Football Stats

[InOffMeLeftShin]

Anybody know of a good place to find Premiership stats?

Things like: Minutes played, goals, open play goals, set piece goals, assists, shots, passes, completed passes, pass completion %, tackles, interceptions, blocks, headers, dribbles, crosses, team goals conceded etc.

 

[kazzah9]

WhoScored or Squawka

 

[ShelfSide18]

Anybody know of a good place to find Premiership stats?

Things like: Minutes played, goals, open play goals, set piece goals, assists, shots, passes, completed passes, pass completion %, tackles, interceptions, blocks, headers, dribbles, crosses, team goals conceded etc.

If you want to go really advanced, go to Man City Analytics...

http://www.mcfc.co.uk/home/the club/mcfc analytics

Beware, they send you a huge XLS document which has literally every action that occurred, by player & match, in the Premiership last season.

Whoscored is excellent as Kazzah9 says, I use that a lot. Also, if you have an iPhone/iPad then you can download the 442 Opta App which is also very good.

 

[InOffMeLeftShin]

Cool, cheers guys. I was reading a really interesting basketball statistical analysis by John Hollinger, who devised a formula to determine a players effect and efficiency. It basically rates each statistical data towards contribution of points vs. team pace and league averages. It's a behemoth of a formula but reading the explanations was fascinating. There is enough data on WhoScored to do a very similar analysis for the Premier League.

 

[ShelfSide18]

Cool, cheers guys. I was reading a really interesting basketball statistical analysis by John Hollinger, who devised a formula to determine a players effect and efficiency. It basically rates each statistical data towards contribution of points vs. team pace and league averages. It's a behemoth of a formula but reading the explanations was fascinating. There is enough data on WhoScored to do a very similar analysis for the Premier League.

Have you read Soccernomics at all? I think you'd be interested in it - they cover this sort of thing and how data analytics are transforming how PL clubs recruit players.

 

[InOffMeLeftShin]

Have you read Soccernomics at all? I think you'd be interested in it - they cover this sort of thing and how data analytics are transforming how PL clubs recruit players.

I haven't, but you're probably right, I would be interested. To be honest I've never been a fan of football statistics, certainly not until recent years where they've improved the way they use them. I'm not convinced by the use of them in finding players, nothing like a pair of eyes.

 

[ShelfSide18]

I haven't, but you're probably right, I would be interested. To be honest I've never been a fan of football statistics, certainly not until recent years where they've improved the way they use them. I'm not convinced by the use of them in finding players, nothing like a pair of eyes.

It's the age old argument isn't it, I'm a big fan of using them, and improving our way of interpreting them - of course our eyes are the main thing here, but there's nothing like a bit of help with numbers. It is described in that book as 'driving with a dashboard'.

There are examples where managers have made poor decisions due to stats, e.g Fergie selling Stam when his no of tackles per game decreased one season - perhaps he was reading the game better? Liverpool recruited Henderson and Downing because they created a lot of 'chances' the previous season, did they then dig into this stat and watch videos of all these chances, were they quality chances?

Sky flashed a stat last night about the no of assists in the EPL since David Silva joined City, and Chris Brunt was in the top 5 - is he a creative genius? It would be a bad move to recruit him on that basis solely, but I'd be interested in watching all those 16 assists, were they quality assists? What kind of assists were they - of course alongside lots of other stats too (passes completed in the final 3rd are a favourite at the moment) you can build up a picture alongside the main thing, using your eyes, and clubs will send scouts to watch many games when thinking of recruiting a player.

Occasionally though you can find a diamond in the rough, an undervalued player for example Flamini first came to the attention of Wenger due to the amount of Kms he was covering per match in France, he then went and watched him and realised he was good, bought him for peanuts then sold him to Milan for 14million after a few good seasons. Stats and eyes together.

I'd give it a read, it's very interesting.

 

[InOffMeLeftShin]

It's the age old argument isn't it, I'm a big fan of using them, and improving our way of interpreting them - of course our eyes are the main thing here, but there's nothing like a bit of help with numbers. It is described in that book as 'driving with a dashboard'.

There are examples where managers have made poor decisions due to stats, e.g Fergie selling Stam when his no of tackles per game decreased one season - perhaps he was reading the game better? Liverpool recruited Henderson and Downing because they created a lot of 'chances' the previous season, did they then dig into this stat and watch videos of all these chances, were they quality chances?

Sky flashed a stat last night about the no of assists in the EPL since David Silva joined City, and Chris Brunt was in the top 5 - is he a creative genius? It would be a bad move to recruit him on that basis solely, but I'd be interested in watching all those 16 assists, were they quality assists? What kind of assists were they - of course alongside lots of other stats too (passes completed in the final 3rd are a favourite at the moment) you can build up a picture alongside the main thing, using your eyes, and clubs will send scouts to watch many games when thinking of recruiting a player.

Occasionally though you can find a diamond in the rough, an undervalued player for example Flamini first came to the attention of Wenger due to the amount of Kms he was covering per match in France, he then went and watched him and realised he was good, bought him for peanuts then sold him to Milan for 14million after a few good seasons. Stats and eyes together.

I'd give it a read, it's very interesting.

The Stam issue is pretty baffling. I find it very hard to believe that someone like Ferguson would first use stats to justify his decisions and second that he would use them so simplistically. The number of tackles per game is a pretty poor indicator of how good a player is defensively used as a single stat like that. The top tackler in the league last season wasn't Kompany but doesn't mean he wasn't the best defender, he just had less to do than other players playing in teams that don't keep the ball as well. It'd be much better using a function of team possession, goals conceded and tackles rather than simply using a raw stat like tackles.

 

[OmarsComing]

I read that Fergie sold Stam because of his comments in his auto-biography.

 

[Kendall]

Early in the 2001–02 season, Stam was controversially sold to Lazio in Italy, after Manchester United manager Sir Alex Ferguson was reportedly furious with allegations Stam had made in his autobiography Head to Head about himself and the club. Stam made numerous statements in the book about his views on opposing players, and notoriously alleged that Ferguson's approach to buy him was done without the permission of PSV Eindhoven.[5] Laurent Blanc was signed as his replacement.
However, Alex Ferguson has since described the decision to sell Stam: "At the time he had just come back from an achilles injury and we thought he had just lost a little bit. We got the offer from Lazio, £16.5m for a centre-back who was 29. It was an offer I couldn't refuse. But in playing terms it was a mistake. He is still playing for Ajax at a really good level."[6] On the financial report Manchester United announced the fee was £15.3 million.[7]

Essentially Ferguson thought 15.3m was a good deal for a nearing 30 year old...

 

[InOffMeLeftShin]

Essentially Ferguson thought 15.3m was a good deal for a nearing 30 year old...

And he was right.

 

[Bowlesinho]

And he was right.

I may be wrong but I thought it had something to do with the drug ban he got later on whilst at Lazio........Fergie found out and got rid asap

 

[ShelfSide18]

The Stam issue is pretty baffling. I find it very hard to believe that someone like Ferguson would first use stats to justify his decisions and second that he would use them so simplistically. The number of tackles per game is a pretty poor indicator of how good a player is defensively used as a single stat like that. The top tackler in the league last season wasn't Kompany but doesn't mean he wasn't the best defender, he just had less to do than other players playing in teams that don't keep the ball as well. It'd be much better using a function of team possession, goals conceded and tackles rather than simply using a raw stat like tackles.

I should have said - I'm pretty sure it wasn't the sole determiner for his departure, but apparently a factor none the less.

I'd imagine 15 million for a 30 year old was a big one too!

 

[Bus-Conductor]

It's the age old argument isn't it, I'm a big fan of using them, and improving our way of interpreting them - of course our eyes are the main thing here, but there's nothing like a bit of help with numbers. It is described in that book as 'driving with a dashboard'.

There are examples where managers have made poor decisions due to stats, e.g Fergie selling Stam when his no of tackles per game decreased one season - perhaps he was reading the game better? Liverpool recruited Henderson and Downing because they created a lot of 'chances' the previous season, did they then dig into this stat and watch videos of all these chances, were they quality chances?

Sky flashed a stat last night about the no of assists in the EPL since David Silva joined City, and Chris Brunt was in the top 5 - is he a creative genius? It would be a bad move to recruit him on that basis solely, but I'd be interested in watching all those 16 assists, were they quality assists? What kind of assists were they - of course alongside lots of other stats too (passes completed in the final 3rd are a favourite at the moment) you can build up a picture alongside the main thing, using your eyes, and clubs will send scouts to watch many games when thinking of recruiting a player.

Occasionally though you can find a diamond in the rough, an undervalued player for example Flamini first came to the attention of Wenger due to the amount of Kms he was covering per match in France, he then went and watched him and realised he was good, bought him for peanuts then sold him to Milan for 14million after a few good seasons. Stats and eyes together.

I'd give it a read, it's very interesting.

I though Ferguson sold Stam because he was unhappy with things he'd written in his autobiography ?

The rest of your post is bang on. It should never be about stats, but a combination of as many facets as possible.

Comolli made the same mistake, I believe, with Bent and Bentley. Basing his purchases on their stats, without giving anywhere near enough thought to the fact that neither suited our team or style of play at all.

I've read Soccernomics. The Billy Beane/moneyball story is even more interesting. It was about Beane using his instinct and nous about the game and players combined with his ability to interpret stats and how they worked and fitted into the dynamics of his team at a given time.

Here is a link to a review of a book about statistical and data analysis, by Nate Silver I believe, and it's evolution and applications (from weather, politics, terrorism, volcanos, gambling etc etc) but there is a section in it that stood for me, particularly the section in bold:

http://www.thesundaytimes.co.uk/sto/newsreview/features/article1121380.ece

As I have studied the habits and techniques of expert forecasters, I have found they tend to share some essential traits — humility, adaptability, a certain tolerance of complexity. But a particular skill kept coming up again and again: they were all masters of a statistical principle called Bayesian reasoning.
Thomas Bayes was an 18th- century English mathematician; Bayes’s theorem is nominally a mathematical formula, but it is really a whole philosophy for what we should do with our wealth of information.
Bayes’s theorem asks the forecaster to begin with an estimate of the probability of a real-world event. It does not require us to believe that the world is intrinsically uncertain — it was constructed in the days when the regularity of Newton’s laws formed the dominant paradigm in science. However, it does require us to accept that our subjective perceptions of the world are approximations of the truth.
Unless we have grown up playing cards or other games of chance, we are not encouraged to think in this way. Maths classrooms spend more time on abstract subjects such as geometry and calculus than probability and statistics. In many areas expressions of uncertainty are routinely mistaken for admissions of weakness. But it is when we are overconfident about our predictions — whether they are about the stock market, a football game or housing prices — that we fail.
In my forecasting work I like to think I have discovered a few of the secrets of prediction. One should never rely on raw data alone — the best forecasters draw on a combination of facts and interpretation, and the more information you are able to gather, the more accurate your predictions will be.

 

[Bus-Conductor]

Edit, replied to SS18 before reading down. Omar and Kendall both mention the book thing. Oh, and I don't believe the offer from Lazio was it at all. Fergy doesn't sell players for money, ever IMO.

 

[yanno]

I've read Soccernomics. The Billy Beane/moneyball story is even more interesting. It was about Beane using his instinct and nous about the game and players combined with his ability to interpret stats and how they worked and fitted into the dynamics of his team at a given time.

From the wiki page on Billy Beane:

http://en.wikipedia.org/wiki/Billy_Beane
When the Athletics ownership group agreed to purchase the reincarnation of the San Jose Earthquakes of Major League Soccer, Beane, who has expressed a passion for soccer, began developing a system for objectively analyzing soccer players. He has agreed to help the Earthquakes front office develop a method for building a cost-effective team, as the salary cap in MLS is even more restrictive than the Athletics' status as a small-market team in Major League Baseball.[48]
He has most recently claimed that Arsenal manager, Arsene Wenger, is a personal idol. His friendship with Damien Comolli and Arsenal owner, Stan Kroenke, allowed him to delve deep into the world of English football. He has also claimed to be a "closeted Arsenal fan".

So, Beane is linked with San Jose Earthquakes, who are our US partner and to whom we've loaned Simon Dawkins - who is a good example of a "Moneyball" value player in that he is relatively effective for his (cheap) price. He's also a friend of Comolli.

However, if Comolli is a disciple of Soccernomics, then he relies too much on the stats and forgets to use his eyes and footballing judgement.

To anyone who understands football, Jordan Henderson is never a £20 million midfielder. Brendan Rodgers has barely played Henderson, and is now picking Shelvey ahead of him after signing Joe Allen for £15 million.

The Billy Beane "Moneyball" Oakland Athletics tale is essentially about identifying effectiveness in a sport and then putting a value on it. In baseball, stats reveal more about effectiveness than they do in football. So, an "effective" player can be identified more easily through stats in baseball, and part of the "Moneyball" myth is about ignoring variables such as a player's lack of charisma or poor body language.

In football, which is far more of a team game than baseball, even stats such as passing need to be carefully analysed. The pass completion rate of a creative player who has the ability to make defence-splitting passes, such as a Hoddle or a VDV, should be lower than the pass completion rate of a defensive midfielder whose job is to break up play and play it simple, such as a Deschamps or a Makelele.

The goal and assist stats of Luka Modric would never suggest he's a £30 million plus player. The scouts of Real Madrid, Chavski, Man Shitty and Manure all had no doubt that he was worth that much and more.

Oscar has just scored two cracking goals for Chavski. ITK claims we were offered him for £3 million a year ago, and obviously that represents ridiculously good "Moneyball" value. However, at £25 million, Chavski overpaid and his two goals tonight do not mean that he will definitely become a £25 million player.

All clubs presumably use their scouts to put a value on a player and then make judgements on whether to buy that player based on how much he is available for in the market.

Our signing of Pienaar for c£3 million was fantastic value in "Moneyball" terms, but for whatever reason, he never got a serious run in our team and didn't settle. Levy ended up making a profit on Pienaar.

Our signing of Bentley for c£16 million was terrible value in "Moneyball" terms, in that the player was never as good or effective as his stats suggested, and he didn't fit our team.

So, ultimately, leaving to one side the cheats like Man Shitty and Chavski, for whom value for money is irrelevant, I suspect every club tries to operate a version of "Moneyball".

If you sign the likes of Modric and Berbatov for around £15-16 million each, and integrate them into your team, then you're playing good "Moneyball". Levy doubled his money on both players.

Ferguson paid £30 million for Berbatov and ultimately failed to use him properly or integrate him into his team, then sold him for £5 million having hardly played him at all for a season. The £30 million for Berbatov might have represented "Moneyball" value if Dimi had become the Bulgarian Cantona. But he didn't, and ultimately Fergie played "Moneyball" badly.

 

[Bus-Conductor]

Beane's model was about value for money, but not just in the buy/sell sense, but in obtaining value in service too, but how he achieved that wasn't just by analysing pure stats. He understood their worth and how to interpret them and only used them in tandem with his own understanding and evaluation of players. There were players who's face didn't fit, were unpopular characters, and he realised that by ignoring this prejudice he could extract what he needed for his team.

We get this all the time in football. Just look at the disparity in popularity over Hudd/jenas or Dawson/kaboul/Gallas (when Kaboul was first here).

The examples you use are good examples. Modric's gaol/assist rate is nothing special, but we use other evidence to understand his strengths, including an appreciation of the role he has been given in the framework and tactics of the team.

Truth is Berbatov is every bit is good as Cantona technically and productively IMO, probably even more use to a team than Cantona in fact, in cold hard productivity, but Cantona had that quality that will appear on no stat record. He had enormous charisma, it effected those around him on and off the pitch, and for that reason he was treated differently to Berbatov, who lets not forget, was Utd's top scorer two seasons ago. This is a prime example of how popularity influences even great minds.

My point was, using the Beane example and the Silver example above, is that it is not about pure stats. It is about knowing how to use them, and using them in conjunction with various other mediums, an important one of which is human intuition.

Comolli may have qualities, and some of his purchases were good, but he showed really poor judgement on several occasions now. Bent, Bentley, Henderson, Carroll to name four massive mistakes which represented four of the worst purchases made in football IMO. Terrible value in both footballing and financial sense.

 

[yanno]

Truth is Berbatov is every bit is good as Cantona technically and productively IMO, probably even more use to a team than Cantona in fact, in cold hard productivity, but Cantona had that quality that will appear on no stat record. He had enormous charisma, it effected those around him on and off the pitch, and for that reason he was treated differently to Berbatov, who lets not forget, was Utd's top scorer two seasons ago. This is a prime example of how popularity influences even great minds.

My point was, using the Beane example and the Silver example above, is that it is not about pure stats. It is about knowing how to use them, and using them in conjunction with various other mediums, an important one of which is human intuition.

Totally agree.

As I've written before on SC, I believe Ferguson bought Berbatov to be his new Cantona - the conductor who orchestrated the team. For a season or two, it loooked like this would work. Then, for whatever reason, Fergie seemed to decide that Rooney needed to play the deep striker role, and the likes of Chicarito or Wellbeck, with their exceptional pace, would play up top. Having made this decision, and snubbed Berba by leaving him out of the CL Final squad, Fergie should have sold Berba whilst he could have got most of his money back. Instead, he let him rot last season, hardly playing him, and then sold him for a paltry £5 million. This for a player like Sheringham, who relies on vision and technique, and could play at the top level till his late 30s, if he still has the desire.

Fergie fucked up with Berbatov. But he's still the Teflon manager - partly because he's the Alpha Dog amongst Alpha Dogs, and bans any journalist who asks him a remotely challenging question.

Spurs paying c£15 million for Berba, and Jol and Ramos in particular making him the hub of the team, playing as the deep lying striker, represents a near perfect "value for money" purchase.

Manure paying £30 million for Berba could have been "value for money", except that Fergie changed his system and after the first two seasons was never able to integrate Berba into the same team as Rooney.

Fulham paying £5 million for Berba, if Jol can keep him interested, could be the "Moneyball" purchase of the season.

Comolli may have qualities, and some of his purchases were good, but he showed really poor judgement on several occasions now. Bent, Bentley, Henderson, Carroll to name four massive mistakes which represented four of the worst purchases made in football IMO. Terrible value in both footballing and financial sense.

I believe it was John W Henry of the Boston Red Sox and now Liverpool who offered to make Beane the highest paid GM (similar to a DoF in that they sign and sell the players for the coach to use) in baseball. (If I'm wrong, please correct me.)

If Comolli was attempting some sort of "Moneyball" philosophy as Henry's DoF at Liverpool, then Comolli's record surely makes him the Anti-Beane.

He is the Rowan Atkinson of Moneyball - buying ineffective players at truly insane prices.

 

[InOffMeLeftShin]

The goal and assist stats of Luka Modric would never suggest he's a £30 million plus player. The scouts of Real Madrid, Chavski, Man Shitty and Manure all had no doubt that he was worth that much and more.

For sure, but we all know that simply goals and assists is a far too simplistic way to analyse a players performance even using stats. Goal production is a function of games played, shots taken, etc. In fact if you made each stat a function of resulting and contributing factors (as Hollinger's PER) you can begin to compare the values of different players in different positions. It can never ever be an accurate reflection as stats simply don't exist for how effectively a player can close down opposition, the quality of off ball movement etc. that obviously effect how good a player is and how valuable they are to a team. Also the raw stats don't express any sort of 'clutch' factor although if the raw data was all analysed it'd be easy enough to distinguish between whether a goal is a game winner (or even a draw earner) with those that are the 4th or 5th goals against a team that is being demolished.

I have managed to sort of produce a rough calculation based off Hollinger's PER to relate to football and I have the data for all players who played at least 20 games last season (as found from WhoScored.com). Fairly interesting the results although it is an absolute shed load of data and calculations gone into it. Modric despite only having 4 goals and 4 assists in 36 games actually came out with the 6th highest PER of the 249 outfield players that played at least 20 games last season when taking into account all facets of the game and their 'rough statistical contribution' to a performance. Note that most of the theories are taken directly from Basketball PER and probably have to be more deeply revised than I bothered to do at this stage. But I think the order the players come out in is interesting considering the 'first impression' of the raw stats.

 

[sbrustad]

For sure, but we all know that simply goals and assists is a far too simplistic way to analyse a players performance even using stats. Goal production is a function of games played, shots taken, etc. In fact if you made each stat a function of resulting and contributing factors (as Hollinger's PER) you can begin to compare the values of different players in different positions. It can never ever be an accurate reflection as stats simply don't exist for how effectively a player can close down opposition, the quality of off ball movement etc. that obviously effect how good a player is and how valuable they are to a team. Also the raw stats don't express any sort of 'clutch' factor although if the raw data was all analysed it'd be easy enough to distinguish between whether a goal is a game winner (or even a draw earner) with those that are the 4th or 5th goals against a team that is being demolished.

Exactly. You can add to that the difference in how teams in football line up against each other based on several different factors. Lets take the purchase of Sigurdsson for example. Last season he played in a newly promoted team, and came from the german Bundesliga. As such he was for the most part an unknown, and was offered lots of time and space on the ball by opposing teams. The opposing teams were playing Swansea, and would generally be pretty attacking minded.

This time around it's different, and although Sigurdsson is clearly very talented he is now up against opposition that will try to contain us and play defensively. While you can obviously account for these things playing the "moneyball" game there are simply too many variables, and low value for money buys are always going to happen simply because some transfers just don't work out.

I just read a thread on a different forum about Moneyball in Baseball and came across this gem. It sort of explains the huge difference between the sports and why it can be hard to apply moneyball principles to football.

"Baseball is great partially because it's the only sport where, on the field, every single player can be completely greedy and do what maximizes their personal value every single play and it's to the team's best interest. Every other sport is a zero sum game, where every shot you take is one your teammate doesn't--in baseball, every hit you get provides another opportunity. There's no clock besides individual failure. If Ayn Rand and Adam Smith fucked, she'd give birth to a baseball"