A look at HSV’s data driven scouting during the Arnesen era

When identifying an effective technical scout for a Bundesliga club with all the thoroughness and detailed statistical analysis skills needed for effective technical scouting, in 2011 you’d have struggled to look passed a man named Steven Houston. In 2011, Houston joined Hamburger SV.

At the time, Houston was an intelligent Brit (and still is, to be fair), educated in America in the field of data analysis with the aptly-named basketball team the Houston Rockets, and had also worked as Head of Technical Scouting and Data Analysis at Chelsea FC. Houston was recruited away from the Rockets European basketball scouting scene by the Blues’ Director of Football Operations, Mike Forde. If you needed more evidence for Houston’s credentials in data analysis, his MIT Sloan Sport Conference profile adds further weight, reading: ‘[C]oming from a consultancy background, Steven previously worked for five years as Lead Analyst with some of the top general insurance companies in the UK’. In short, Houston works very effectively with data.

As HSV’s contemporary historians will tell you, he was not the only Chelsea FC employee to make his way over to the Imtech Arena in 2011. In the same summer, Chelsea’s Sporting Director Frank Arnesen and Chief Scout Lee Congerton relocated to the Bundesliga club, and the two greeted Steven Houston and his huge collection of data on arrival.

Frank Arnesen, who had appointed Congerton as Hamburg’s Technical Director and Houston as the team’s Chief Technical Scout, was looking to improve the player recruitment, analysis and scouting department of the German club in a long-term project by using data (among other things) to help inform the club’s training sessions, match performances, and manage their squad more effectively. Arnesen had set up what he thought was the dream team of football analysis. Lee Congerton worked alongside Steven Houston, who was specifically asked to help with player recruitment, and use his arsenal of data, both visual and statistical, to find players who could help Hamburg improve. Congerton also set up effective data analysis for opposition scouting, and assessing and improving Hamburg’s current roster of players, including monitoring players who were out on loan.

The word ‘Moneyball’ is largely unavoidable when talking about sports clubs adopting a data-driven approach, especially in regards to player recruitment, and in 2011 Hamburg faced similar comparisons. The only other Bundesliga club at that time who significantly invested in data analysis in terms of player recruitment was Borussia Dortmund, who used various statistical models to buy players, but mainly to analyse their own team and individual performances. Like Dortmund, Hamburg would now rely on individuals like Congerton and Houston to improve the side, widen the knowledge gap, hopefully giving them a competitive advantage in the transfer window too. In Germany, Houston in particular was also especially aware of the resonance that any potential signing could have with Hamburg’s fan base, and in 2012 he stated, “50% of every team is owned by some sort of fan ownership group so you tend to find them more rigorous in the way they’re evaluating their outgoings and incomings.” It was under these strict parameters that Arnesen’s Hamburg would operate.

2011, in truth, was a bad year for data analysis in football. Liverpool FC, with Damien Comolli as Director of Football Strategy, purchased Stewart Downing and Andy Carroll driven by the data analysis on crossing, and it spectacularly back-fired, costing the club millions. Unfortunately, Arnesen, Houston and Congerton suffered similar set-backs and their time at Hamburger SV was largely viewed as unsuccessful. The data driven approach did not appear to work. There were positives, but, in general, there were far more negatives. Lee Congerton and Steven Houston still work together, now at their third club side-by-side. Both are employed at Sunderland AFC in the English Premier League, which shows the strength of opinion they have in each other and they methods they use. Frank Arnesen is currently unemployed, having most recently had an unsuccessful spell in the Ukraine, with FC Metalist Kharkiv.

In fairness, recruitment when Frank Arnesen and his team arrived at Hamburg was very different from what the Chelsea contingent has envisaged when they had signed their contracts with the German club , to such an extent as to make Congerton and Houston almost redundant. The transfer balance the statistical duo assumed they would work with was taken away from them before they even got to the club. At board level, the Hamburg CEO Bernd Hoffman, who Arnesen had signed the contract with in February 2011 whilst still at Chelsea, left in March, and was replaced by Carl-Edgar Jarchow, who took one look at the club’s accounts and decided cut backs had to be made. These cutbacks almost cost Hamburg their precious Bundesliga status. Lee Congerton, who was interviewed by the Guardian in 2012, explains:

‘Once we knew the financial constraints, our priority from day one changed completely and it was very much ‘wir müssen in der Bundesliga bleiben. We were unable to go for the big players that we had been talking to because the salaries were coming down and we had no money to spend, so we had to go for players that we knew and whom Frank could convince to come here. And, ultimately, the owner at Chelsea was very supportive and prepared to help.’

The support from the Chelsea owner came in the form of player transfers from their old club, and in lieu of spending heavily on established stars and the remit under Hoffmann, Hamburg had to recruit under stricter financial constraints until Houston and his data could become established to inform their purchasing. Arnesen brought personnel from his old club on his arrival in the summer of 2011, and Jacopo Sala, Slobodan Rajkovic, Gökhan Töre, Jeffrey Bruma, and Michael Mancienne followed swiftly after. Although they did not add quality to the side, these players padded out a squad who had lost 14 players, including Ruud van Nistelrooy, the Brazilian Zé Roberto, veteran goalkeeper Frank Rost and the Germany international Piotr Trochowski, as a Guardian article explains.

Hamburg had a poor season in 2011/2012, Arnesen’s first. Finishing 15th, their lowest finishing position in the Bundesliga since its inception, the size of the task for the Dane and his data team was becoming apparent. In the background though, Houston and Congerton had a year to establish their networks, watch the players that they thought would be helpful to the club in the long-term, and gather data to help inform signings for a clean slate in the 2012/2013 season.

Artjoms Rudņevs and René Adler joined Hamburg at the beginning of Arnesen’s second full season based on what Houston and Congerton’s data showed they could contribute to the team, but with differing levels of success. Rudņevs, who spent the second half of the 2013/2014 season on loan at Hannover 96, was signed thanks to striking insight revealed by data, as explained to the Guardian by Lee Congerton:

‘Rudņevs’ numbers, sourced from a database of 10,000 European players controlled by Steven Houston, another former Chelsea employee, were “exceptionally high”. The next stage is to view video clips and, if Congerton likes what he sees, as was the case with Rudnevs, a scout will go to watch the player and compile a dossier listing everything from language skills to alcohol consumption. Congerton and Arnesen then traveled to see Rudņevs in action before discussing the merits of signing the player with the manager, Thorsten Fink’.

In regards to Rudņevs, it was a thumbs up from the whole team and the signing, sanctioned by Thorsten Fink, made Rudnevs the first Latvian to play in the Bundesliga when he arrived at Die Rothosen in the summer of 2012 for around £3 million.

Adler joined at the same time. Speaking to Christoph Biermann in 2012, Steven Houston told the interviewer that Adler was bought because of his peak statistical values in terms of shots saved, when referencing the quality of the shots he faced, in the whole of Europe. Adler had also been injured for almost a year, rendering him undervalued. In the immediate, the transfer appeared highly successful. As ESPN journalist Uli Hesse stated in his article about data analysis in October 2013, ‘many people questioned the logic behind signing Adler, considering Hamburg already had a pretty good goalkeeper in Jaroslav Drobny. But now, after six rounds of games, the transfer looks like the most inspired deal Hamburg have made in a long time.’ And so, Arnesen’s job grew more secure and with that, faith and trust in Congerton and Houston’s recruitment models gained value and bore fruit. By the end of the 2012/2013 season, Hamburg finished 7th – a placing that the fans and ownership were more accustomed too.

Steven Houston has frequently said that it wasn’t so much the data they used at that time, but the intelligent way the data they was applied, and the fact that they coupled it with traditional scouting methods, rejecting the opinion that adopting wholesale the Moneyball model was possible in elite European football. Speaking to Sky Sports in November 2012, he stated, ‘stats are central to how we work but you need great scouts. You need them to watch players live. What technical scouting can do is allow you to be more efficient. Scouts can’t watch every game; they can’t watch every team. There are only so many resources you have and it’s about trying to use those resources efficiently. Are we watching the right games and the right players?’ In a nutshell, this was Houston’s job.

Lee Congerton also explained exactly how their data helped individual players that Hamburg already had in their development, ‘Jeffrey [Bruma] has a problem with crosses, with his body position, so we have done work with him on the video showing him clips. We set the model up, the coach understands what we want to do and his assistants will sit down with the player and go through the details.

At its best, the data project at Hamburg bread many imitators, particularly in regards to the player development that Lee Congerton described when helping Bruma (although there is no suggestion this practice was actually sped up by the German team). Video-based analytical coaching in specific scenarios is now commonplace at elite, established first division clubs (and below) across the whole of Europe and beyond. At its worst though, the data project at HSV did not significantly help Hamburg during a period where they needed help the most. The players signed did not contribute to sustained success, or anything like it, nor did they provide a necessary source of income in resale, despite suitors for players such as Milan Badelj and presently Hakan Çalhanoğlu, both bought in the Arnesen era. Frank Arnesen and Hamburg parted ways in late May of 2013, and with the departure his data-driven approach at Hamburger SV concluded.

There are some valuable lessons from what Arnesen, Houston and Congerton tried to do, however. Steven Houston said in 2012 that a mentor told him, ‘99% of your scouting is who you don’t sign’, and it’s clear that mantra means a lot to him. It’s also strikingly similar to the Moneyball sound bite of ‘you can always recover from the player you didn’t sign, but you may never recover from the player that you did sign at the wrong price’. This principle remains a cornerstone of any form of scouting. It is also vital that, if data is used to recruit players, it has context. In a translated quote from the Biermann interview, Houston identified just how easy it is to get carried away by data deprived of context: ‘I had once found the new Messi and I was very excited for 30 seconds, until I realised I had not set the filter of the playing time on the software. Unfortunately, the new Messi had only played four minutes.’ Data won’t show you what you want to see unless you ask it the right questions.

The only source of frustration with what Hamburg tried to achieve during Arnesen’s time there was that they did not have access to the sort of funds that they were initially promised. The timing of the Chelsea contingents arrival and Hamburg’s subsequent league position reflects badly on the trio, but it wasn’t their fault at all. As one Sunderland AFC fan pointed out on the Sunderland AFC fan forum www.readytogo.net when it was announced that Congerton would now be working with the club, ‘has somebody pointed this out to Ellis Short? It looks like Congerton is single handedly responsible for destroying Hamburg.

The only defence to this is to point out, as any effective technical scout would, that correlation does not imply causation.

If you’re interested in this story, click the links below to learn more:

http://www.readytogo.net/smb/threads/2012-article-about-congerton-one-of-his-assistants-and-stats-in-football.892705/

http://www1.skysports.com/football/news/11096/8293811/scouting-revolution

http://www.espnfc.com/story/1180420

http://www.theguardian.com/football/2012/may/21/lee-congerton-interview

http://www.sloansportsconference.com/?p=4558

http://www.11freunde.de/artikel/hsv-scout-steven-houston-im-interview

Be the first to comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.