Moneyball

[COYS]

Somebody mentioned the book Moneyball on this forum a couple of days ago, and it got me thinking about how it could be related to football. I found this article this morning, and thought it was fairly interesting:

Arsene Wenger's Moneyball Strategy, EPL Talk

When Michael Lewis published Moneyball in 2003, it instantly became one of the most influential books about American sports in a generation. Michael Lewis, normally a financial writer, analyzed how baseball’s Oakland A’s managed to consistently produce highly competitive teams despite a payroll that was a fraction of their biggest rivals. From 1999 -2006, the A’s finished either first or second in their division, and won over 100 games in both 2001 and 2002. They won more games than any team in baseball other than the Yankees during that period, but had one of the smallest payrolls in baseball.

The protagonist in Moneyball was A’s general manager Billy Beane, who looked at potential baseball players in a completely different way from other GMs. From Lewis’ perspective, Beane was doing what any good executive should do – trying to bring rationality to a completely irrational market. Beane saw that the market for high-end free agents was over-inflated, and the rational tools (primarily baseball statistics) for properly valuing players were underutilized. Beane was able to use those tools to find outstanding players that were completely overlooked by other teams at a fraction of their market cost.

It would be difficult for Football managers to mimic Beane’s methods in their player selection. Beane relied far more on stats than on scouting to make his decisions, and football does not have the same voluminous stats to hand over to some PhD to find hidden gems. However, there is one manager who has replicated Beane’s unconventional thinking and has consistently produced winning teams for a fraction of what his rivals spend – Arsenal’s Arsene Wenger.

Arsenal has finished in the top four every year for the past decade, and under Wenger has collected three Championships, six FA Cups, and went undefeated for the 2003-2004 EPL season. He has done this while spending a small fraction of what his rivals have spent in the transfer market. If you add up all of Wenger spending since he joined the team in 1996 and subtract his sales in the transfer market over that time, his net expenditures are a measly £17 million – less than £2 million per year. Among the legendary players that Wenger has bought at a fraction of their market value are:

Patrick Vieira – £3.50 million
Robin van Persie – £2.75 million
Freddie Ljungberg – £3.00 million
Sol Campbell – 0 (free transfer)
Nicolas Anelka – £0.50 million
Cesc Fabregas – 0 (signed as youth)
Kolo Toure – £0.15 million
Thierry Henry – £10.50 million

As recently as this calendar year, Wegener seems to have strengthened his team by selling Toure and Emmanuel Adebayor for £39 million and buying Thomas Vermaelen and Andrei Arshavin for £25 million. When you consider that Arshavin cost Arsenal significantly less than what Liverpool spent on Glen Johnson, it really does seem like Wenger is attacking the market in a completely different way than his peers. So how does this manager, who not coincidently has a Masters Degree in Economics, maneuver so well in a completely irrational marketplace?

While Beane had some basic rules in analyzing players (value walks as much alalalalas, discount achievements made in high school, foot speed is overrated, body shape is meaningless, etc.), Wenger has not been so kind as to sit down with an author and lay out his techniques. However, looking over his career at Arsenal, you can come to see Wenger’s own Moneyball strategy. These rules include:

Never buy a star from the EPL, La Liga or Serie A. In fact, try to avoid buying players from these leagues that are in the starting XI of their team. The market has already overinflated their value and you wind up paying too much. There have only been a small handful of players that Wenger has plucked from these leagues, (not counting Sol Campbell on a free transfer and William Gallas as the ballast for the Ashley Cole deal), and two of his biggest transfer blunders (Francis Jeffers and Jose Reyes) have come via this route. Wenger, who has never paid more than £15 million for a player (but has sold a handful for more than £25 million), sees all big stars as drastically over-valued whose price has been driven up by teams like Chelsea and Real Madrid for whom money is no obstacle. Even if Chelsea or Real Madrid is not interested in the “massive” league player you are looking at, their buying power has artificially inflated that player’s costs.

Instead, look for great players in lesser leagues. Big bargains can be found in the Dutch Eredivisie, the Bundesliga, French Ligue 1, the Russian league and many other leagues that do not regularly broadcast their games into English managers’ living rooms. If the French, Dutch, German and Russian national teams are good, it should stand to reason that their domestic leagues should contain some of their future stars. Scouting these “lesser” leagues like the Swedish league (Ljungberg), Eredivisie (van Persie, Vermaelen, Overmars), Bundesliga (Rosicky, Lehman, Hleb), and Croatian League (Eduardo) can yield cheap gold. This is especially true in the French League, with its high emphasis on skill training and its low popularity in the European football world. Wenger has mined France for a wealth of talent over the year, including Manu Petit, Nicolas Anelka, Patrick Vieira, Gael Clichy, and Bacary Sagna among others.

Of course, to find these gems, you must scout them. Wenger has spent money to create the most detailed, professional scouting network of any team in the world. While most major clubs rely on youtube as their main scout, Wenger and his Chief Scout Steve Rowley have built an extensive network which scours Europe and the world looking at players in obscure places dozens of times before making a move. There is no team too minor for this network. Gael Clichy was discovered and signed playing for AS Cannes in France’s 3rd division. Arsenal even employs a full time scout in the US looking for the next opportunity. Wenger firmly believes that every great player started off somewhere obscure, and if he can find those players before they become valuable, the savings will pay for this scouting network many times over.

Once you scout them and sign them, you must train them. Arsenal’s investment in training is massive. Their London Colony training ground, built under Wenger’s close supervision, is state of the art and has become the training ground for the English National Team. From strength and fitness training to nutrition to the best video room set-up in England, Wenger has spared no expense in creating a facility to pour the skill into his hidden gems. Moreover, Wenger is notorious for integrating young players into the first team training at very young age. Fabregas has repeatedly talked about the experience of being a 15 year old trainee playing alongside Patrick Vieira every day in training.

Few teams want to invest in this type of scouting and training effort. It costs millions of pounds to build and maintain this system, and most teams are content to use that money to chase some well-known player from a major team. Yet, from Wenger’s perspective, this investment is cheap. Take just one player as an example. Kolo Toure was discovered in Africa and bought for £150,000. On the training ground, he was converted from a holding midfielder to a central defender. He spent two seasons learning his craft spending every day on a professional-style pitch at the Colony playing against Henry, Bergkamp and the rest of the Arsenal scorers. He then served six outstanding seasons as part of the Arsenal starting XI before being sold for £14 million, an increase of 9333% from his original purchase price. That one player, and his profit of £13.85 million, more than pays for the training ground and scouting network alone.

Thus, when a team like Liverpool spends £20 million for a known commodity like Robbie Keane and sells him again him six months later for £12 million, Wenger must shake his head in wonder. The loss on Keane could pay for a scouting network that would unearth several Robbie Keane’s over the next few years for a small fraction of what the Reds spent on Keane in the first place.

So why doesn’t every team follow Wenger’s economic model? In baseball, many teams copied Beane’s system to great affect (and to the determent of the A’s, which have not been as successful since Moneyball was published). Luckily for Wenger, the market for managers is just as irrational as it is for players. The Wenger system takes an investment of time as well as money that does not immediately create wins, and few teams seem willing to give their managers either of those things. Wenger had the good fortune to be hired by Arsenal vice-Chairman David Dein, a man who made his money as a commodities trader and who understood the utility of finding hidden value in the market. After winning the double in his first full year and profiting over £25 million in the forced sale of Anelka to Real Madrid, the Arsenal board gave Wenger carte blanche to set up this system and has had the patience (like any good investor) to avoid panicking when the results have dipped a little.

As Wenger looks for his third great team at Arsenal to finally be ready to battle for the big trophy, other mangers must wonder how he does it. But they don’t wonder too hard. They do not want to be distracted from their main task – badgering their board to free up money to buy the next big name that comes on the market.

If anybody has read the book, or is aware of any other examples of how the principles outlined in the book have been applied to football (Olympique Lyonnais for instance), then feel free to contribute.

[dena]

I was the one who mentioned Moneyball, as a matter of fact, it's right on my computer shelf!

The subtitle to the book is "The Art of Winning an Unfair Game" and I think it fits perfectly in baseball and football where the money can get pretty crazy and the gulf between payroll/wage bills/budget can get ridiculously unfair.

I don't think it would be that hard to use sabermetrics (the statistical analysis of baseball players) in football players, because every position can be statistically broken down, like this:

Goalkeeper: saves, goals conceded, penalty kicks conceded, corner kicks conceded, passing/distribution %, balls won
Defender: goal kicks forced, balls won/lost, corner kicks conceded, fouls conceded, penalty kicks conceded, passing %
Midfielder: passing %, balls won/lost, assists, fouls conceded, fouls won, corner kicks forced, shots, goals
Striker: shots, goals, passing %, balls won/lost, corner kicks forced, fouls won

or

1. Number of touches
A measurement of how often a player is involved on the ball
What it reveals Player's fitness level, the number of times he gets into a position to receive the ball and team-mates' willingness to pass to him
2. Shot creation
The number of times a player participates in a possession leading to a shot (both on target and off)
Reveals The attacking effectiveness of a player, especially attacking midfielders and forwards. Measures ability to balance ball retention with creating scoring chances
3. Ability to retain the ball
A measurement of the probability that the next player who touches the ball will be a member of your team
Reveals Contribution of players who are less directly involved in attack
4. Balls won per 90 minutes
Measures defensive effectiveness
Reveals Attacking players' willingness to defend; defenders' ability to tackle, intercept passes and position themselves well

Beane also had some rules that he would abide by, like not taking the big flashy name in the draft, or signing someone immediately after a big year (ala signing someone for 5 million + more than he would be after a good show at the World Cup, or Euro's)

I can't really think of any major clubs who have used a similar system, other than Sevilla and Villarreal, both clubs know what types of players they want, and either groom them, scout them cheaply, and sell for profit (Pepe Reina, Diego Forlan, Riquelme, Sergio Ramos, Dani Alves, Keita) and from what I've read Manuel Pellegrini has used statistics in the past to determine who he signs.

[Luxz0rz]

^ How often is that information available though when scouting smaller clubs? I doubt they had all those stats on Clichy

Also it stats don't always mean much...baseball is a more individual sport, especially the pitcher and hitter.

[dena]

A European coach could have an entirely different set of criteria that he wants a prospect to meet if he is to be signed, that was just an example, but if you scout a kid long enough (which you want to do to prevent a bust, right?) Then over a month, two months, even a whole season, you are able to find out "yeah, this kid does what I'm looking for, let's get him into the academy" through the stats, and of course looking at him personally.

And yeah football is more individual but one or two players in each area of the field can drastically change the outcome of a game, and when teams are without their star players they suffer greatly, so even though it is more of a team game than baseball, the individuals still have the same affect.

[Luxz0rz]

Originally Posted by dena

A European coach could have an entirely different set of criteria that he wants a prospect to meet if he is to be signed, that was just an example, but if you scout a kid long enough (which you want to do to prevent a bust, right?) Then over a month, two months, even a whole season, you are able to find out "yeah, this kid does what I'm looking for, let's get him into the academy" through the stats, and of course looking at him personally.

And yeah football is more individual but one or two players in each area of the field can drastically change the outcome of a game, and when teams are without their star players they suffer greatly, so even though it is more of a team game than baseball, the individuals still have the same affect.

This kind of ruins the whole point of it though.

Wasn't this baseball guy just looking at stats and making his decision based on that?

Wenger actually goes out and finds talent, but he doesn't analyse every footballer for a few months and find out who is best.

[dena]

He did both (with a bit of common sense), which I think was the right way to do it. Instead of going for the unproven high school draft pick (which happens a lot in baseball) he went for more established college players who other teams looked over, which is similar to not signing a big name prospect out of South America or something because his club wants 15 million for him, and instead getting a player who is just as serviceable, or you think can be trained in your academy to be just as serviceable, for much cheaper.

For players out in the market he also used the statistcs as a way of determining who got the job done, instead of saying "I'll go out and buy the big name" everyone knows Steven Gerrard, Frank Lampard, and Cesc Fabregas are good, but if there is someone in the Premier League who creates the same chances (not necessarily assists) and covers just as much ground or more in the same position, or something like that, why not buy him instead? He'll be much cheaper which is less of a risk, and if you scouted him well enough he will be able to duplicate those results for your club.

[MFC_LukeCarter]

It's basic Economics and technical efficiency at it's best.

Wenger gets a basic commodity (the player) for minimal input and gets the best out of them by streamlining them through training with the likes of Vieria and Gallas and gains maximum output.

It's a long run process but which remains ongoing throughout his reign as Arsenal manager.

[Gianni]

Originally Posted by dena

I was the one who mentioned Moneyball, as a matter of fact, it's right on my computer shelf!

The subtitle to the book is "The Art of Winning an Unfair Game" and I think it fits perfectly in baseball and football where the money can get pretty crazy and the gulf between payroll/wage bills/budget can get ridiculously unfair.

I don't think it would be that hard to use sabermetrics (the statistical analysis of baseball players) in football players, because every position can be statistically broken down, like this:

or

Beane also had some rules that he would abide by, like not taking the big flashy name in the draft, or signing someone immediately after a big year (ala signing someone for 5 million + more than he would be after a good show at the World Cup, or Euro's)

I can't really think of any major clubs who have used a similar system, other than Sevilla and Villarreal, both clubs know what types of players they want, and either groom them, scout them cheaply, and sell for profit (Pepe Reina, Diego Forlan, Riquelme, Sergio Ramos, Dani Alves, Keita) and from what I've read Manuel Pellegrini has used statistics in the past to determine who he signs.

milan has started using the system you mentioned to determine who plays and who doesn't, as well as for scouting opposition, and determining in which direction training needs to go for a player. it's becoming more widespread here, from what i've heard.

[SG Cesc]

Very nice read, thanks for sharing

[Scuba Steve]

Problem is, the A's have never won the World Series with the moneyball tactic. For their market size, they have been wildly successful in the past decade. But on a larger scale, the fact that the did not return a winner is somewhat damning.

Especially when you compare it to the Yankees of this year. Yeah, the Yankees always spend big, but no one ever wins every year. What was their payroll in 09? $200m? That'll just go to show other owners that you have to spend to win it all.

Maybe if someone won the World Series in baseball with the moneyball strategy, or if several teams used it successfully, it would become more widely accepted.

In terms of football, I am sure there are plenty of ways to break down stats to find out who the most productive players are. Actim, for example, surely has a lot of information on plenty of players.

There are just a couple of logistical problems when it comes to football. Firstly, there are way too many players to fully evaluate. Sure, you could break it down to just the top leagues, but then, how useful would it be?

Another problem is the style of play of each league. In baseball in the US, there is one style of play. Pitch, field, hit, run. That's it. In football (and that's what is great about it), there are so many different styles of playing. Samba Brazilian, tactical Italy, fast paced England, Spain with it's flair, etc etc. It would be hard to put a number on a guy who stars as a winger in Italy, and say that if he reached a certain number, that he would be a success in England.

Finally, I think the style of play within each team makes it hard to find a use for such a system. Sure, the top players in the world could easily slot in on any team. But for the players below that level, there is no way of knowing whether or not a single player would fit in with another team. In fact, it's totally possible that a player works better in a bigger side like Madrid, rather than a side like Real Betis (and that's before we even factor in changes in managers and such).

I suppose the best usage of a statistical system would be for a manager in evaluating his own team. That way, he could analyze how well players perform on the pitch, and possibly adjust according to that.

But, that doesn't account for chemistry and the like, which is a huge factor in football.

[Strange Pirate]

I started reading that book, then I got bored.

[dena]

That's because you don't have an advanced mind.

Another problem is the style of play of each league. In baseball in the US, there is one style of play. Pitch, field, hit, run. That's it. In football (and that's what is great about it), there are so many different styles of playing. Samba Brazilian, tactical Italy, fast paced England, Spain with it's flair, etc etc. It would be hard to put a number on a guy who stars as a winger in Italy, and say that if he reached a certain number, that he would be a success in England.

Finally, I think the style of play within each team makes it hard to find a use for such a system. Sure, the top players in the world could easily slot in on any team. But for the players below that level, there is no way of knowing whether or not a single player would fit in with another team. In fact, it's totally possible that a player works better in a bigger side like Madrid, rather than a side like Real Betis (and that's before we even factor in changes in managers and such).

That's my biggest issue as well, every league is different in it's own way, and you can't add "Premier League multipliers" to stats and say "This guy scored 15 Premier League goals, he'd score 21 in La Liga", which is why it restrics things a bit for players outside of the country, but if you are looking for players within the league and within the football league system, I think it could work out.

[nzgangster]

gonna have to go have a read of this book.

[COYS]

Originally Posted by dena

I was the one who mentioned Moneyball, as a matter of fact, it's right on my computer shelf!

I'm now half-way through it.

It's quite a difficult book to understand if somebody dosen't understand the rules of Baseball or the terminology.

[dena]

The first time I read it I didnt understand some terminology either, I knew batting average and the simple things like that, but once it got into WHIP and stuff like that I was lost. Since that book came out when you go on baseball stat websites like MLB or ESPN.com you find all of these stats next to the player names, it really changed the way people looked at the game.

[Andrés]

I'm suprised nobody has mentioned Sporting CP & Ajax.

They don't do it in the same way as Arsenal but they do pick up players from a young age, develop them and then sell them on for a big profit whilst having success in their own domestic leagues. The main difference between these teams and Arsenal is the way the money is re-invested. I don't think Ajax or Sporting CP put the same percentage back into the club as Arsenal do and it limits them to only being able to look for players locally, which they do very succesfully.

[Senor_Smurf]

This is why I love Wenger. Now he should stop talking before they're on to us. It's fun when they spend £100m+ and we still pose a huge threat to their dreams.

[st3adman]

Originally Posted by Senor_Smurf

This is why I love Wenger. Now he should stop talking before they're on to us. It's fun when they spend £100m+ and we still pose a huge threat to their dreams.

One of the many joys of being an Arsenal supporter

[blueisthecolour8]

This thread is way too smart for me

[polskaGOLA]

This is why Arsenal is by far my favorite club in the top 4.

I think like Wenger to, while playing Fifa Manager atleast. (and in real life as well)

Maybe I'll become a successful manager in the far future.

---- Ok that isn't going to happen.

[nzgangster]

Originally Posted by COYS

I'm now half-way through it.

It's quite a difficult book to understand if somebody dosen't understand the rules of Baseball or the terminology.

i found the book and i am going to start reading it soon

[Zzonked]

Originally Posted by Senor_Smurf

This is why I love Wenger. Now he should stop talking before they're on to us. It's fun when they spend £100m+ and we still pose a huge threat to their dreams.

Originally Posted by st3adman

One of the many joys of being an Arsenal supporter

But you'll never know what it feels like to buy a massive name!

You made get a Ferrari part a day for 5 years and for free and eventually get a Ferrari, but you wont notice it because it happens over time. Whereas we just straight up buy a BMW.

Oh I went all out with my analogies.

[st3adman]

Originally Posted by Zzonked

But you'll never know what it feels like to buy a massive name!

You made get a Ferrari part a day for 5 years and for free and eventually get a Ferrari, but you wont notice it because it happens over time. Whereas we just straight up buy a BMW.

Oh I went all out with my analogies.

Wierd analogy but anyway, i'd rather make good players than buy the best in the world. The feeling of knowing this player has trained with your team since early/late teens and now he is a star is uncomparable to just splashing out on them while they're in their peak.

:Edit: Oh btw, Arshavin.

[Scuba Steve]

Originally Posted by blueisthecolour8

This thread is way too smart for me

The maths is actually pretty simple. Most of it is simple division. Once you get past the terminology, it's easy to understand.

What is boggling is the amount of depth the statistics go to. For a particular baseball player, they could probably tell you what his on base percentage is against left handed pitchers, at night, when the temperature is below sixty degrees, and with one runner on base.

[COYS]

FT.com / weekend columnists / Simon Kuper - Baseball's love of statistics taking over football

Baseball’s love of statistics taking over football

It has taken too long, but at last European football clubs are starting to learn from American sports.

Mike Forde, Chelsea’s performance director, visits the US often. “The first time I went to the Red Sox,” he says of the Boston baseball team, “I sat there for eight hours, in a room with no windows, only flipcharts. I walked out of there saying, ‘Wow, that is one of the most insightful conversations on sport I have ever had.’ It was not: ‘What are you doing here? You do not know anything about our sport.’ That was totally irrelevant. It was: ‘How do you make decisions on players? What information do you use? How do we approach the same problems?’”

Forde, holding forth excitedly from his comfy chair at Chelsea’s health club, is tapping the statistical revolution that has swept American sports. The revolution’s manifesto was Michael Lewis’s 2003 baseball book Moneyball.

Earlier this year, Lewis proclaimed: “The virus that infected professional baseball in the 1990s, the use of statistics to find new and better ways to value players and strategies, has found its way into every major sport.” In soccer, Forde is spreading the virus.

Forde worked at Bolton Wanderers before Chelsea, and he looks like a football man: trim, graying, regional accent, nice suit. That helps him deal with hidebound football men who are wary of fancy numbers spouted by dowdy statisticians. “Letting even a top-level statistician loose with a more traditional football manager is not really the right combination,” says Forde.

He studied psychology in San Diego, and that early American experience proved key. He often visits Billy Beane, hero of Moneyball, general manager of the Oakland A’s baseball team and a soccer fan who grills him on English football’s latest goings-on.

Recently, though, Forde has been studying basketball, a sport more like soccer. “Basketball is ahead of us,” Forde admits. However, he says England’s biggest football clubs now have people in roles like his. “We as a nation are probably more open to the American experience than maybe the French are, the Italians are. Maybe we will be quicker to adapt the Moneyball ideas because of that.”

Adapting those ideas began a decade ago, when clubs started to buy data on the number of passes, tackles and kilometres run for each player. Forde remembers the early hunt for meaning in numbers. “Can we find a correlation between total distance covered and winning? The answer was invariably No.”

People from England rugby told Forde possession won matches. But that did not work in soccer. “If you had 55 per cent possession, the chances of winning were less than if you had 35 per cent possession.”

But the data can help clubs evaluate individual players. After all, says Forde, “most of the elite clubs are probably spending 70 per cent of their revenues on 2.5 per cent of their workforce. Really all we have got is talent.”

Forde sees his task as “risk management”. For instance, he studies data covering a player’s career to avoid the trap of signing someone when he is in top form. A footballer, explains Forde, spends minimal time in the ideal state of flow. “The player thinks that is his normal standard. It is not. My job is to see what form he regresses to.”

The search is still on for the best data to evaluate players. If a forward is tearing up the French or Dutch league, you need to predict his strike-rate in the tougher Premier League. Forde says: “We created our own algorithm: if the guy scores 15 goals in France, is that 10 in England?” Finding criteria to assess defenders is harder. “Is it tackles? Well, look at Paolo Maldini: he made one tackle every two games.”

The holy grail would be discovering the key to victory. “I do not think we are there yet,” Forde admits. But he says: “If you look at 10 years in the Premier League, there is a stronger correlation between clean sheets and where you finish than goals scored and where you finish.” Billy Beane would have been proud.