For the first time since the 1940s Brentford FC have been promoted to play in the Premier League. I’ve been a Brentford supporter since being taken to the games as a 5 year old by my father, and I was even the team mascot one weekend. The club’s former ground Griffin Park, where I watched my first matches had a pub on every corner, a capacity of 12,000 and was walking distance from my home. On the roof of the stand opposite, it had written
NEXT TIME FLY KLM
For years I looked at the sign and it made no sense to me as a child, even when I learned that KLM was a Dutch airline it still didn’t make sense. Why would football supporters watching a game be interested in thinking about which airline they flew with? And why next time? We weren’t flying with anyone this time, we were watching a football game.
I must admit my interest in The Bees waned as the West London club oscillated between the Championship and the lower division. What now makes this story interesting is the owner, Matthew Benham a physicist / derivatives trader who set up his own sports betting company (Smartodds) bought Brentford FC in 2012, when the club was experiencing financial difficulties. But even before he bought Brentford FC, Benham had also been sending Smartodds statistical analysis to a young Thomas Tuchel who was interested in any data or analysis which would help him to win games. The two men got on when they met, both somewhat maverick outsiders. Tuchel had never played football professionally and only having managed a youth team squad (ie not a senior team) in his career when he was appointed to manage Mainz 05 before replacing Jurgen Klopp at Dortmund. He’s gone on to manage Paris St Germain, then coming to Chelsea and winning the Champions League (on the same day that Brentford won the play-off finals at Wembley).
Generating profits and winning matches
Buying the team that he supported as a boy, has allowed Benham to put his data skills into action, not just risking money in the betting market, but implementing tactics, buying (and selling) players, and looking for other marginal gains (nutrition, injury reduction etc) to try to get the best results on the pitch. He also cut back inessentials, relying on a relatively small squad and axing the youth system. High potential youngsters were bid away by the bigger clubs before they had a valuable contract, so instead Benham has inverted and focused on picking up talented players from the big clubs’ youth academies who aren’t quite good enough to make it into their first team. I used to work for Keith Harris at Seymour Pierce, well known for being the corporate finance adviser for those wanting to buy football clubs, which I wrote about in more detail here. Keith’s view was that you could make money as a football club owner, but only if you did not get involved in trying to pick the team, or carried away with your ego bidding for high profile players. This looks different, because Matthew Benham’s background means that the owner really is likely to bring insights that the manager should listen to.
Most football clubs are loss making, but a few are profitable (roughly an 80/20 pareto distribution). I’ve taken this graphic from Kieran Maguire’s site, which shows the losses for the Championship in 2018/19, Brentford did not win promotion to the Premier League that year, but when it comes to profitability they finished first with £20m of profits. Only 6 other teams in the league did better than breakeven, and of those Brentford’s £20m almost exceeded the profits of the 6 other profitable teams added together (£21m). (Click on the image to enlarge)
Don’t mention Moneyball!
Benham is clearly focused on financial and pitch success. But he hates the term Moneyball. That’s because, according to this quote in a 2015 Guardian newspaper article:
“Moneyball’s idea wasn’t about using any old statistics but statistics as an academic and scientific exercise to see what stats actually helped predicted things. The Moneyball label can be confusing because people think it is using any stats rather than trying to use them in a scientific way.”
That said he is a fan of Michael Lewis the Moneyball author, gifting employees of Smartodds Lewis Undoing Project about Tversky and Kahneman, as well as Kahneman’s Thinking Fast and Slow. If I ever met him, I’d be tempted to ask what he makes of Nassim Taleb’s Fooled by Randomness and William Poundstone’s Fortune’s Formula (both better books in my view). The latter has a great explanation of Kelly’s system of position sizing when gambling, where you work out your edge/odd ratios as well as a profile of Ed Thorp and Claude Shannon.
Benham has a degree in physics from Oxford University, and was a derivatives VP at an investment bank (Bank of America I think) but left around the time when the first internet bubble burst (good market timing!) He then made his second fortune with Smartodds . Looking at the vacancies they seem to offer jobs for people who work in C#.
Love means nothing in tennis
This sounds very similar to Jim Simmons at Renaissance Technology, the top performing quantitative hedge fund. Football, like investing, is a game where the best strategies and tactics tend to win over time. But being a low scoring game there is some luck involved too. Tennis is a high scoring game, you can win a set (6 games) in as little as 24 points, winning each game to love. But on average a set has 60 points played, and it’s first to win two sets (or first to three sets in the big tournaments). Hence a player with superior skill is statistically far more likely to win, even if they are unlucky with a few points. Love means nothing in tennis, but it’s better to be a lucky football team than a lucky tennis player. That’s because football is a low scoring game (one nil is not uncommon, particularly if you’re watching Arsenal play) meaning that luck plays a larger part. Investors also find it very hard to separate luck and skill, Howard Marks has written about it here. Much of his inspiration comes from Ed Smith’s (a professional cricketer) book on luck. So I’m not the only one to notice the similarities between sports betting, luck, and financial markets.
The table does lie
There is an old football cliche, ‘the table doesn’t lie’, suggesting that at the end of the season the best teams are at the top, because the Law of Large Numbers, means that luck is evenly distributed after 38 games have been played. However Benham is happy to insist that sometimes the table does lie, according to his model the year VfB Stuttgart won the Bundesliga they were actually the fourth best team. Similarly a couple of years later when VfL Wolfsberg won, they were the third best team. I think that’s probably right, it’s a binomial distribution problem, although I don’t know the underlying assumptions. Benham also ignores recent runs of form, “mostly noise” he says.
Like hedge funds, in order to maintain an “edge” both Tuchel and Benham are reluctant to talk in detail about their methodology. Just as in finance they have been collecting data sets that others haven’t thought of, or perhaps they have the same data, but their interpretation is superior viewpoint – or probably a bit of both! It depends somewhat on how you define “data”, everyone can watch the same football matches, there is no “inside information”. But Smartodds office is apparently full of employees watching games recording datapoints that get fed into their database. But the skill is knowing the data to keep track of in advance that will then have predictive value.
A game of two halves
But I reckon that the proprietary data sets is only the first half of the story. The second half is the ability to look at the same data that everyone else has from a new perspective. I often see this with investing. I use Sharepad, and write a weekly column for the company. I’m often amazed how investors can look at the same financial releases, analyse the same ratios, but come up with different interpretations of the same data.
Or put another way, perspective matters as much as data/information. Many years after I watched Brentford, I was flying into Heathrow, the airplane following the River Thames as a flight path, and I looked out of the window for a view of West London. Looking down, was a football ground, which I realised was Griffin Park, with the stand “NEXT TIME FLY KLM” written on it…the same information but suddenly viewing it from 32 000 feet it made much more sense. Rather than looking at the stand from ground level, if you’re on an airplane flying into Heathrow, you immediately “get” the joke:
NEXT TIME FLY KLM
Bruce Packard writes a weekly column for Sharepad (hence I’m biased), but I think the product which offers data and analysis for private investors is excellent. For a free trial click on this link or the image below.
Much of this background is taken from
Football Hackers: The Science and Art of a Data Revolution by Christoph Biermann.
Biermann first came across Matthew Benham in 2008. The book isn’t about Benham, but he features heavily because his data driven approach is the subject of the book.
Photo at the top is from Brentford FC.
Kieran Maguires’s site http://priceoffootball.com/