Betting 360 Ep014: Soccer Data Analysis with Dr. Howard Hamilton

Betting 360 Podcast - Betting From All AnglesEpisode 014 features Dr. Howard Hamilton from Soccermetrics Research chatting with Dave about the most valuable data analysis metrics available in the round ball game. Dr. Hamilton is a former aerospace engineer who uses his mathematical ability to view sports data in a way that few others can. In this episode they chat about the past, present, and future of Soccer analysis.

Punting Insights You’ll Find

  • How to break soccer down into an actionable set of data.
  • Which stats to look at to get a supreme advantage.
  • The most valuable high-level measurements of a defensive player.
  • Dr. Hamilton’s biggest recommended change to the soccer analysis industry.
  • How to identify promising players that most people overlook.
  • A crystal ball view into the future of soccer betting.

Today’s Guest:

Dr. Hamilton’s Closing Tip:

Any innovation that happens in football analytics will happen outside of the football clubs. “

Episode 014: Dave Chats About Soccer Data Analysis With Dr. Howard Hamilton

Welcome to Betting 360, your number one source for horse racing and sports betting insights. Coming around the bend is your host David Duffield, with another expert view to give you the winning edge.

David: Hi this is David Duffield and welcome to another edition of the Betting 360 Podcast. Our special guest today is Dr Howard Hamilton. He’s the founder and the CEO of Soccermetrics Research, that’s And we’ve looked at, or we know about the growth of analytics across all the different sports, and also just the increased importance of analysis of data in horse racing as well. So I wanted to get his take on the boom of analytics in soccer, because it’s such a fluid game, it’s such a complex game, and also such a low scoring one. So it’s important to find out how data can be used and to get some insights and from our perspective if there’s any betting pointers. To see undervalued players, undervalued teams, and really narrow down on that, so let’s have a chat with Dr Howard Hamilton.

Thanks very much for joining us today Howard.

Howard: Hi Dave. Thanks for having me on and hello to everyone else in Australia.

David: Yeah looking forward to having a chat. Certainly a different perspective to a lot of what we cover. So just to get started, what motivated you to move from being an aerospace engineer through to football analytics?

Howard: Well, football’s always been an interest of mine. I’ve had a couple of sites on soccer in North America. And I had always come to football with an analytical view, come to sport in general with an analytical view. And I took a look at what passed for statistical analysis in football, and I thought it was very primitive at the time. And I had my own ideas given my background as an aerospace engineer, and software programming, mathematical modelling, things like that. And I think the issue with doing analysis in sports and other areas, is that if you think of the problem and how to set it up, you can express the problem in the language of mathematics. And I was more comfortable, I was a lot more comfortable in that world than a lot of other people in the football analytics space. I thought had something to contribute, and that’s how Soccermetrics got started.

David: You guys over there have the same battle as us in Australia. That ‘football’ to us is often Australian Rules or Rugby League, football to many in America is NFL, but it’s that interchangeable term soccer versus football.

Howard: Yeah. It’s an interesting question, when I started Soccermetrics, Soccermetrics is a contraction of Soccersabermetrics. I could use the word Sabermetrics because it’s specific to baseball. But there is some relationship, there is some relationship to that. I think Soccermetrics, the word Soccermetrics, just reflects the fact that a lot of these statistical analysis that happens in sport, originates from North America. I think there’s much more of a culture of statistical analysis in North America than anywhere else. As far as using the word soccer, my perspective on it is it’s a British word, it’s a legitimate word for soccer, I feel there’s nothing wrong. And if people have a problem with it then that’s their issue. But I’ve never had an issue using the word Soccermetrics in Europe, or anywhere else. It has to be said, whenever I’m in Europe I call it football, and when I’m in North America I call it soccer, and that seems to be a reasonable compromise for everyone.

David: Makes sense. So you mentioned the prevalence of analytics in the big American sports. In a sport like NBA obviously there’s only 5 players on the court at the one time. In Major League Baseball it’s close to a one on one battle with the pitcher and the batter. Obviously there’s the fielding aspect as well, but that’s far less important. Then you’ve got soccer where it’s such a complex game and scoring is so rare, how do you break that down into meaningful data?

Howard: I suppose for, I guess, so if you were to understand about Soccermetrics, we’re not in the data collection business per se, we’re more of the data analysis business. Data collection in soccer is a very labour intensive, very time intensive business. And there are companies such as Opta Sports and Prozone and some others who did a really good job of collecting a lot of data, a lot of highly granular data within a football match. The issue with football analysis is, that until about 10 or 12 years ago, the most data we had surrounding a football match was who started, who scored, who was substituted, maybe who was sent off and that’s about it. So now through analysis, either watching on live feed, or cameras positioned above a stadium, or camera position above a pitch, observing every motion of the player and the ball we have information that describes not only every touch-by-touch event that happens on the pitch, but also all the player, ball, and referee motions during play. So soccer’s moved from a period of time where we had very little data to work with, to now clubs, leagues, and the entire industry is literally swimming in data.

David: What stats do you consider to be important when you’re assessing either player performance or team performance? Because you mentioned that in years gone past, there was a very limited range of stats. But if you were to look at key indicators, key performance indicators right now, are there certain stats that you tend to focus on?

Howard: I tend to look at the game a lot more holistically. As far as how a player contributed to the likelihood of a shot taking place, or how a player contributed to a shot not taking place, if you have a defensive wall. I think that what’s, it’s described as what’s currently done now in this industry. You look at number of passes or number of entries, whether into the opposing half or the final third or penalty area. Then you look at passes within the area, to assess some sort of conversion rate. You look at different areas of the pitch and where players shot the ball, or where players either made blocks or interceptions or things like that. I think all of those stats in soccer can be ultimately misleading without an understanding of context. Because I think that the underlying assumption within say, tabulating the number of passes, or computing a number of shots in certain areas, is that you are saying that every pass is the same. You’re saying that every shot is the same, and I think we all know that that’s not the case. Because at a different oint in time an action is dependent on who initiated the action, when it occurred, where on the pitch it occurred.

David: You mentioned defensive stats, again how do you try and measure the defensive
performance of a player?

Howard: I think, thinking about it fundamentally, I think the object of a defensive player, or the object of any player that is in a defensive role, that’s not just defenders, but also midfielders, even strikers if they’re forced to drop back is to prevent the opposing team from creating opportunities to score. So I guess at a much higher level you could look at either an interception, or tackle attempt at different areas of the pitch. But with the amount of spatial data that we have, and also temporal data’s that we know, so that not only do we know when an event occurred, but also where it occurred, and who initiated it, we can work on bringing those events together in order to figure out who contributed to breaking up those offensive plays, and what impact they had on the match. But I would say that right now, the current state of the art involves tabulating tackles, and interceptions, and clearances, and shot blocks as well, and a lot of times that’s just all you have to work with, but it’s a little disappointing that we’re at a stage now, where we have more temporal and spatial data that’s available, and we still persist in doing that type of analysis. I just feel there’s a lot deeper work that we can do.

David: In what way? What would you suggest as being more meaningful data, or analysis anyway?

Howard: I’m really encouraged by some of the work that’s being done in network in applying network analysis to football. As far as viewing each play as a bunch of interconnected plays between players and actions, identifying which players are more influential, which player had a role in either creating opportunities or destroying opportunities. Which could be good or bad if you have an offense role or a defence role. Network analysis, not to get too technical, just, you’re looking at interconnectedness between events, between people, between actions, and even applied in social network analysis and biological analysis. And some other technical fields that are similar to soccer in that divert interconnected, they’re very dynamic, they resist a lot of conventional analysis. And I think that the transition in football analysis, to incorporating some elements of network analysis is a really positive thing, and I hope that continues. We have some ideas on that, and we hope to have something to say about that in the coming months.

David: How do you identify promising players? I know part of the services you offer at Soccermetrics are packages on certain leagues, and identifying players that show a lot of promise, or undervalued, or both. So how do you go about doing that?

Howard: The way we do it right now, is that we look at the statistical record of players over the entire season. And we’ve, we being myself and my partner Erin Neilson, have done some work on, have done some work on assessing expected performance as a function of current statistical output. Their age, position, and expected market value. So we take all those quantities and characteristics together, and work out an internal rating for those players, and we use that in the ranking of our prospect list.

Now one caveat with that is that the prospect list is not necessarily best player list. For example in our prospect list in Spain, I think most people, well I think just about everyone would say that Lionel Messi and Cristiano Ronaldo are the best players in the Spanish Primera. But it’s also clear that those two players aren’t going anywhere. Maybe Ronaldo if you believe the papers, but those two players aren’t going
anywhere, so presenting them in a prospect report isn’t very useful.

But I think what is useful, is presenting those players who are up and coming players, who have some, based on their statistical performance, and knowing how that might relate to performance and other leagues, that they could do very well at a larger club or in a higher profile league. So essentially what it is, we focus on younger players, most of the players, I think 95% of players on our prospect list are under 25. And we take a look at their yearly statistical output, and going on trends, and previous performance of other players, we determine whether this player has potential, or repeating that kind of performance in the future.

David: Okay. And what about by position. Do you find that there are certain positions that might be overrated, in terms of the salaries they attract or the transfer fees?

Howard: I haven’t done a lot of work in that kind of analysis to say something really definitive on that. But I do believe that for the most part, defensive players are very underrated. If you look at the championship winning teams in the domestic leagues in Europe, this is not just Europe, it’s other parts of the world. It’s not necessarily the teams that score the most goals that win league titles, it’s teams that have really consistent defences. If you want to look at it in a technical way, if they allow a low amount of goals and they do that consistently in every game, if their variance in goals allowed is very low, that’s a really good indicator that that team has a strong chance of winning the league.

So I think this is something that Chris Anderson has hit on in his book Numbers Game, I think he has a chapter devoted to it. I think if you want to have a very strong team, you should try to minimize the goals that you allow, and minimize the goals that you allow in every match and do that consistently, and to that end it goes, to that end it leads to bringing on a very solid goalkeeper, and also a very solid defensive line in front.

David: So if you were the manager of a club that would be your first focus, defense?

Howard: Yes.

David: And what about in terms of managers tactics. Have you looked at all at, you find some content with a draw late in the game, as opposed to risking that to go for a win. Do you look at that at all as far as the ideal strategy in that situation, or is that not something that you cover?

Howard: It’s not something I cover at the moment, it’s something of interest. It is a topic of interest to me, but it’s not something I can really focus on lately. I’ve done some work on squad rotation, and player utilisation and some of those results were quite interesting, particularly in the shorter term say European Championship or African Nations Cup or things like that. Even through the course of a season, it’s interesting to see the practices, the varying practices that managers in top flight leagues have when it comes to rotating players. But as far as in match strategies, we haven’t done a lot of work on that at this time, but that’ll change in the future.

David: I just know that in the metrics movement in the US, in a sport like NFL they’ve said that the coaches there can be quite conservative. The risk of public ridicule, they won’t go for it on fourth down certain situations. And I was wondering if soccer was much the same late in the game, just being content with a draw, but that’s okay it doesn’t sound like it’s been a focus.

What about measuring the strength of one league against another? With the analysis that you do, I suppose being self-contained on a particular division or league, can you actually measure across different competitions and assess the performance of one team against another, or one league against another in terms of their strength?

Howard: I think there have been some various attempts of doing that. I think a really crude attempt is through the Continental Coefficient set. UEFA has one, I created one for CONCACAF, well they use something very similar to what I developed. You know Federations like those will tell you until their blue in the face that it’s not a ranking system. They strictly use it to seed teams and determine slots or competitions. But people see it as a comparative ranking of different leagues anyway, I think that’s, I think of right now that’s really the extent of it. I think there are some people who look at, again cumulative statistics.

Of either passing performance or shot performance, or shot conversion, performance between leagues and try to work out some sort of comparison between competitions. Here in North America that’s really common with Major League Soccer seeking to compare itself against other leagues, or other people trying to compare Major League Soccer against other leagues for different purposes.

Where to pump up the image of MLS, or knock down the image of MLS. I think the way that’s done right now isn’t particularly useful, because again you’re just looking at cumulative, you’re looking at a bunch of cumulative events or lumping together, and I think the context is lost. I think you can look at characteristics of leagues, and I think beyond the big five European leagues, there are some very, I think there are some significant differences in points, styles, and characteristics. But I would say right now we haven’t really scratched the surface of doing any, I think true comparative analysis of different domestic leagues and competitions.

David: Okay, just to finish up then. Obviously there’s been a massive growth in your industry, I mean just analytics or metrics in general rather than specific to soccer. Over the last, well 10 to 15 years it’s really exploded. Where do you see it going in the soccer industry, say over the next 3 to 5 years? How do you think it will be integrated into assessing player performance, team performance, and how do you think managers will use it?

Howard: Well I think there’s going to be a lot of consolidation in the soccer analytics industry. There’s just a lot of small fragmented analytic firms, and you also have the large data providers. So I think there will be some tough consolidation, either smaller consulting forms dying out, or merging, or being bought out by bigger companies. As far as analysis, I think the football analytics community is at a crossroads right now because analysis needs data, but to do that kind of data you need to work with either clubs or data companies. And that’s always been proprietary, and clubs by their very nature have a very short term point of view. So I think that any innovation that happens in football analytics, will happen outside of the football clubs.

But I believe that, I think if you wanted to go through some areas where you might see some innovation, perhaps incorporation of psychology with on-field performance, some more work in front office analytics , from player evaluation, to player recruitment, talent identification, talent tracking. I think if FIFA allows the use of GPS units, like the RFID tags that you see in AFL or Rugby, that that would open things up completely. To be able to track player movements within the game, and I think that would make a lot more data available to the football analytics community. But again the issue gets back to you, how do you do analysis on it, and how do we make better use of that data than just coming up with tabulations and averages? So I think we have a lot of challenges, but I also think there are a lot of opportunities as well for some innovation to be made, if we take full advantage of them.

David: Excellent. Well I appreciate your time today, the promise of the podcast is a look at everything from different angles and it’s certainly been that today. And the first doctor we’ve had on the podcast, so it’s been an interesting insight and really appreciate your time.

Howard: Thanks Dave, it was a pleasure being on, really appreciate it.

David: Thank you

Get More Betting 360

Make sure you don’t miss out punting tips to come! Subscribe on:

or you can directly download this episode by right-clicking, Save As Here.

What Do You Think of the Show?