets work well together.
0:16:51 They are in fact human beings.
0:16:52 They are.
0:16:53 Yes.
0:16:56 That is, you know, the, the kind of joke where it’s like, well, the problem with football
0:16:59 clubs is that it’s full of human beings.
0:17:01 So let’s talk about what you can model.
0:17:06 Like, what are you good at modeling in, in this context, like in helping, in scouting,
0:17:10 essentially, in helping clubs value, you know, players they might acquire.
0:17:11 Yeah.
0:17:16 So we’re pretty good at modeling everything that happens when a player has the ball.
0:17:21 And unfortunately in football, that’s like a very, very small portion of the game.
0:17:22 Yeah.
0:17:27 And so there was this really cool development in the last four or five years where a number
0:17:33 of companies have come out with a data source that’s basically taking the video feeds from
0:17:38 TV and turning that into tracking data.
0:17:43 So basically they’re able to track the location of every player that’s on screen.
0:17:45 So obviously you, you don’t know all the players.
0:17:51 And then, you know, what we’ve learned is that really what’s going on off screen tends
0:17:56 to not be as relevant and the location of those players isn’t as relevant.
0:18:01 And so now all of a sudden you can start doing modeling on what people are doing off the
0:18:06 ball and use it for recruitment because prior to this, similar to the NBA, there were cameras
0:18:13 in all the stadiums and you would get this like full data set, but only for your league.
0:18:14 And so you couldn’t really use it for recruitment.
0:18:16 And so this has been like a really big change.
0:18:23 So, you know, we can get much better views into what is a player doing physically?
0:18:28 What are they doing defensively in terms of cutting off passing lanes, things like that?
0:18:31 There’s still a lot we’re not good at with that.
0:18:35 But what it’s also done is allowed us to say, well, if we’re going to move a player from
0:18:40 this league to that league, how different is it physically?
0:18:41 How do we think they’re going to adapt?
0:18:44 And so this has really opened up a lot of things.
0:18:49 So if we want to talk about like risk adjusted pricing of assets, now we’re starting to be
0:18:52 able to quantify a little bit.
0:18:56 Like what’s the risk of bringing in somebody from a really, really different environment
0:19:00 into this one versus one that’s more similar or…
0:19:06 So you should apply a larger discount when you’re bringing a player from a very different
0:19:08 league, presumably because you’re taking more risk.
0:19:09 Yes.
0:19:10 Yes.
0:19:15 And if not, then it’s not an undervalued asset and maybe walk away.
0:19:22 So tell me more, like is the output you have just here is what you should pay for each
0:19:23 of these players?
0:19:28 Like you have whatever, 10,000 players or something and you put a dollar value on each
0:19:30 one or like, what’s the output?
0:19:31 Yeah.
0:19:32 I mean, I wish it was that simple.
0:19:37 We’re still like pretty far from putting it all together and saying like, buy this player
0:19:44 at this price, you know, a lot of the difficulties is that there’s no good data set on pricing
0:19:45 information.
0:19:48 So then what do you just have a relative like kind of a stack rank?
0:19:51 I mean, there are like estimates.
0:19:57 So in the media, they’ll say like, oh, this guy went for 10 million, but then a different
0:20:02 media source will say this guy went for 15 million because like the selling club wants
0:20:07 to report a higher price and the buying club wants to report lower price.
0:20:15 So there’s real no truth and then salaries for players are not public for most leagues.
0:20:16 So okay.
0:20:19 So that’s a very important variable that you don’t really have access to.
0:20:20 That’s a problem.
0:20:22 So what is the output of your model then?
0:20:23 Yeah.
0:20:29 I mean, so we kind of do like a stack ranking and like a, you know, projected like, what
0:20:31 do we think this player would do in this situation?
0:20:35 But like the error bars on these things are like are pretty big.
0:20:40 So we’re still kind of in the like subjective realm of like based on these factors.
0:20:42 We think this or we think that.
0:20:48 And then, you know, a lot of the like the markets change every year in football as well.
0:20:53 So even if we had like a really precise model that said this guy should be worth 5 million,
0:20:56 well, he should, he’d be worth 5 million last year.
0:20:57 Yeah.
0:21:02 I appreciate your candor, you know, so the market changes, you don’t have pricing data.
0:21:06 There’s big uncertainty even on the outputs you do have.
0:21:12 Like all of this seems, so what is the use of what you’re doing?
0:21:14 How is it valuable to people?
0:21:19 That’s the point is that we’re competing with people who, you know, have gone to maybe
0:21:26 three or four games, watch this guy in, you know, unknown conditions.
0:21:33 The human brain is like not conditions to be like very objective.
0:21:38 And so like, you know, there’s a host of like very known biases that like people can can
0:21:39 fall for.
0:21:46 And so really what we’re trying to do is just give you like a much more fair and even view
0:21:52 of a player taking into account a lot of, you know, the factors that these scouts are
0:21:58 trying to account for as well, but just doing it more objectively, more rigorously, you
0:22:01 know, over a longer period of time.
0:22:08 It seems from what you’ve been saying, like soccer is behind certainly baseball and perhaps
0:22:09 other sports in analytics.
0:22:13 I mean, your first paper, you were following somebody who had written a paper in American
0:22:16 football and you’re like, oh boy, if we do that for soccer, is it true that soccer is
0:22:17 behind it?
0:22:18 And if so, why?
0:22:19 Yeah, it’s true.
0:22:23 And I think, you know, the debate used to be like, well, are we making any progress?
0:22:29 I think we’ve made a ton of progress, but we’re still really far behind other sports.
0:22:34 And so one answer is that we also don’t get the funding that other sports get.
0:22:37 So so the level of investment just isn’t there.
0:22:44 And so if we were behind 15 years ago, we’re certainly not keeping pace, you know, and then
0:22:47 I think there’s other structural issues with it.
0:22:53 So I love to use this image that somebody else made, but it shows the relative size
0:22:56 of a basketball court to a soccer pitch.
0:23:01 And basically a basketball court can fit into like the little tiny penalty area on a soccer
0:23:02 pitch.
0:23:08 And so, you know, the distances and spacing of players is so much more variable in soccer.
0:23:13 So you can’t say like, oh, 10 meters is a good distance between me and a teammate because
0:23:18 it depends on the situation that’s happening where on the pitch, like, is it a transitional
0:23:19 moment?
0:23:20 Is it kind of a more controlled moment?
0:23:25 And so there’s a lot of complexities like that, you know, the game doesn’t stop.
0:23:29 There aren’t a lot of these set pieces where it’s like we have a very choreographed idea
0:23:32 of what we want to do.
0:23:38 And then, you know, the sad reality is that most leagues only play 38 matches a season.
0:23:43 And so you never see a team in the same context twice.
0:23:46 You either play them at home or away, and then that’s it.
0:23:50 Or maybe you play them in a cup game, which is like a really different environment.
0:23:55 And so all of these things just kind of like add up and it’s like, well, we have fewer resources.
0:23:57 It’s much more complex.
0:24:01 Like of course we’re behind on this.
0:24:09 Still to come on the show, the one big strategic change that Sarah really, really wants someone
0:24:16 in soccer to try and why nobody has tried it yet.
0:24:27 The Medal of Honor is the highest military decoration in the United States, awarded for
0:24:33 gallantry and bravery in combat at the risk of life above and beyond the call of duty.
0:24:42 Since it was established in 1861, there had been 3,517 people awarded with the medal.
0:24:47 I’m Malcolm Gladwell, and our new podcast from Pushkin Industries and iHeartMedia is
0:24:53 about those heroes, what they did, what it meant, and what their stories tell us about
0:24:57 the nature of courage and sacrifice.
0:25:02 Without him and the leadership that he exhibited in bringing those boats in and assembling them
0:25:09 to begin with and bringing them in, I saved a hell of a lot of lives, including my own.
0:25:15 Listen to Medal of Honor Stories of Courage on the iHeart Radio App, Apple Podcasts, or
0:25:26 wherever you listen to podcasts.
0:25:33 So we talked a lot about essentially scouting, evaluating players as one of the big things
0:25:35 you do.
0:25:36 What else do you do?
0:25:37 Yeah.
0:25:43 I mean, so there’s a number of different services that we provide, so it could be doing retrospectives
0:25:49 of how did the team play this weekend, or maybe some kind of, I guess in U.S. terms they
0:25:54 would call it advanced scouting, but delivering kind of like a data profile on their upcoming
0:25:59 opponent, and then going into like on-field stuff, like we’re doing a lot of research
0:26:06 now around, you know, various things in terms of, you know, how can we maximize set pieces.
0:26:10 A set piece is like a play analogous to sort of a play.
0:26:11 Yeah.
0:26:13 So, you know, football is really fluid.
0:26:19 It never really stops, except for these certain moments that are called dead balls.
0:26:23 And when a dead ball happens and kind of like the attacking area.
0:26:26 And you have like a throw-in or a penalty kick or something.
0:26:27 Yeah.
0:26:28 I mean, penalties are pretty straightforward.
0:26:29 They’re kind of exceptional.
0:26:32 It’s like, well, just kick it into the goal.
0:26:39 But corner kicks are kind of like the most common one, because they all happen from the
0:26:40 same location.
0:26:41 They happen fairly frequently.
0:26:43 You can analyze them.
0:26:44 You can prepare for them.
0:26:49 So you have this set piece opportunity, and then it’s, you know, what’s the strategy we’re
0:26:52 going to use to maximize it?
0:26:54 So a lot of research into that.
0:26:58 Has analytics in a general way changed the way people take corners?
0:26:59 Yeah.
0:27:06 I mean, I think it has influenced how much time they spend thinking about it and preparing
0:27:07 for it.
0:27:12 So that’s another one, Ted Knudsen, who is the CEO of Statsbom.
0:27:18 He’s been shouting this from up high for the longest, but set pieces for a long time were
0:27:23 really valuable, but teams would only train them for like 10 minutes a week.
0:27:25 It’s an interesting level of analytics.
0:27:26 That’s an analytical claim, right?
0:27:31 It’s like your training time is measurable, and you should be allocating the right proportion
0:27:32 of it to the right things.
0:27:37 And he’s essentially arguing that you’re underfunding in time corners.
0:27:38 Yeah.
0:27:39 Yeah.
0:27:43 Like the number of goals scored from corners or conceded from corners was not at all proportional
0:27:48 to the amount of time spent training them.
0:27:54 And so that’s been a big focus on teams lately, and I still don’t think they spend enough
0:28:00 time training it, but now you’ll see certain clubs even hiring set piece specialists.
0:28:05 So they’ll have a member of coaching staff whose job, whose only job is really thinking
0:28:08 about these things and helping prepare the players for it.
0:28:11 Is that person basically the corner kick coach?
0:28:15 I mean, they, you know, there is a guy who’s like a throw-in coach as well.
0:28:19 So you can even have like some specialties to them, but…
0:28:21 That guy dreams of being the corner kick coach.
0:28:22 Yeah.
0:28:24 Gets promotion too.
0:28:30 So I mean, just thinking more generally, like it seemed, well, other sports here, my U.S.
0:28:36 centric sports knowledge, which itself is somewhat limited, like clearly analytics have
0:28:42 changed, you know, football, you see people going for it on fourth down, more basketball.
0:28:46 Obviously the demise of the mid-range shot, right?
0:28:50 Somebody just realized analytically some years ago, like, don’t shoot from whatever 12 feet
0:28:51 out, right?
0:28:52 Shoot from under the basket or take a three.
0:28:54 And then baseball, they was the shift.
0:28:59 Like there are these big, big changes in the way the games look because of analytics.
0:29:01 Is there anything analogous in soccer?
0:29:06 Yeah, probably like the earliest one was that the distance from which people are shooting
0:29:08 has changed.
0:29:13 So you’re not going to see a Charlie Adam taking a shot from 40 yards too often unless,
0:29:17 you know, there’s like a, oh, the goalkeeper is off his line, he’s out of position, I can
0:29:18 get it in there.
0:29:20 Is that because of you?
0:29:24 You know, I can’t take credit for it.
0:29:27 It feels like, you know, the invention of calculus where there were a lot of people kind of coming
0:29:28 to this conclusion.
0:29:33 So you’re saying you might not be Newton, but if you’re not Newton, you’re Leibniz?
0:29:34 Yes.
0:29:35 Okay, fair enough.
0:29:39 Yeah, that’s probably like the earliest one.
0:29:46 And then again, like, you can’t prove causality, but there’s been a shift to this tactic of
0:29:48 teams putting a lot of pressure up high.
0:29:54 So if they have the ball in the attacking area of the pitch and they lose it, they immediately
0:29:58 put pressure on the other team to try to win it back as quickly as possible.
0:30:00 Kind of a full court press.
0:30:01 Yeah, exactly.
0:30:05 And so I don’t know if the origin of that is based in analytics, but like analytics will
0:30:11 tell you, yeah, you’re more likely to score a goal if you win the ball back up high.
0:30:12 So that’s another one.
0:30:13 Huh.
0:30:14 Why do you think people didn’t do it before?
0:30:17 Is it not intuitively obvious that that’s good?
0:30:19 Yeah, because if you don’t want to…
0:30:20 Are you so tired?
0:30:21 Well, yeah.
0:30:25 So one, it makes you tired and there is like a risk of injury trying to kind of like make
0:30:28 the guys super fit to do this.
0:30:33 But two, kind of going back to like the fourth down analogy, if you don’t win the ball back,
0:30:38 then you’ve committed so many players upfield that you’re going to concede a kind of like
0:30:40 silly looking opportunity.
0:30:41 It’s risky.
0:30:42 It’s risky.
0:30:43 Yeah, it’s a calculated risk, but it’s risky.
0:30:46 The expected value is positive, but the variance is high.
0:30:47 Exactly.
0:30:50 What are some interesting frontier problems?
0:30:53 I mean, they’re sort of getting people to listen to you, which we’ve talked about, but
0:31:00 just on the analytics side, like what are you trying to figure out on the analytics side?
0:31:03 What’s a big problem you’re trying to solve?
0:31:08 One area that we haven’t really exploited yet, but I’m curious about, but we always have
0:31:13 this problem that like we can’t always learn from the data because everybody’s doing the
0:31:14 same thing.
0:31:17 You can’t really do an A/B test because everybody does A.
0:31:19 Yeah, exactly.
0:31:20 And so it’s really frustrating.
0:31:24 I mean, like the biggest example would be substitution patterns.
0:31:28 Everybody makes roughly the same type of substitution at the same time.
0:31:31 And presumably it’s not optimal, right?
0:31:34 It’s just conventional wisdom, but you’re like, “There could be a whole better world
0:31:35 that nobody’s ever tried.”
0:31:36 Yeah, exactly.
0:31:41 And so I think, you know, with generative AI, like now you’re kind of talking about like,
0:31:45 “Well, can we do really smart, really realistic simulations on these sorts of things?”
0:31:47 Kind of synthetic data.
0:31:48 Yeah.
0:31:49 Yeah.
0:31:50 That’s what we’re hoping.
0:31:54 So are you trying to figure out a better way to do substitutions?
0:31:55 Yes.
0:31:56 Yeah.
0:31:57 You feel like you got it?
0:32:00 Do you feel like you have one in your pocket or you’re not quite ready to say?
0:32:01 I mean, we have some theories.
0:32:05 We haven’t proven them, but like we also need somebody to say like, “Yeah, go ahead.
0:32:07 Like mess with our substitution patterns.
0:32:08 We feel comfortable with this.”
0:32:09 Uh-huh.
0:32:10 Uh-huh.
0:32:14 It is amazing how, I mean, like if you think of the shift in baseball or whatever, like
0:32:20 a game can just exist for a hundred years and there can be a way, better way to do it.
0:32:25 And nobody ever does it just because they don’t have the imagination or the courage.
0:32:26 Yeah.
0:32:27 It’s wild.
0:32:31 And I think, you know, that sort of thing, like I hope in soccer is less common because
0:32:35 it’s more of like an adversarial, uh, game, but-
0:32:37 But it maps to the world more generally, right?
0:32:42 Like, I think a lot of it is just fear, like you won’t get in trouble if you do the thing
0:32:45 everybody else did and you have a bad outcome.
0:32:48 Like you see it in, you know, in finance certainly, right?
0:32:54 Like financial advisors just exhibit herd behavior because of the way the incentives
0:32:55 are structured.
0:32:57 Like if you do what everybody does and you lose money, you’re like, “I was just doing
0:32:59 what everybody else did.”
0:33:00 Yeah.
0:33:01 Yeah.
0:33:02 And it’s really frustrating.
0:33:07 And so, um, you know, I think that again goes back to like why we want to get control
0:33:15 of a club because, you know, I think there is a lot of ways that we can optimize things.
0:33:20 And if we’re only reporting to ourselves, then it’s like, well, you know, I’m not going
0:33:23 to fire myself just because the theory didn’t work out.
0:33:24 Yes.
0:33:26 It’s hard to get to a big enough end though, right?
0:33:30 It’s hard to get to a big enough sample size or just aren’t that many games.
0:33:35 Um, so, so let’s talk about the sort of happy outcome for you.
0:33:41 Say it’s whatever, five years from now, and you have found capital, you’ve found somebody
0:33:44 who wants to essentially bet on you, which is, if I understand correctly, what would
0:33:46 have to happen.
0:33:53 And you are, you know, with your finance-y, finance-y or partners running a club.
0:33:54 What’s that look like?
0:33:55 Yeah.
0:33:59 I mean, I think it would look really, really different.
0:34:03 Um, I think that the typical profile of who would be working in that club would be quite
0:34:04 different.
0:34:05 Yeah.
0:34:06 Oh, yeah.
0:34:07 Nerdier.
0:34:08 For sure.
0:34:09 For sure.
0:34:14 It’s hard because like, uh, you know, I’ve played a lot of sports growing up and so it’s
0:34:16 like, well, you know, I’m not that nerdy.
0:34:19 And then it’s like, oh, yeah, yeah.
0:34:25 Um, but yeah, I mean, I think it would look nerdier, but I think it would be a lot more
0:34:29 like just a totally different mindset, really a lot more people who are like, yeah, let’s
0:34:31 do things differently.
0:34:35 Let’s have the courage to, to try new things.
0:34:39 And then let’s have kind of the, the thinking power behind it to not just like, let’s try
0:34:45 random things, but let’s try really well thought out, creative, but courageous things.
0:34:50 When you talk about like, you know, being bold and creative and different, is there
0:34:53 some particular thing you have in mind?
0:34:57 Is there’s like one thing where you’d like that one thing, I wish somebody would just
0:34:58 try it.
0:35:02 I mean, right now it’s probably the substitution thing because that’s like the lowest hanging
0:35:06 fruit and it like really, really drives me nuts.
0:35:08 What’s your secret theory?
0:35:09 What do you think would be better?
0:35:10 Oh, I think you should do early.
0:35:14 I think you should kind of treat it like line changes in hockey.
0:35:19 So I mean, like right now what they do is they, they typically try to bring on maybe like a
0:35:24 fast player late in the game that, you know, would theoretically be running against tired
0:35:25 legs.
0:35:28 But the problem is that like it’s too late.
0:35:33 And so, you know, our idea is basically like bring on people who like know that they only
0:35:39 have 30 minutes, run as, as hard as you can, as crazy as you can, get, get that one goal
0:35:42 advantage or two goal advantage and then adapt.
0:35:44 And nobody has tried that.
0:35:45 That doesn’t seem crazy.
0:35:48 Like there’s a hundred soccer teams.
0:35:51 All you need is one person to try and it’s like a free idea.
0:35:52 I know.
0:35:57 Well, you know, if we see like a change in substitution patterns after this, that’s right.
0:35:58 Yeah.
0:36:02 But, you know, like the crazy thing is there used to only be three substitutions in soccer.
0:36:06 And so you always wanted to keep at least one in case a player gets injured.
0:36:08 So that really leaves you with two.
0:36:11 But during COVID, they upped it to five.
0:36:14 And so now it’s like, well, you have two, three, and that has persisted.
0:36:15 Yeah.
0:36:19 So they’ve almost doubled the number of substitutions allowed.
0:36:21 Has the strategy and using them changed?
0:36:22 No.
0:36:25 It’s like a very slightly change.
0:36:30 And a lot of coaches don’t even take advantage of all substitutes.
0:36:36 Why do you think nobody has, you know, dramatically significantly changed their strategy around
0:36:38 substitutions, even though the rule has changed so significantly?
0:36:39 Yeah.
0:36:41 One, I mean, I guess there’s like two explanations.
0:36:45 The easiest one is like, well, it’s always been the way that we’ve done it.
0:36:48 So, you know, why change?
0:36:53 I think part of what’s driving that, though, is that the way that rosters are constructed
0:36:55 hasn’t changed.
0:37:00 And so, you know, depending on like the finances of a team, you might not have as many like
0:37:05 quality players to go that deep onto the bench.
0:37:09 But you know, if this is your strategy, then you can change how you recruit.
0:37:12 You can change the profile of that squad composition.
0:37:16 The second answer is interesting, right, because it requires you to think more systematically.
0:37:17 It’s like, oh, the rules are different.
0:37:21 So therefore, we should build a different team, which is kind of next level.
0:37:22 Yeah.
0:37:29 And yeah, that’s not typically how teams operate.
0:37:33 We’ll be back in a minute with the lightning round.
0:37:47 The Medal of Honor is the highest military decoration in the United States, awarded for
0:37:53 gallantry and bravery in combat at the risk of life above and beyond the call of duty.
0:38:01 Since it was established in 1861, there had been 3,517 people awarded with the medal.
0:38:06 I’m Malcolm Gladwell, and our new podcast from Pushkin Industries and iHeartMedia is
0:38:12 about those heroes, what they did, what it meant, and what their stories tell us about
0:38:16 the nature of courage and sacrifice.
0:38:22 Without him and the leadership that he exhibited and bringing those boats in and assembling
0:38:27 them to begin with and bringing them in, I saved a hell of a lot of lives, including
0:38:28 my own.
0:38:34 Listen to Medal of Honor Stories of Courage on the iHeart Radio app, Apple Podcasts, or
0:38:45 wherever you listen to podcasts.
0:38:50 We’re going to finish with the lightning round, soccer or football?
0:38:56 I say football, fries or chips, cookies or biscuits.
0:39:00 So my husband is from India, and he will kill me if I say cookies, but cookies.
0:39:02 Are there other Britishisms?
0:39:03 You sign your emails.
0:39:04 Cheers.
0:39:07 Sometimes I hate doing it.
0:39:12 Mate is another one because that’s you call everything brilliant.
0:39:15 I love talking to British people because they say everything I say is brilliant, but they
0:39:16 don’t mean it the way I think.
0:39:21 Yeah, brilliant, bloody, yeah, just little things like that.
0:39:22 Blimey?
0:39:23 British guy once said blimey to me.
0:39:24 I loved that.
0:39:27 Yeah, I’ve not done that.
0:39:31 But yeah, I think when I was little, my brother and I used to antagonize my dad until he
0:39:37 would yell at us in British slang, and that was success for us.
0:39:39 Give me your impression of your dad yelling at you in British slang.
0:39:42 He would just say, “Oh, bugger off.”
0:39:46 A bugger off is good.
0:39:50 What’s one thing you learned working at Microsoft?
0:39:53 It’s hard to get things done in large companies.
0:39:58 Do you have any tips for getting things done in large companies, or is it just leave?
0:40:02 I mean, my answer was leave.
0:40:07 But yeah, I mean, I think there’s certain behaviors that maybe I don’t necessarily possess
0:40:13 or love, but aggressive, loud people tend to get things done there.
0:40:16 Were you there in the Ballmer area?
0:40:17 Oh, yeah.
0:40:18 Ballmer era?
0:40:19 That’s very Steve Ballmer vibes.
0:40:26 Who’s the most underrated player you’ve ever seen?
0:40:29 It might be this guy named Manu Trigueros.
0:40:36 He plays for a club in Spain called Virial, and I think he’s brilliant.
0:40:41 He was kind of born at a time when Spain had loads of really, really talented players in
0:40:43 his position.
0:40:46 He never got called up to the national team or anything like that.
0:40:48 But his running style is a little bit nerdy.
0:40:51 He runs very much on his toes.
0:40:55 He doesn’t look like an athlete.
0:41:00 He actually is like, while he was playing professionally, he was doing his masters in education.
0:41:02 So he was like a student teacher.
0:41:03 It’s just that you love him.
0:41:06 It’s not that he’s an amazing player, it’s that you love him?
0:41:10 I love him, but he was also an amazing player.
0:41:16 But I think having this air of nerdiness around him kind of kept him a little bit under the
0:41:17 radar.
0:41:26 Who’s the most overrated player you’ve ever seen?
0:41:28 That’s a really hard one.
0:41:30 Are you afraid of getting in trouble?
0:41:33 Is there a name in your mind and you just don’t want to say it because you don’t want
0:41:34 to antagonize anybody?
0:41:35 Yeah.
0:41:36 Definitely.
0:41:37 There’s some of that.
0:41:40 And then there’s like a little bit of like hindsight bias where it’s like, there were
0:41:46 guys who were overrated at the moment and then like kind of crumbled and failed.
0:41:53 So it’s like, well, if I say like Marwan Shamak, who was a kind of like notoriously disastrous
0:41:56 signing for Arsenal, everyone would say like, oh yeah.
0:41:57 Too easy.
0:41:58 Yeah.
0:42:06 What’s one piece of advice you have for women working in male dominated fields?
0:42:10 That’s a really good one.
0:42:15 It would probably be that just you belong and know that you belong and don’t let people
0:42:19 try to make you feel that you don’t.
0:42:23 And one of the ways that you can do that is talk to other women.
0:42:27 When I got started 15 years ago, like there weren’t any other women.
0:42:32 Now there’s loads of women working in all different sports and all different sorts of
0:42:34 roles.
0:42:39 And a couple of years ago, this conference started called Women in Sports Data.
0:42:43 This year it’s going to be held in Philadelphia, September 7th, I think.
0:42:48 But that’s like a great place to connect with people because to me, I find it really powerful
0:42:54 to be in like a gymnasium with like a room literally full of women who are interested
0:42:57 in sports and data and technology.
0:43:02 And yeah, just, I don’t know, it’s good for your mental health.
0:43:05 It’s good for, you know, your self-esteem and everything.
0:43:11 But yeah, just know that you belong, that you can do it and that, yeah, nobody should
0:43:13 tell you otherwise.
0:43:16 So Brad Pitt played Billy Bean and Moneyball.
0:43:18 Who’s going to play you in Moneyball 2?
0:43:21 Don’t call it soccer.
0:43:23 Oh, yeah.
0:43:25 I have no idea.
0:43:30 I don’t watch a lot of movies, so like I couldn’t even name an actress right now.
0:43:34 Maybe Kira Knightley because she was in Bend at Lake Beckham.
0:43:38 So I’ll go with Kira Knightley.
0:43:43 Sarah Rudd is the co-founder and CEO of Source Football.
0:43:46 Today’s show was produced by Gabriel Hunter Chang.
0:43:51 It was edited by Lydia Jean Cotte and engineered by Sarah Brugger.
0:43:54 You can email us at problem@pushkin.fm.
0:43:59 I’m Jacob Goldstein, and we’ll be back next week with another episode of What’s Your Problem?
0:44:17 The Medal of Honor is the highest military decoration in the United States.
0:44:25 Since it was established in 1861, there have been 3,517 people awarded with the medal.
0:44:29 I’m Malcolm Gladwell, and our new podcast from Pushkin Industries and iHeartMedia is
0:44:35 about those heroes, what they did, what it meant, and what their stories tell us about
0:44:38 the nature of courage and sacrifice.
0:44:44 Listen to Medal of Honor stories of courage on the iHeart Radio app, Apple Podcasts, or
0:44:46 wherever you listen to podcasts.
0:00:12 Since it was established in 1861, there have been 3,517 people awarded with the medal.
0:00:17 I’m Malcolm Gladwell, and our new podcast from Bushkin Industries and I Heart Media is
0:00:19 about those heroes.
0:00:25 What they did, what it meant, and what their stories tell us about the nature of courage
0:00:26 and sacrifice.
0:00:32 Listen to Medal of Honor Stories of Courage on the iHeart Radio App, Apple Podcast, or
0:00:44 wherever you listen to podcasts.
0:00:57 Around 2010, Sarah Rudd was rooting for this British soccer team.
0:00:59 It was a team called Blackpool.
0:01:01 It had been her great grandfather’s favorite team.
0:01:03 She was really into it.
0:01:06 But she started getting annoyed by this one player.
0:01:08 His name was Charlie Adam.
0:01:14 She just loved to take shots from really long distance, and it drove me crazy.
0:01:19 Every possession just kind of ended with him taking these random shots, and it’s like,
0:01:20 “What a waste.
0:01:21 What are you doing?”
0:01:25 Sarah was living in Seattle, working for Microsoft as a software engineer.
0:01:28 But she really wanted to get into sports analytics.
0:01:32 Basically, she wanted to do moneyball but for soccer.
0:01:38 And then she discovered that this sports data company called StatDNA was hosting a contest.
0:01:44 A contest where people could use the company’s soccer tracking data to generate analytical
0:01:45 insights.
0:01:48 And she thought, “This could be my chance.
0:01:51 This could be my chance to break into the industry.”
0:01:57 And also, at the same time, that guy, Charlie Adam, really should not be taking those long
0:01:58 shots.
0:02:04 I’m a big fan of searching for inspiration elsewhere, so I started just looking around
0:02:07 and seeing, “Well, what have other people done in this space?”
0:02:10 And it turns out for soccer, not much.
0:02:14 But there was some really interesting papers in other sports, and the one that caught my
0:02:17 eye was by a gentleman named Keith Goldner.
0:02:24 He had done something using NFL data, where he looked at kind of this, “If the ball is
0:02:29 on the 35-yard line, it’s second and 10 with X amount of time on the clock.
0:02:32 How likely are you to score?”
0:02:36 And I thought that was really interesting, and I thought, “Well, isn’t that kind of
0:02:40 similar to Charlie Adam having the ball 40 yards out?”
0:02:42 So Sarah took this methodology.
0:02:47 It’s called a Markov chain, just FYI, and she applied it to soccer.
0:02:52 She built a model that let you look at a moment when a player had the ball in a given location
0:02:58 on the field, and the model could evaluate what the player did at that moment.
0:03:03 Specifically, the model let you ask, “Did the choice the player made?”
0:03:08 Passing forward, passing to the side, shooting, whatever.
0:03:15 Did this choice the player made increase or decrease the probability of the team scoring
0:03:17 a goal on that possession?
0:03:25 And, yeah, it really hadn’t been done in soccer before, which was kind of why it was kind of
0:03:26 a revolutionary paper.
0:03:27 I hate to use that term, but —
0:03:28 I’ve seen it listed.
0:03:29 I’ve seen it listed.
0:03:34 538 had a list of the 10 big papers in figuring out this kind of thing, and yours was the
0:03:36 first one on the list.
0:03:40 So it was chronological, to be fair, but still, big first is a big deal.
0:03:44 Yeah, and it was cool, because I was kind of like, “Oh, well, I’m just doing what this
0:03:50 guy in the other football is doing,” but yeah, it was pretty impactful.
0:03:53 And by the way, what did you find out about Charlie Adam?
0:04:01 Yeah, not good to shoot from four yards out was what I found out, which was, yeah, really
0:04:09 soul soothing for me, or it was like, “I was right.”
0:04:12 I’m Jacob Goldstein, and this is What’s Your Problem.
0:04:17 Today we have the third and final episode in our series of interviews with people who
0:04:22 are working at the frontiers of technology to help elite athletes perform better.
0:04:25 My guest today is Sarah Rudd.
0:04:30 She’s the co-founder and CEO at Source Football, a soccer analytics company that works with
0:04:34 professional soccer teams in the US, the UK, and Europe.
0:04:39 There are different ways to frame the core problem that Sarah is trying to solve.
0:04:44 For today’s show, I’m going to go with a problem that generalizes way beyond sports.
0:04:46 The problem is this.
0:04:53 How do you translate analytical insights into meaningful changes in the real world?
0:04:57 In our conversation, Sarah and I talked about analytics and the challenge of getting people
0:05:02 to change their behavior, change the way they make decisions based on analytics.
0:05:06 But before we got to all that, we talked a little bit more about that first paper that
0:05:11 Sarah wrote to prove that Charlie Adam should stop shooting from 40 yards out.
0:05:15 Sarah told me that paper also led to some less obvious insights.
0:05:16 Yeah.
0:05:23 I think one of the first ones was that crossing from wide situations, which was really still
0:05:24 very popular.
0:05:28 So if you think of a player coming down the wings, and then they’re going to kick the
0:05:33 ball high in the air, and then hopefully a teammate is going to head it into the goal,
0:05:36 that is also pretty low value.
0:05:37 Interesting.
0:05:42 Because you have a really high chance of turning the ball over from that.
0:05:46 And was that finding contrary to the conventional wisdom of the time?
0:05:47 Yeah.
0:05:48 At that time, it was.
0:05:53 At that time, there were still a lot of teams using that tactic.
0:06:00 And I think going back, people were starting to just intuitively be like, “Maybe there’s
0:06:01 a better way.”
0:06:08 So it’s not like it kind of broke soccer or football, but it just kind of I think reinforced
0:06:11 some of the intuitions that people had where it’s like, “Maybe there’s a better way to
0:06:12 do this.”
0:06:17 And just to be clear, in the time since then, in the whatever, 15 years or so since then,
0:06:22 have analytics shown clearly that that’s a bad call and has that style of play decreased
0:06:23 as a result?
0:06:24 Yeah.
0:06:29 And now it’s actually seen as a sign of the offense is struggling or we’re resorting
0:06:32 to this really low probability tactic.
0:06:37 So this paper you write gets you hired by this analytics company.
0:06:43 And then the analytics company gets acquired by Arsenal, the famous London soccer team.
0:06:48 And you wind up working at Arsenal for many years.
0:06:54 And one of the things I’ve heard you talk about, about that time, and that clearly is
0:06:59 a big important theme that goes beyond soccer, is trying to figure out how to get the people
0:07:06 who make real world decisions to actually listen to you, to you and the analytics people.
0:07:07 Yeah.
0:07:13 It’s really hard because if you remove sports from it, just getting anybody to change their
0:07:18 opinion based on facts or evidence is really, really difficult.
0:07:24 And now you’re talking about people who are in incredibly stressful jobs where they can
0:07:29 lose their job really based on some random occurrence that happens over the weekend.
0:07:32 Like result doesn’t go their way.
0:07:33 You’re fired.
0:07:34 Good luck to you.
0:07:36 Which is not the way analytics works, right?
0:07:39 Like the way analytics works is you need a large end.
0:07:43 You need to do the thing a hundred times and then 60 times it’ll go the way the analytics
0:07:44 says it.
0:07:45 40 times it won’t.
0:07:48 So that’s just the nature of the probabilistic world we live in.
0:07:49 Yeah.
0:07:54 And analytics is also all about separating process from outcomes and yet these decisions
0:07:56 are still being made on outcomes.
0:08:00 So I have a lot of empathy for people who are in this situation.
0:08:05 And a coach will get fired more likely if they do the thing that the analytics says that’s
0:08:08 contrary to the sort of conventional wisdom in the sport, right?
0:08:13 It’s like why in American football coaches for a long time didn’t go for it enough on
0:08:14 fourth down, right?
0:08:18 Yeah, analytics clearly said they should, but everybody would get pissed at them if they
0:08:19 went for it and didn’t get it.
0:08:21 Now that’s changed, right?
0:08:24 Because of analytics, interestingly, although was it this last Super Bowl where the coach
0:08:28 kept going for it on fourth down and not getting it and people were like, “He did it too much.”
0:08:30 I was like, “No, he didn’t!
0:08:33 Just because it didn’t work out doesn’t mean it was the wrong decision.”
0:08:34 Yeah.
0:08:38 And so, you know, this just kind of goes back to like you need to have kind of top-top
0:08:42 decision makers saying like, “Yeah, it’s okay if you’re going to do something that looks
0:08:44 a little bit unconventional.
0:08:45 We believe in it.
0:08:46 We trust it.”
0:08:48 So, okay, so we’ve talked about why it’s hard.
0:08:52 Did you figure anything out about how to get people to change their decision?
0:08:57 I would say the best thing would be building trust through a common language.
0:09:00 And so for us, the common language was video.
0:09:05 So we could build a model, but like until we could show them the model outputs on video
0:09:10 and walk them through it and say, “This is what the model sees.
0:09:12 This is what the model predicts.
0:09:13 What do you think?
0:09:14 Like walk me through it.”
0:09:19 And so then it became kind of like these collaborative, iterative, model-building processes.
0:09:24 It’s sort of a version of having an idea and then convincing your boss that that idea is
0:09:26 actually their idea.
0:09:28 And once they think it’s their idea, then they’ll do it.
0:09:29 It sounds like that.
0:09:30 It is.
0:09:32 And like maybe a little bit less cynical, but yeah.
0:09:33 Yeah.
0:09:34 Yeah.
0:09:37 So eventually you left Arsenal and you started Source Football, this company that you’re
0:09:39 running now.
0:09:41 What led you to make that leap?
0:09:47 So part of starting this company was that I wanted to learn and to experience as much
0:09:48 as possible.
0:09:55 But the other part is that we felt like there’s a huge need for clubs to get help in terms
0:10:00 of getting started on this analytics journey where a lot of them just don’t even know where
0:10:01 to begin.
0:10:04 They don’t know what’s good, what’s bad.
0:10:10 So our kind of main value-add is helping them get set up, helping them get started.
0:10:16 And then once that happens, doing all of the hard work, hard thinking, that’s impossible
0:10:19 to get done within a football club.
0:10:24 If anyone has ever been in a training ground, I’m sure it’s the same across all sports,
0:10:31 but they’re chaotic, they’re noisy, they’re loud, they’re not good places to do deep,
0:10:32 hard things.
0:10:33 A lot of testosterone.
0:10:39 Yeah, a lot of really loud, bad music, like humping.
0:10:42 If your office is anywhere near the gym, forget it.
0:10:48 You’ve got to wait until everybody goes home before you can actually have a clear thought.
0:10:51 That was one of the things that we realized is you can’t work on solving these really
0:10:53 hard problems.
0:10:57 Football is really hard to analyze in terms of analytics.
0:10:59 You can’t do it within a club.
0:11:01 So that’s why we wanted to be a little bit outside.
0:11:07 And then kind of our long-term vision is actually we’re using this consulting business to kind
0:11:13 of fund the development of our intelligence platform and then the idea is eventually to
0:11:19 look for a set of investors or maybe an ownership group and take control of the club and run
0:11:24 it in kind of like the modern progressive way that we think they should be run.
0:11:29 So what you really want to do is basically buy a soccer team and run it smarter than
0:11:30 everybody else.
0:11:31 Is that what you’re telling me?
0:11:32 Yeah, of course.
0:11:34 Tell me more about the big dream.
0:11:38 We’ll do the pieces, but I’m curious, like that’s a big audacious dream and it’s fun.
0:11:39 Tell me about it.
0:11:42 Yeah, I mean, I think, you know, one of the things that probably everybody in analytics
0:11:47 has experienced is that unless you’re kind of like the key decision maker, it’s always
0:11:53 going to be kind of difficult to influence decisions because there’s always going to
0:11:58 be somebody that has, you know, kind of their own perspective and everything.
0:12:02 And, you know, at Arsenal, we were really lucky because we had like a lot of resources.
0:12:07 But what I see at a lot of clubs is that, you know, they only hire one or two people.
0:12:13 The work that they do is good, but like there’s a limit to what two human beings can do.
0:12:17 And so like a lot of what they produce doesn’t necessarily like match the gut.
0:12:21 And so it’s hard to get this buy-in.
0:12:24 And so then they just say like, well, okay, thanks for that.
0:12:26 I’m only going to listen to you if I agree with it.
0:12:28 If I disagree with it, I’m going to go with my gut.
0:12:31 And so it just becomes difficult.
0:12:36 But I think also like there’s so many issues within football clubs that go beyond just
0:12:39 analytics or, you know, making the right decision on players.
0:12:46 I mean, there’s so much in terms of like building good, good cultures, you know, kind of making
0:12:50 sure that everything is kind of set up in like a professional way.
0:12:53 Because if you think about it, clubs are kind of run by people who’ve never worked outside
0:12:59 of football, but I think there’s a lot of lessons that you can take from working outside
0:13:02 of sports, bring them into into football.
0:13:08 And we’ve seen it in every other major American sport, not to be like super ruthless and say
0:13:12 like, we got to maximize profit, maximize profit, because I think like they are these
0:13:17 weird social institutions that have a lot of meaning to a lot of people.
0:13:19 And so you obviously want to respect that.
0:13:23 And you also have these amazing platforms to bring positive change into the world.
0:13:25 And so you want to take advantage of that as well.
0:13:31 But like there certainly can be a lot more kind of professionalism in them than what we
0:13:33 experience at a lot of places.
0:13:38 So now you have this company and you want to use it eventually to take over the world.
0:13:39 Correct.
0:13:44 Before you use your company to take over the world, like what are the what are the services
0:13:46 you’re selling to teams, to clubs?
0:13:52 Yeah, I mean, so clubs are in really different situations.
0:13:56 And then also like, you know, it’s not like the U.S. where like every single major league
0:14:00 baseball team is like loaded with cash.
0:14:04 Typically in European leagues, like you’re going to have several divisions.
0:14:10 So like you’re talking about anywhere from like a single A baseball team to like a major
0:14:14 league baseball team, like there’s huge difference in resources.
0:14:20 So like orders of magnitude in terms of revenue, how much presumably they want to spend on
0:14:21 your services, etc.
0:14:23 Yeah, exactly.
0:14:27 And so like the main area where we can help them is really recruitment.
0:14:30 So helping them kind of find the best players.
0:14:35 So unlike American sports where you kind of like trade players and draft them, everything
0:14:39 here is kind of like an open market where you can buy and sell the contracts of these players.
0:14:43 And so if you’re a small team and you have a really, really good player, like you can
0:14:49 just sell him, keep that cash, or you can invest it back into your club.
0:14:52 This is this is really echoes of of Moneyball.
0:14:54 I mean, I’m sure I don’t know if you’re tired of hearing about Moneyball.
0:14:58 That book came out 20 years ago, but like that was the basic idea there, right?
0:15:02 It was better, better scouting, essentially, right?
0:15:05 Like you had these scouts who were sort of using their guts and these kind of conventional
0:15:07 wisdom, heuristics.
0:15:12 And then you had what I’m sure now seem like primitive analytics coming in and basically
0:15:16 doing a better job of predicting player success, right?
0:15:17 At some level.
0:15:18 Yeah, exactly.
0:15:22 And so, you know, I hate to say that like soccer is 20 years behind baseball, but like
0:15:24 in a lot of ways we are.
0:15:29 And so like this is still like the main area of value add for a lot of clubs because you’re
0:15:34 competing in this open market with people with varying amounts of knowledge.
0:15:39 So sometimes that knowledge is just, hey, like we had scouts at that game, we’ve seen
0:15:43 this player once, like this is our opinion of him.
0:15:46 And then you have like the more sophisticated teams that are like, okay, well, we have a
0:15:50 database of, you know, 600,000 matches in the world.
0:15:55 We know the 50 best players in every league at every position.
0:15:57 We know how much they’re worth.
0:16:02 So we help teams kind of go from like the former to the latter and just be a little bit more
0:16:05 sophisticated in terms of the amount of information that they have.
0:16:09 I mean, it’s sort of crass to put it this way, but it’s really like pricing assets.
0:16:10 Right.
0:16:11 Like the players.
0:16:13 And again, I realize this is kind of dehumanizing.
0:16:14 I’m sorry.
0:16:22 But it is analogous to people valuing a stock or valuing anything they might buy, right.
0:16:28 And the better you can model the value of the asset, the better you’re going to be at
0:16:30 finding mispriced assets, right.
0:16:33 You want to, you want to find bargains.
0:16:37 You want to, you want to get the most you can for your, for your dollar.
0:16:38 Yeah.
0:16:39 Absolutely.
0:16:44 And so, you know, there’s a lot of techniques that we can take from other industries because
0:16:47 it’s not that different from it where it, where it gets hard is that you have to have
0:16:50 those assets work well together.
0:16:51 They are in fact human beings.
0:16:52 They are.
0:16:53 Yes.
0:16:56 That is, you know, the, the kind of joke where it’s like, well, the problem with football
0:16:59 clubs is that it’s full of human beings.
0:17:01 So let’s talk about what you can model.
0:17:06 Like, what are you good at modeling in, in this context, like in helping, in scouting,
0:17:10 essentially, in helping clubs value, you know, players they might acquire.
0:17:11 Yeah.
0:17:16 So we’re pretty good at modeling everything that happens when a player has the ball.
0:17:21 And unfortunately in football, that’s like a very, very small portion of the game.
0:17:22 Yeah.
0:17:27 And so there was this really cool development in the last four or five years where a number
0:17:33 of companies have come out with a data source that’s basically taking the video feeds from
0:17:38 TV and turning that into tracking data.
0:17:43 So basically they’re able to track the location of every player that’s on screen.
0:17:45 So obviously you, you don’t know all the players.
0:17:51 And then, you know, what we’ve learned is that really what’s going on off screen tends
0:17:56 to not be as relevant and the location of those players isn’t as relevant.
0:18:01 And so now all of a sudden you can start doing modeling on what people are doing off the
0:18:06 ball and use it for recruitment because prior to this, similar to the NBA, there were cameras
0:18:13 in all the stadiums and you would get this like full data set, but only for your league.
0:18:14 And so you couldn’t really use it for recruitment.
0:18:16 And so this has been like a really big change.
0:18:23 So, you know, we can get much better views into what is a player doing physically?
0:18:28 What are they doing defensively in terms of cutting off passing lanes, things like that?
0:18:31 There’s still a lot we’re not good at with that.
0:18:35 But what it’s also done is allowed us to say, well, if we’re going to move a player from
0:18:40 this league to that league, how different is it physically?
0:18:41 How do we think they’re going to adapt?
0:18:44 And so this has really opened up a lot of things.
0:18:49 So if we want to talk about like risk adjusted pricing of assets, now we’re starting to be
0:18:52 able to quantify a little bit.
0:18:56 Like what’s the risk of bringing in somebody from a really, really different environment
0:19:00 into this one versus one that’s more similar or…
0:19:06 So you should apply a larger discount when you’re bringing a player from a very different
0:19:08 league, presumably because you’re taking more risk.
0:19:09 Yes.
0:19:10 Yes.
0:19:15 And if not, then it’s not an undervalued asset and maybe walk away.
0:19:22 So tell me more, like is the output you have just here is what you should pay for each
0:19:23 of these players?
0:19:28 Like you have whatever, 10,000 players or something and you put a dollar value on each
0:19:30 one or like, what’s the output?
0:19:31 Yeah.
0:19:32 I mean, I wish it was that simple.
0:19:37 We’re still like pretty far from putting it all together and saying like, buy this player
0:19:44 at this price, you know, a lot of the difficulties is that there’s no good data set on pricing
0:19:45 information.
0:19:48 So then what do you just have a relative like kind of a stack rank?
0:19:51 I mean, there are like estimates.
0:19:57 So in the media, they’ll say like, oh, this guy went for 10 million, but then a different
0:20:02 media source will say this guy went for 15 million because like the selling club wants
0:20:07 to report a higher price and the buying club wants to report lower price.
0:20:15 So there’s real no truth and then salaries for players are not public for most leagues.
0:20:16 So okay.
0:20:19 So that’s a very important variable that you don’t really have access to.
0:20:20 That’s a problem.
0:20:22 So what is the output of your model then?
0:20:23 Yeah.
0:20:29 I mean, so we kind of do like a stack ranking and like a, you know, projected like, what
0:20:31 do we think this player would do in this situation?
0:20:35 But like the error bars on these things are like are pretty big.
0:20:40 So we’re still kind of in the like subjective realm of like based on these factors.
0:20:42 We think this or we think that.
0:20:48 And then, you know, a lot of the like the markets change every year in football as well.
0:20:53 So even if we had like a really precise model that said this guy should be worth 5 million,
0:20:56 well, he should, he’d be worth 5 million last year.
0:20:57 Yeah.
0:21:02 I appreciate your candor, you know, so the market changes, you don’t have pricing data.
0:21:06 There’s big uncertainty even on the outputs you do have.
0:21:12 Like all of this seems, so what is the use of what you’re doing?
0:21:14 How is it valuable to people?
0:21:19 That’s the point is that we’re competing with people who, you know, have gone to maybe
0:21:26 three or four games, watch this guy in, you know, unknown conditions.
0:21:33 The human brain is like not conditions to be like very objective.
0:21:38 And so like, you know, there’s a host of like very known biases that like people can can
0:21:39 fall for.
0:21:46 And so really what we’re trying to do is just give you like a much more fair and even view
0:21:52 of a player taking into account a lot of, you know, the factors that these scouts are
0:21:58 trying to account for as well, but just doing it more objectively, more rigorously, you
0:22:01 know, over a longer period of time.
0:22:08 It seems from what you’ve been saying, like soccer is behind certainly baseball and perhaps
0:22:09 other sports in analytics.
0:22:13 I mean, your first paper, you were following somebody who had written a paper in American
0:22:16 football and you’re like, oh boy, if we do that for soccer, is it true that soccer is
0:22:17 behind it?
0:22:18 And if so, why?
0:22:19 Yeah, it’s true.
0:22:23 And I think, you know, the debate used to be like, well, are we making any progress?
0:22:29 I think we’ve made a ton of progress, but we’re still really far behind other sports.
0:22:34 And so one answer is that we also don’t get the funding that other sports get.
0:22:37 So so the level of investment just isn’t there.
0:22:44 And so if we were behind 15 years ago, we’re certainly not keeping pace, you know, and then
0:22:47 I think there’s other structural issues with it.
0:22:53 So I love to use this image that somebody else made, but it shows the relative size
0:22:56 of a basketball court to a soccer pitch.
0:23:01 And basically a basketball court can fit into like the little tiny penalty area on a soccer
0:23:02 pitch.
0:23:08 And so, you know, the distances and spacing of players is so much more variable in soccer.
0:23:13 So you can’t say like, oh, 10 meters is a good distance between me and a teammate because
0:23:18 it depends on the situation that’s happening where on the pitch, like, is it a transitional
0:23:19 moment?
0:23:20 Is it kind of a more controlled moment?
0:23:25 And so there’s a lot of complexities like that, you know, the game doesn’t stop.
0:23:29 There aren’t a lot of these set pieces where it’s like we have a very choreographed idea
0:23:32 of what we want to do.
0:23:38 And then, you know, the sad reality is that most leagues only play 38 matches a season.
0:23:43 And so you never see a team in the same context twice.
0:23:46 You either play them at home or away, and then that’s it.
0:23:50 Or maybe you play them in a cup game, which is like a really different environment.
0:23:55 And so all of these things just kind of like add up and it’s like, well, we have fewer resources.
0:23:57 It’s much more complex.
0:24:01 Like of course we’re behind on this.
0:24:09 Still to come on the show, the one big strategic change that Sarah really, really wants someone
0:24:16 in soccer to try and why nobody has tried it yet.
0:24:27 The Medal of Honor is the highest military decoration in the United States, awarded for
0:24:33 gallantry and bravery in combat at the risk of life above and beyond the call of duty.
0:24:42 Since it was established in 1861, there had been 3,517 people awarded with the medal.
0:24:47 I’m Malcolm Gladwell, and our new podcast from Pushkin Industries and iHeartMedia is
0:24:53 about those heroes, what they did, what it meant, and what their stories tell us about
0:24:57 the nature of courage and sacrifice.
0:25:02 Without him and the leadership that he exhibited in bringing those boats in and assembling them
0:25:09 to begin with and bringing them in, I saved a hell of a lot of lives, including my own.
0:25:15 Listen to Medal of Honor Stories of Courage on the iHeart Radio App, Apple Podcasts, or
0:25:26 wherever you listen to podcasts.
0:25:33 So we talked a lot about essentially scouting, evaluating players as one of the big things
0:25:35 you do.
0:25:36 What else do you do?
0:25:37 Yeah.
0:25:43 I mean, so there’s a number of different services that we provide, so it could be doing retrospectives
0:25:49 of how did the team play this weekend, or maybe some kind of, I guess in U.S. terms they
0:25:54 would call it advanced scouting, but delivering kind of like a data profile on their upcoming
0:25:59 opponent, and then going into like on-field stuff, like we’re doing a lot of research
0:26:06 now around, you know, various things in terms of, you know, how can we maximize set pieces.
0:26:10 A set piece is like a play analogous to sort of a play.
0:26:11 Yeah.
0:26:13 So, you know, football is really fluid.
0:26:19 It never really stops, except for these certain moments that are called dead balls.
0:26:23 And when a dead ball happens and kind of like the attacking area.
0:26:26 And you have like a throw-in or a penalty kick or something.
0:26:27 Yeah.
0:26:28 I mean, penalties are pretty straightforward.
0:26:29 They’re kind of exceptional.
0:26:32 It’s like, well, just kick it into the goal.
0:26:39 But corner kicks are kind of like the most common one, because they all happen from the
0:26:40 same location.
0:26:41 They happen fairly frequently.
0:26:43 You can analyze them.
0:26:44 You can prepare for them.
0:26:49 So you have this set piece opportunity, and then it’s, you know, what’s the strategy we’re
0:26:52 going to use to maximize it?
0:26:54 So a lot of research into that.
0:26:58 Has analytics in a general way changed the way people take corners?
0:26:59 Yeah.
0:27:06 I mean, I think it has influenced how much time they spend thinking about it and preparing
0:27:07 for it.
0:27:12 So that’s another one, Ted Knudsen, who is the CEO of Statsbom.
0:27:18 He’s been shouting this from up high for the longest, but set pieces for a long time were
0:27:23 really valuable, but teams would only train them for like 10 minutes a week.
0:27:25 It’s an interesting level of analytics.
0:27:26 That’s an analytical claim, right?
0:27:31 It’s like your training time is measurable, and you should be allocating the right proportion
0:27:32 of it to the right things.
0:27:37 And he’s essentially arguing that you’re underfunding in time corners.
0:27:38 Yeah.
0:27:39 Yeah.
0:27:43 Like the number of goals scored from corners or conceded from corners was not at all proportional
0:27:48 to the amount of time spent training them.
0:27:54 And so that’s been a big focus on teams lately, and I still don’t think they spend enough
0:28:00 time training it, but now you’ll see certain clubs even hiring set piece specialists.
0:28:05 So they’ll have a member of coaching staff whose job, whose only job is really thinking
0:28:08 about these things and helping prepare the players for it.
0:28:11 Is that person basically the corner kick coach?
0:28:15 I mean, they, you know, there is a guy who’s like a throw-in coach as well.
0:28:19 So you can even have like some specialties to them, but…
0:28:21 That guy dreams of being the corner kick coach.
0:28:22 Yeah.
0:28:24 Gets promotion too.
0:28:30 So I mean, just thinking more generally, like it seemed, well, other sports here, my U.S.
0:28:36 centric sports knowledge, which itself is somewhat limited, like clearly analytics have
0:28:42 changed, you know, football, you see people going for it on fourth down, more basketball.
0:28:46 Obviously the demise of the mid-range shot, right?
0:28:50 Somebody just realized analytically some years ago, like, don’t shoot from whatever 12 feet
0:28:51 out, right?
0:28:52 Shoot from under the basket or take a three.
0:28:54 And then baseball, they was the shift.
0:28:59 Like there are these big, big changes in the way the games look because of analytics.
0:29:01 Is there anything analogous in soccer?
0:29:06 Yeah, probably like the earliest one was that the distance from which people are shooting
0:29:08 has changed.
0:29:13 So you’re not going to see a Charlie Adam taking a shot from 40 yards too often unless,
0:29:17 you know, there’s like a, oh, the goalkeeper is off his line, he’s out of position, I can
0:29:18 get it in there.
0:29:20 Is that because of you?
0:29:24 You know, I can’t take credit for it.
0:29:27 It feels like, you know, the invention of calculus where there were a lot of people kind of coming
0:29:28 to this conclusion.
0:29:33 So you’re saying you might not be Newton, but if you’re not Newton, you’re Leibniz?
0:29:34 Yes.
0:29:35 Okay, fair enough.
0:29:39 Yeah, that’s probably like the earliest one.
0:29:46 And then again, like, you can’t prove causality, but there’s been a shift to this tactic of
0:29:48 teams putting a lot of pressure up high.
0:29:54 So if they have the ball in the attacking area of the pitch and they lose it, they immediately
0:29:58 put pressure on the other team to try to win it back as quickly as possible.
0:30:00 Kind of a full court press.
0:30:01 Yeah, exactly.
0:30:05 And so I don’t know if the origin of that is based in analytics, but like analytics will
0:30:11 tell you, yeah, you’re more likely to score a goal if you win the ball back up high.
0:30:12 So that’s another one.
0:30:13 Huh.
0:30:14 Why do you think people didn’t do it before?
0:30:17 Is it not intuitively obvious that that’s good?
0:30:19 Yeah, because if you don’t want to…
0:30:20 Are you so tired?
0:30:21 Well, yeah.
0:30:25 So one, it makes you tired and there is like a risk of injury trying to kind of like make
0:30:28 the guys super fit to do this.
0:30:33 But two, kind of going back to like the fourth down analogy, if you don’t win the ball back,
0:30:38 then you’ve committed so many players upfield that you’re going to concede a kind of like
0:30:40 silly looking opportunity.
0:30:41 It’s risky.
0:30:42 It’s risky.
0:30:43 Yeah, it’s a calculated risk, but it’s risky.
0:30:46 The expected value is positive, but the variance is high.
0:30:47 Exactly.
0:30:50 What are some interesting frontier problems?
0:30:53 I mean, they’re sort of getting people to listen to you, which we’ve talked about, but
0:31:00 just on the analytics side, like what are you trying to figure out on the analytics side?
0:31:03 What’s a big problem you’re trying to solve?
0:31:08 One area that we haven’t really exploited yet, but I’m curious about, but we always have
0:31:13 this problem that like we can’t always learn from the data because everybody’s doing the
0:31:14 same thing.
0:31:17 You can’t really do an A/B test because everybody does A.
0:31:19 Yeah, exactly.
0:31:20 And so it’s really frustrating.
0:31:24 I mean, like the biggest example would be substitution patterns.
0:31:28 Everybody makes roughly the same type of substitution at the same time.
0:31:31 And presumably it’s not optimal, right?
0:31:34 It’s just conventional wisdom, but you’re like, “There could be a whole better world
0:31:35 that nobody’s ever tried.”
0:31:36 Yeah, exactly.
0:31:41 And so I think, you know, with generative AI, like now you’re kind of talking about like,
0:31:45 “Well, can we do really smart, really realistic simulations on these sorts of things?”
0:31:47 Kind of synthetic data.
0:31:48 Yeah.
0:31:49 Yeah.
0:31:50 That’s what we’re hoping.
0:31:54 So are you trying to figure out a better way to do substitutions?
0:31:55 Yes.
0:31:56 Yeah.
0:31:57 You feel like you got it?
0:32:00 Do you feel like you have one in your pocket or you’re not quite ready to say?
0:32:01 I mean, we have some theories.
0:32:05 We haven’t proven them, but like we also need somebody to say like, “Yeah, go ahead.
0:32:07 Like mess with our substitution patterns.
0:32:08 We feel comfortable with this.”
0:32:09 Uh-huh.
0:32:10 Uh-huh.
0:32:14 It is amazing how, I mean, like if you think of the shift in baseball or whatever, like
0:32:20 a game can just exist for a hundred years and there can be a way, better way to do it.
0:32:25 And nobody ever does it just because they don’t have the imagination or the courage.
0:32:26 Yeah.
0:32:27 It’s wild.
0:32:31 And I think, you know, that sort of thing, like I hope in soccer is less common because
0:32:35 it’s more of like an adversarial, uh, game, but-
0:32:37 But it maps to the world more generally, right?
0:32:42 Like, I think a lot of it is just fear, like you won’t get in trouble if you do the thing
0:32:45 everybody else did and you have a bad outcome.
0:32:48 Like you see it in, you know, in finance certainly, right?
0:32:54 Like financial advisors just exhibit herd behavior because of the way the incentives
0:32:55 are structured.
0:32:57 Like if you do what everybody does and you lose money, you’re like, “I was just doing
0:32:59 what everybody else did.”
0:33:00 Yeah.
0:33:01 Yeah.
0:33:02 And it’s really frustrating.
0:33:07 And so, um, you know, I think that again goes back to like why we want to get control
0:33:15 of a club because, you know, I think there is a lot of ways that we can optimize things.
0:33:20 And if we’re only reporting to ourselves, then it’s like, well, you know, I’m not going
0:33:23 to fire myself just because the theory didn’t work out.
0:33:24 Yes.
0:33:26 It’s hard to get to a big enough end though, right?
0:33:30 It’s hard to get to a big enough sample size or just aren’t that many games.
0:33:35 Um, so, so let’s talk about the sort of happy outcome for you.
0:33:41 Say it’s whatever, five years from now, and you have found capital, you’ve found somebody
0:33:44 who wants to essentially bet on you, which is, if I understand correctly, what would
0:33:46 have to happen.
0:33:53 And you are, you know, with your finance-y, finance-y or partners running a club.
0:33:54 What’s that look like?
0:33:55 Yeah.
0:33:59 I mean, I think it would look really, really different.
0:34:03 Um, I think that the typical profile of who would be working in that club would be quite
0:34:04 different.
0:34:05 Yeah.
0:34:06 Oh, yeah.
0:34:07 Nerdier.
0:34:08 For sure.
0:34:09 For sure.
0:34:14 It’s hard because like, uh, you know, I’ve played a lot of sports growing up and so it’s
0:34:16 like, well, you know, I’m not that nerdy.
0:34:19 And then it’s like, oh, yeah, yeah.
0:34:25 Um, but yeah, I mean, I think it would look nerdier, but I think it would be a lot more
0:34:29 like just a totally different mindset, really a lot more people who are like, yeah, let’s
0:34:31 do things differently.
0:34:35 Let’s have the courage to, to try new things.
0:34:39 And then let’s have kind of the, the thinking power behind it to not just like, let’s try
0:34:45 random things, but let’s try really well thought out, creative, but courageous things.
0:34:50 When you talk about like, you know, being bold and creative and different, is there
0:34:53 some particular thing you have in mind?
0:34:57 Is there’s like one thing where you’d like that one thing, I wish somebody would just
0:34:58 try it.
0:35:02 I mean, right now it’s probably the substitution thing because that’s like the lowest hanging
0:35:06 fruit and it like really, really drives me nuts.
0:35:08 What’s your secret theory?
0:35:09 What do you think would be better?
0:35:10 Oh, I think you should do early.
0:35:14 I think you should kind of treat it like line changes in hockey.
0:35:19 So I mean, like right now what they do is they, they typically try to bring on maybe like a
0:35:24 fast player late in the game that, you know, would theoretically be running against tired
0:35:25 legs.
0:35:28 But the problem is that like it’s too late.
0:35:33 And so, you know, our idea is basically like bring on people who like know that they only
0:35:39 have 30 minutes, run as, as hard as you can, as crazy as you can, get, get that one goal
0:35:42 advantage or two goal advantage and then adapt.
0:35:44 And nobody has tried that.
0:35:45 That doesn’t seem crazy.
0:35:48 Like there’s a hundred soccer teams.
0:35:51 All you need is one person to try and it’s like a free idea.
0:35:52 I know.
0:35:57 Well, you know, if we see like a change in substitution patterns after this, that’s right.
0:35:58 Yeah.
0:36:02 But, you know, like the crazy thing is there used to only be three substitutions in soccer.
0:36:06 And so you always wanted to keep at least one in case a player gets injured.
0:36:08 So that really leaves you with two.
0:36:11 But during COVID, they upped it to five.
0:36:14 And so now it’s like, well, you have two, three, and that has persisted.
0:36:15 Yeah.
0:36:19 So they’ve almost doubled the number of substitutions allowed.
0:36:21 Has the strategy and using them changed?
0:36:22 No.
0:36:25 It’s like a very slightly change.
0:36:30 And a lot of coaches don’t even take advantage of all substitutes.
0:36:36 Why do you think nobody has, you know, dramatically significantly changed their strategy around
0:36:38 substitutions, even though the rule has changed so significantly?
0:36:39 Yeah.
0:36:41 One, I mean, I guess there’s like two explanations.
0:36:45 The easiest one is like, well, it’s always been the way that we’ve done it.
0:36:48 So, you know, why change?
0:36:53 I think part of what’s driving that, though, is that the way that rosters are constructed
0:36:55 hasn’t changed.
0:37:00 And so, you know, depending on like the finances of a team, you might not have as many like
0:37:05 quality players to go that deep onto the bench.
0:37:09 But you know, if this is your strategy, then you can change how you recruit.
0:37:12 You can change the profile of that squad composition.
0:37:16 The second answer is interesting, right, because it requires you to think more systematically.
0:37:17 It’s like, oh, the rules are different.
0:37:21 So therefore, we should build a different team, which is kind of next level.
0:37:22 Yeah.
0:37:29 And yeah, that’s not typically how teams operate.
0:37:33 We’ll be back in a minute with the lightning round.
0:37:47 The Medal of Honor is the highest military decoration in the United States, awarded for
0:37:53 gallantry and bravery in combat at the risk of life above and beyond the call of duty.
0:38:01 Since it was established in 1861, there had been 3,517 people awarded with the medal.
0:38:06 I’m Malcolm Gladwell, and our new podcast from Pushkin Industries and iHeartMedia is
0:38:12 about those heroes, what they did, what it meant, and what their stories tell us about
0:38:16 the nature of courage and sacrifice.
0:38:22 Without him and the leadership that he exhibited and bringing those boats in and assembling
0:38:27 them to begin with and bringing them in, I saved a hell of a lot of lives, including
0:38:28 my own.
0:38:34 Listen to Medal of Honor Stories of Courage on the iHeart Radio app, Apple Podcasts, or
0:38:45 wherever you listen to podcasts.
0:38:50 We’re going to finish with the lightning round, soccer or football?
0:38:56 I say football, fries or chips, cookies or biscuits.
0:39:00 So my husband is from India, and he will kill me if I say cookies, but cookies.
0:39:02 Are there other Britishisms?
0:39:03 You sign your emails.
0:39:04 Cheers.
0:39:07 Sometimes I hate doing it.
0:39:12 Mate is another one because that’s you call everything brilliant.
0:39:15 I love talking to British people because they say everything I say is brilliant, but they
0:39:16 don’t mean it the way I think.
0:39:21 Yeah, brilliant, bloody, yeah, just little things like that.
0:39:22 Blimey?
0:39:23 British guy once said blimey to me.
0:39:24 I loved that.
0:39:27 Yeah, I’ve not done that.
0:39:31 But yeah, I think when I was little, my brother and I used to antagonize my dad until he
0:39:37 would yell at us in British slang, and that was success for us.
0:39:39 Give me your impression of your dad yelling at you in British slang.
0:39:42 He would just say, “Oh, bugger off.”
0:39:46 A bugger off is good.
0:39:50 What’s one thing you learned working at Microsoft?
0:39:53 It’s hard to get things done in large companies.
0:39:58 Do you have any tips for getting things done in large companies, or is it just leave?
0:40:02 I mean, my answer was leave.
0:40:07 But yeah, I mean, I think there’s certain behaviors that maybe I don’t necessarily possess
0:40:13 or love, but aggressive, loud people tend to get things done there.
0:40:16 Were you there in the Ballmer area?
0:40:17 Oh, yeah.
0:40:18 Ballmer era?
0:40:19 That’s very Steve Ballmer vibes.
0:40:26 Who’s the most underrated player you’ve ever seen?
0:40:29 It might be this guy named Manu Trigueros.
0:40:36 He plays for a club in Spain called Virial, and I think he’s brilliant.
0:40:41 He was kind of born at a time when Spain had loads of really, really talented players in
0:40:43 his position.
0:40:46 He never got called up to the national team or anything like that.
0:40:48 But his running style is a little bit nerdy.
0:40:51 He runs very much on his toes.
0:40:55 He doesn’t look like an athlete.
0:41:00 He actually is like, while he was playing professionally, he was doing his masters in education.
0:41:02 So he was like a student teacher.
0:41:03 It’s just that you love him.
0:41:06 It’s not that he’s an amazing player, it’s that you love him?
0:41:10 I love him, but he was also an amazing player.
0:41:16 But I think having this air of nerdiness around him kind of kept him a little bit under the
0:41:17 radar.
0:41:26 Who’s the most overrated player you’ve ever seen?
0:41:28 That’s a really hard one.
0:41:30 Are you afraid of getting in trouble?
0:41:33 Is there a name in your mind and you just don’t want to say it because you don’t want
0:41:34 to antagonize anybody?
0:41:35 Yeah.
0:41:36 Definitely.
0:41:37 There’s some of that.
0:41:40 And then there’s like a little bit of like hindsight bias where it’s like, there were
0:41:46 guys who were overrated at the moment and then like kind of crumbled and failed.
0:41:53 So it’s like, well, if I say like Marwan Shamak, who was a kind of like notoriously disastrous
0:41:56 signing for Arsenal, everyone would say like, oh yeah.
0:41:57 Too easy.
0:41:58 Yeah.
0:42:06 What’s one piece of advice you have for women working in male dominated fields?
0:42:10 That’s a really good one.
0:42:15 It would probably be that just you belong and know that you belong and don’t let people
0:42:19 try to make you feel that you don’t.
0:42:23 And one of the ways that you can do that is talk to other women.
0:42:27 When I got started 15 years ago, like there weren’t any other women.
0:42:32 Now there’s loads of women working in all different sports and all different sorts of
0:42:34 roles.
0:42:39 And a couple of years ago, this conference started called Women in Sports Data.
0:42:43 This year it’s going to be held in Philadelphia, September 7th, I think.
0:42:48 But that’s like a great place to connect with people because to me, I find it really powerful
0:42:54 to be in like a gymnasium with like a room literally full of women who are interested
0:42:57 in sports and data and technology.
0:43:02 And yeah, just, I don’t know, it’s good for your mental health.
0:43:05 It’s good for, you know, your self-esteem and everything.
0:43:11 But yeah, just know that you belong, that you can do it and that, yeah, nobody should
0:43:13 tell you otherwise.
0:43:16 So Brad Pitt played Billy Bean and Moneyball.
0:43:18 Who’s going to play you in Moneyball 2?
0:43:21 Don’t call it soccer.
0:43:23 Oh, yeah.
0:43:25 I have no idea.
0:43:30 I don’t watch a lot of movies, so like I couldn’t even name an actress right now.
0:43:34 Maybe Kira Knightley because she was in Bend at Lake Beckham.
0:43:38 So I’ll go with Kira Knightley.
0:43:43 Sarah Rudd is the co-founder and CEO of Source Football.
0:43:46 Today’s show was produced by Gabriel Hunter Chang.
0:43:51 It was edited by Lydia Jean Cotte and engineered by Sarah Brugger.
0:43:54 You can email us at problem@pushkin.fm.
0:43:59 I’m Jacob Goldstein, and we’ll be back next week with another episode of What’s Your Problem?
0:44:17 The Medal of Honor is the highest military decoration in the United States.
0:44:25 Since it was established in 1861, there have been 3,517 people awarded with the medal.
0:44:29 I’m Malcolm Gladwell, and our new podcast from Pushkin Industries and iHeartMedia is
0:44:35 about those heroes, what they did, what it meant, and what their stories tell us about
0:44:38 the nature of courage and sacrifice.
0:44:44 Listen to Medal of Honor stories of courage on the iHeart Radio app, Apple Podcasts, or
0:44:46 wherever you listen to podcasts.
Sarah Rudd is the co-founder and CEO of the soccer analytics company src | ftbl (It’s pronounced “Source Football.”) Sarah’s problem is this: How do you model a sport as fluid and complex as soccer and translate the analytical insights from the model into meaningful changes on the pitch?
This is the third and final episode of our series about people who are working at the frontiers of technology to help elite athletes perform better.
See omnystudio.com/listener for privacy information.