Month: July 2013

Since 2008, 100 different MLS squads have taken the field and 96 of them have played better at home than on the road (measured by goal differential).  The average home team over the past five and a half years has a positive goal differential of +0.47 goals a game.

Home field advantage is a big deal.  But in a league as diverse in geography and supporter culture as MLS is— think 5,000 people at a Chivas USA game vs. 40,000 at a Seattle Sounders game—you would expect some teams to have a greater advantage than others.

In the table below you can see every team’s home field advantage for every season from 2008-2013.  These years were chosen as they represent the contemporary Soccer Specific Stadia-version of MLS.  Home field advantage is measured by taking a team’s home goal differential and subtracting their road goal differential.  This is not necessarily the most accurate possible measure (doesn’t adjust for SOS, etc.) but it represents the most sensible proxy.

Some Takeaways:

The variation from year to year can be quite staggering.  The LA Galaxy, despite playing in the same stadium, went from having basically no home field advantage in 2009 and 2010 to having the largest differential so far in 2013.  Because of this variability, among other reasons, only limited conclusions can be drawn. For example, although Montreal, Vancouver, and Portland all have great home field advantages—both statistically and by the eye test—the limited sample size for all three in terms of games played precludes any broad declarations.

Perhaps the one factor that does appear to have an effect, as Ted Knutson correctly hypothesized, is altitude.  Of the 14 teams that have been in the league from 08-13, the two teams that play at altitude—RSL and Colorado—have the best and third best home field advantages, respectively.  Somewhat curiously, the New York Red Bulls have the second best home field advantage over that same time frame, though it noticeably depreciated once they moved into Red Bull Arena in 2010.

We’ve gained a lot of readers just in the first week, which is pretty awesome. In a way, I guess there were a lot of people who were interested in analytics just waiting for a site like this to crop up. However, because so many of you are new, you might have missed out on some things we wrote back before this site existed, when we only published in the dark corners of the internet.

I’ve migrated some of my writing on football and stats to StatsBomb, and today I want to recap what I’ve been working on this summer.

I spent most of May and June working on a framework to statistically evaluate players. Most of the writing has been focused on attackers, partly because they are far more interesting to write (and read) about, and partly because midfielders, defenders, and especially goalkeepers are tougher to evaluate just using public stats. The thought behind all of it was that most football teams seem objectively bad at finding young, undervalued players. Is that happening because those players don’t exist? Because building reliable scouting networks is hard?  Or maybe because they just don’t understand how to use stats to do some of their scouting for them? There’s now enough public data that I figured I could help out with at least the first and the last one, so I began donking around with the data.

It really kicked off with two different pieces. The first one was called Walking It Into the Net. In it, I postulate that “assists are the game of football’s walks,” meaning assists are quietly a very useful little stat in finding potentially undervalued footballers. There’s also a bit in there about key passes and Schrodinger’s Goals, which has guided my general thinking on this subject since. This search for value also lead to the next piece, which was about one particular undervalued player. That piece was called The Max Kruse Problem.

From there I started writing transfer dossiers. I didn’t want to discuss the big names everyone knew about. I wanted to find guys most people didn’t know about, and maybe draw some attention to them before their names appeared everywhere in the transfer rumors.

Following the Kruse piece, I stayed in Germany, discussing Farfan, Kiyotake, Gonzalo Castro, and Aaron Hunt. Kiyotake has briefly been linked to Aston Villa, but I haven’t heard a whisper about Castro or Hunt.  I’m linking that one first, because it has a methodology outline there, which at this point is pretty obviously a work in progress.

Anyway, I don’t want to go through these step by step. I just wanted to introduce new readers to them – they mostly highlight what the stats say about all of these players, and provide a number of Youtube videos for your viewing pleasure. (Yay, now we can all shop like Sven Goran Erikson!) For the rest, I’ll just link to them, and if you want to check out what I said about some of your (potentially new) favourite players, you can.

Think you might be interested?

Here's a sample of what I wrote on Manchester City newboy Stevan Jovetic.

Name: Stevan Jovetic

Team: Fiorentina

Age: 23

Position: Wide Forward (Left)

Potential Buyers: Juventus, Arsenal, sundry other interested clubs

Looking for a young player who played a lot smarter at age 23 than he did at age 22? Look no further

Passing percentage up. Turnover numbers (Disp+Trn) waaaaay down. Often you’d think that would come at the cost of dribbling stats and key passes, but dribbling only dropped a touch, while key pass numbers went up. This is exactly the type of improvement you want to see from precocious young talents. Basically, Jovetic got quite a bit better at doing the important good things, and better and not doing the important bad things (like giving the ball away).

The only worrying thing about Jovetic for potential buyers is that he only gets 38% of his shots on target. This is somewhat understandable for players who play on teams with lesser talent, as they are often not in great positions for shooting, but Fiorentina were actually quite good. I find myself wondering what his shot position chart looks like.  That said, this seasons’s non-penalty goal rate is at the edge of elite at age 23, and Jovetic has a ton of stuff going for him. Even clubs at the top levels of the Champions’ League would benefit from having Jovetic around, especially if he keeps getting better.

Update: Thanks to the miracle of the internet, Colin Trainor immediately provided me with Jovetic's shot chart. BAM!

What you see there are waaaaay too many shots outside the good shooting areas. Speculative shots from range have terrible expected value, and Jovetic takes a ton of them. Get him with a good manager who can correct his shooting areas, and teams will have a real gem on their hands.

Likely cost: £25M

(Click the player names to go to their transfer dossier.)

Serie A: Icardi, Erik Lamela, Miralem Pjanic, Lorenzo Isigne, josep Illicic, El Shaarawy, Marek Hamsik

Serie A: Stevan Jovetic, Marco Sau, Adem Ljajic, Gomez, Luca Cigarini

La Liga: Christian Tello, Antoine Griezmann, Diego Costa, Thiago

La Liga: Ivan Rakitic, Pizzi, Koke, Iago Aspas, Patrick Ebert

Ligue 1: Ben Yedder, Alessandrini, Ryad Boudebouz, Saber Khelifa

Ligue 1: Dimitri Payet, Yohan Mollo, Remy Cabella, Julien Feret

When the transfer window closes, I’ll do a review of the players that moved, where, and at what price to see how the projections fared. Assuming time allows, I’m also going to look back at the data and see which players would have been flagged last year, and check how they did/whether they moved anywhere as well.

Welcome aboard. I hope you enjoyed the first week around here, and thanks again for all the support.

Cheers,

--TK

14 runs. In the narrative of a Test match, five days of attrition and back and forth momentum, it is a total low enough to ensure that fans on both sides watched its conclusion with nerves shredded. Whilst not quite matching the parallels of Edgbaston 2005, it was a finish close enough to have left both sides harbouring genuine hopes of victory right until the very last ball was bowled.

Australia will be despondent, having hauled themselves back into a game more than once that looked as though it had slipped their grasp, whilst England will breathe a mighty sigh of relief at having avoided dropping the opening game of the a series that for all intents and purposes should have been theirs around an hour or so into the second day's play. Having now drawn breath and with the dust having settled, we can take a look at a key area of the game – partnerships – to assess where and how that absorbing Test may have been won. Heading into the series, England were very much written up as favourites. A stronger side in all areas was aligned with a more settled preparation as opposed to Australia's rather disjointed lead up, having done away with coach Mickey Arthur just a fortnight or so before the first Test. The Australian hopes were centred primarily around their promising pace attack. With England's occasional habit of collectively batting below par, particularly in the first Test of a series, Australia sensed an opportunity. The difficulty for the visitors is that, in comparison to their bowling, their batting line up does not possess either the depth of quality nor the solidity that England's does. In fact, heading into the Test, their order was still not fully known, perhaps even to themselves. Could their top six really put the runs on the board to make a sustained effort at getting the better of England's attack? Here is the fall of wicket breakdown for both sides over both innings:

Aggregate Runs And Fall Of Wicket Table

We can see a breakdown of how the runs for each sides two innings were accumulated:

Runs And Fall Of Wicket Graph

It is evident that with the exception of Australia's last wicket partnership in both innings, the lower orders struggled to put together a sustained partnership at any point, with the line of progress almost flat. This will concern England, whose lower order (after Matt Prior) have struggled of late. It shows that although steady progress was made for the early wickets, without the Ian Bell/Stuart Broad partnership on day three and four, England would have been setting Australia a far more reachable total. If Australia had concerns heading into the series regarding their top six, the first Test will have done nothing to allay them. In the game as a whole, only the opening partnership in the second innings between Shane Watson and Chris Rogers was of any note. While England were grateful for Bell and Broad, without the heroics of Ashton Agar along with Philip Hughes, Australian hopes would have been sunk long before Brad Haddin edged to Prior off James Anderson. We can now also see the breakdown of runs per wicket, which further highlight where both teams struggled

Runs Per Wicket Partnership Graph

  Six partnerships in the game exceeded 50; the aforementioned tenth wicket Australian ones, Bell/Broad, Watson/Rogers and the measured third wicket partnership between Alastair Cook and Kevin Pietersen.  In a game that saw the sides separated by just 14 runs, partnerships such as these were key. In the end, despite Agar and Hughes, England's two key wickets proved enough. If we split each innings into two parts - the first six wickets (when the tail end enters) and the final five wickets - we can look at the weighting of runs that the top six wickets contributed:

Percentage Of Runs By Top 6 Wickets Table

Moving forward into the remainder of the series, the two first innings percentage totals will concern both sides. For England, the over reliance on the top six is not sustainable over a series to consistently post match winning totals. For Australia, unless their top six can improve and post totals to compete with England, they will be playing catch up too often to have a realistic chance of recovering the series. It was a Test match that provided wonderful narratives throughout, but in the cold light of day as we head to Lords, batting will remain a major concern for both sides.

“Negredo to City seems imminent. Any time you can pay £22M for a 28-year-old forward, you gotta take it, right?”

I make snarky comments like this a lot, but talking to readers yesterday, it was made abundantly clear that this whole “value” crusade I thought was totally obvious is actually clear as mud.

It makes an excellent excuse for an article though, right?

Today I'm going to talk generally about age and transfer values, and then specifically about how Manchester City are going about this summer all wrong. One of the problems I have with who Manchester City have purchased this summer is the age of their targets. Ignore the fact that they seem to be overpaying for players no other teams were in for, I want to focus on why paying big money for players age 28 or older is just bad strategy, even when your owners swim through vaults of gold every morning.

Performance Value and the Aging Curve

This seems like a good place to start because it’s one of the areas that seems to cause the most confusion. What is the aging curve? The aging curve is an explanation of how player performance changes at different ages. The human body matures and deteriorates over time, and that will have a strong impact on the performances of elite athletes. Baseball’s aging curve is a fairly long one for sport, and looks something like this (from Baseball Prospectus):

And it’s commonly used to project how players will continue to produce as they age. Baseball is lucky in that they had hundreds of thousands of player seasons to analyse and project future player development. Unfortunately for football, it lived in an information stone age for most of the sport’s life. Despite the fact the game has been played professionally for over 100 years, the only detailed data we can access on player seasons goes back about um... five. Ow. Anyway, the curve for EPL forwards looks like this:

This type of curve actually changes based on what position you are examining. Forwards have the earliest peak and decline, then come midfielders, defenders, and finally goalkeepers. There have been a couple of different studies on this, most recently by Ben Pugsley, who concluded the that the peak for Forwards in the English Premier League starts as early as 23 and lasts through age 26, at which point performances start to decline.

(The earlier study I read - and unfortunately can’t find at the moment - pegged forwards at 24-27, so this mostly agrees.)

This isn’t to say that every player will experience this decline.

As Ben noted in his Top 10 Forward Seasons article, the best performance from the last six years was from a 31-year-old. However, the aging curve is real, and it’s something teams definitely need to keep in mind when buying transfer targets. Example: 24 years old vs. 28 Let’s take the exact same average forward and break him into two time eras. In one era, he is transferred after a breakout season when he is 23 years old. In the second era, he transfers after yet another good season when he’s 27 years old, so the buying team is getting a 24-year-old in one era, and a 28-year-old in the other.

Both times our example player signs a four-year contract. I’ve done a bit of smoothing to Ben’s data, but those are the basic averages for goals per 90 across a broader population, multiplied by a 38 game season. Ignoring price completely, buying the older player lowers the production you are likely to receive by about 22-25%. Elite clubs expect more from their forwards though.

What about if we apply this to someone with a higher level of output?

This would take a forward who averaged 20 goals per season during their peak (an excellent rate), and instead return 15 goals per year. That difference turns what was an elite forward during their peak into merely a good one. World class players are in short supply and fantastically expensive. Good ones are prevalent, and cost far less. Thus it would also lop a massive amount of money off any potential transfer fee for performance reasons alone.

Survivor Selection Bias

What’s interesting is that the actual output expectation of the 30 and older years is likely far worse for the entire forward population you measured back at age 22. The seasons recorded in the studies are biased toward guys who were still playing solid minutes at forward age 30 and beyond. Most of them were outstanding performers in their earlier years and still playing at a high level in the Premier League as they passed through 30.

That doesn’t include all of the guys who dropped off the map due to injury, transferred to easier leagues, or who were replaced by better, younger candidates on their own teams. The PL is now hugely competitive for positional spots, and most teams show no sentiment in chucking older players aside for ones that can generally help them win now.

Asset Value

This is the other big reason transfer values should decrease massively as players get older. A 26-year-old player with two years left on their contract can usually be resold for the full amount if a club finds them to be surplus requirements.

A 30-year-old in the same situation cannot. Example: At age 27, Dimitar Berbatov cost Manchester United £30M to secure his services. Aged 31, and only one season removed from one of the best performances by a forward in recent EPL history, he moved to Fulham for only £4M.  David Villa is a similar case study, and these are the guys who were still pretty good at goalscoring at the time of their moves. For ones that suffer a sharper age-related decline, their salaries are sunk costs.

Wage and Performance Opportunity Cost

Every season, every football team in the world has finite resources.  Even if their owners are willing to spend limitless amounts of money on transfer fees and wages, they have at least one thing they cannot change: playing time. League seasons have the exact same amount of minutes to be filled. Cup runs will be of varying lengths, but even those have a hard maximum of minutes to be played.

In reality, every player budget is fixed as well. Teams have a total budget they are willing to spend on player fees and potential transfers, which is then adjusted by the players they sell, and that’s it. This is especially true now that the Premier League has instituted a new soft cap on salary growth.

What this means is that signing any one player has an opportunity cost in the form of a) their wages and b) playing time. Minutes are finite. Money to be spent on wages is also finite. Buying and playing one guy excludes you from playing another at the same position.

For Champions League purposes, roster spots are also finite. You only get 25 dudes, and even the ones you can name there have some restrictions about locality. Why do we care about this from a value perspective? Because wages and playing time given to declining players over the age of 30 cannot be given to their time-shifted, better performing dopplegangers who are only 25 or 26 years old.

Would you rather your team played the guy who will return 100% on his performance potential or the exact same guy returning 75%? Premier League titles are won by the finest of margins. A 22-25% decline in goal difference last year would have been enough to take the league leaders to third. This stuff matters.

Manchester City’s Summer

As I write this, Alvaro Negredo has been reported as signed by Spanish media for 25M Euros plus 3M in bonuses, so somewhere between £21.5 and £24M. He turns 28 in August. They also signed Jesus Navas this summer for £15M plus bonuses (turns 28 in November), and Fernandinho for £30M, who turned 28 in May.

That’s a total minimum outlay on transfer fees alone of £76M for players who are already in decline.

Look, I think all of them are good players, but that £76M paid will be more like £20M in retained value when these guys turn 30, which is only two summers from now. Additionally, by that time they will likely take minutes from players who could contribute better performances for the team as a whole, simply because they are younger.

There is a correct way to buy older players that you think can help your team - at a discount.

What City have been doing this summer ain’t it. If City wanted to things correctly, they should have paid more for guys who were 23-25, but who would have been entering the peaks of their aging curves. They would have retained a ton of value at the end of their contracts, and not forced replacements for performance reasons 2-3 seasons from now (and at additional cost, but one that is hard to calculate).

Instead they overpaid for guys that are 28 now, will need to be replaced sooner rather than later, and will have almost no residual value halfway through their contracts. City’s owners may be richer than God, but their recent buying spree isn’t logically or analytically sound. It practically ignores the fact that they’ll need every bit of edge they can get this season and in the years to come to further stock the trophy case.

--TK

Earlier in the day Ted Knutson published this (link) on age and value in the transfer market. It's a terrific piece with a focus on Man City's Summer spending and the problems with signing players who, at 28 years old, may have already peaked. Now I know that using words like already 'peaked' when talking about 28 year old athletes may seem like crazy talk but some of the information I collected on this topic shows that the average forward is already starting to decline. To start us off let's look at the age curve. To do this I want to use Time On Pitch % which is the percentage of minutes a player features in divided by the total available minutes (league only).

Age Curve

    This chart shows us some cool information although I must repeat the warnings we posted in the previous article that this is only a 5 year sample. More data is desperately required. Here are some quick thoughts from the Age Curve graph:  

• Young players do not log heavy minutes in the Premier League. An apprenticeship is required, and it is only at the age of 22 that forwards start to play anything approaching regular minutes.
• The age 29 bucket is strange. It does feature a smaller sample, but it could also be the death rattle of desperate forwards struggling against the dying of the light. Or, maybe it's something to do with power forwards/target men.
• If we cast aside that 29 year old bucket, which is a smaller sample anyhow, the we see a peak for ToP% between the ages of 24-27 with a slight Autumn peak at age 28. Depending on your outlook peak years in terms of percentage of playing time could be 24-27 or 24-28.
• If you are a forward aged 30 or over then you can expect the ToP% to fall away quite rapidly. In fact, on average, forwards over this age haven't played this little football since they were 21 or under.

We have the age curve figured out, now I want to show you Shots per 90 and Shots On Target Per 90 performance by age.  

Shots Per 90 And Age

    Yet again we see weird peaks and troughs in the early and late age buckets due to small sample size and talent outliers. The bigger samples are between the ages of 22 and 30.  

• The ability of forwards to generate shots falls between the ages of 23 and 26.
• From age 27 shots per 90 is in decline.
• Arguably shots per 90 is in terminal decline but for the the 32 year old bucket, which has some survivor bias and small sample issues.
• For reasons unknown, players from the age of 27 onwards just cannot generate shots per 90 at the same rate they once did.

Shots On Target Per 90 And Age

    This chart looks at Shots on Target per 90 performance of forwards. We see very similar patterns to shots per 90 with volatile spikes for younger players.  

• Peak years are between 23 and 26 years of age.
• From 27 years old we see a gentle decline in shots on target per 90 output.  Then we have an out of place spike at age 29 and a huge small sample spike at age 32.

Final Thoughts

After an extensive amount of research into the performance of EPL forwards over the last 5 seasons I now have quite the historical database. I recorded player age, shots, team, their scoring percentages. Almost everything, in fact.

~370 players played over 500.000 minutes, took over 19.000 shots and scored over 2000 goals. . There are two simple rules I put in place when logging the data for strikers:

• Penalties are stripped out.
• Penalty shots and shots on target are stripped out

There is also a firm but fair rule which denied many a striker a place in this Top 10 list which we are focusing on today:

• Player must have played a minimum of 60% of the available league minutes.

The minimum minutes requirement nixed quite a few outstanding seasons, but it was necessary to have this limit and guard against smaller samples and phenomenal half season performances (P. Cisse).

The one player who suffered from this minutes (ToP%) rule was Fernando Torres. Torres, in 2009/10, was 25 years old and at the peak of his powers and posted a Goals per 90 of 0.95 and a scoring % of 56.2%. Alas, Torres only played ~47% of the available minutes. It was a tremendous season from the Spaniard but I had to omit it. Rules are rules.

Top 10 Forwards By Goals Per 90

Here is the Top 10 forward performances of the last 5 seasons:

10 - Jermain Defoe (2009/10)

To kick us off, Jermain Defoe did this in 2009/10:

0.60 Goals per 90 and 1.62 SoT per 90 are strong numbers. Defoe's stellar season was powered by strong shooting accuracy and scoring percentage numbers. Defoe certainly played a vital role in securing 4th place (ahead of Man City) in 2009/10.

9 - Gareth Bale (2012/13)

Another Tottenham player, this time Gareth Bale's sensational season from 2012/13.

0.65 goals per 90 from a player who only played a portion of the season in a central position is a mind boggling number. Bale wasn't much of a creator and his Scoring% is below a forward's average, but there is just so much to like here: Durable (ToP%), a shots monster who is above average in shooting accuracy and the best shots on target per 90 number I have on record. All at the age of 23.

8 - Robin van Persie (2012/13)

Another from 2012/13

Robin van Persie's transfer to Man United was not without risk with injury history and advancing age among the question marks against the Dutchman. RvP cast aside those concerns with a season for the ages.

Goals and Assists per 90 were both historically elite as was van Persie's SoT per 90. What of the concerns about van Persie's injury history? A time on pitch percentage of 87.4% will answer that question. Without this performance Man United probably don't win the 2012/13 Premier League title.

7 - Wayne Rooney (2011/12)

As good as van Persie was in 2012/13 Wayne Rooney topped it in 2011/12.

Rooney was sensational in 2011/12 posting 2 SoT per 90, a near 80% ToP% and a strong shots accuracy number. With Berbatov struggling and Welbeck and Hernandez only playing support roles, Wayne Rooney carried this United team in 2011/12.

6 - Sergio Aguero (2011/12)

As good as Rooney was in 2011/12 it was Aguero's goals, and goal, that brought the title to East Manchester.

The second 23 year old on the list, Aguero's first season in the blue of City is the greatest debut season I have on record. Goals and Assists per 90 are strong. Aguero's SoT per 90 is the second lowest on this Top 10 list but it barely matters for the Argentine's performance was powered by an elite Scoring%.

This excellent season was capped by the most famous moment in Premier League history.

5 - Luis Suarez (2012/13)

The third player on the list from 2012/13

Suarez gets a bad rap from some folk for poor shot discipline and shots locations. Despite this, Suarez still posts an accuracy number which sits above average. Scoring% is sub par but the important numbers - Goals and SoT - are Suarez's strongest numbers. In fact, Suarez's 2.19 Shots on target per 90 is the 4th best number I have on record. If Suarez could be finally taught to shoot from better locations then, my word, this would be some player.

4 - Robin van Persie (2011/12)

The second appearance on the list from RvP.

In 2011/12, van Persie was finally fit and healthy and the league was finally able to see what a healthy van Persie could do when an offensive scheme was tailored to his strengths. RvP blew the rest of the league away and that included Rooney's amazing 2011/12 season. Goals, SoT and shooting accuracy were all at a very high elite level. For me, van Persie's Time on Pitch percentage told us this was a player that could finally, and I mean finally, stay healthy.

It was a one-man show of skill and finishing with a unique skill for scoring important (-1,Tied, +1) goals.. Arsenal replaced van Persie's goal tally via committee and many a blowout of the league's bottom feeders.

3 -Wayne Rooney (2009/10)

Rooney was pretty darn good in this World cup season, scoring 40 goals in 50 competitions and often whilst playing up front on his own. Tevez and Ronaldo had flown the nest the previous summer and Rooney was the only remaining superstar forward. And he played like a superstar.

Rooney's 0.73 goals per 90 is the 3rd best I have on record for players who played over a ToP% of over 60%. The goals were powered by a very good Scoring% (40.7%) and sheer volume of shots (3rd highest on record). Rooney's shooting accuracy was below average as were his assists per 90 but none of this mattered much, Wayne was a true shots beast in this season.

United were pipped to the title by Chelsea and lost out in the CL semi to Bayern, the foot injury suffered by Rooney in the run-in was a huge blow to Man United. How times change.

2 - Dimitar Berbatov (2010/11)

What?!

This actually happened.

Berbatov's 29-year-old season is the second best Forwards performance of the last five years. An 0.82 goals per 90 was powered not by overwhelming shots volume, but by stunning scoring% (47.6) and shooting accuracy numbers (42.8%).

Berbatov's 2011/12 season wasn't a consistently outstanding one, but when he was on, he was on. 3 against Liverpool, 2 against Blackpool 5 against Blackburn and 3 against Birmingham. Now these goals are hardly against the highest quality of competition and this may be seen as a knock against the player. The simple fact is, all of the players above have a flat truck bully element to their game, and all strikers are streaky and tend to score in bunches.

The numbers are the numbers and they have Dimitar Berbatov at age 29 as having the second best season on record.

1 - Didier Drogba (2009/10)

Didier is our champ. That Drogba performed this well at age 31 is incredible. Strikers tend to be well in decline by this age, but Drogba rolled back the years and this outstanding league performance enabled Chelsea to pip Man United on the last day of the season.

0.91 goals per 90 is by far the best on record. Drogba's assist numbers are off the chart too. Shots on target per 90 and Scoring% are super elite. Right across the board Drogba just kills it. This was a truly special Premier League forward who in 2009/10 recorded the best individual performance of the last 5 seasons.

Drogba was truly a rare player in that his peak years came much later than is normal for forwards. A warhorse, a beast, a bully. Drogba was all these things, but my he could play.

Visualized

A quick note: I could only go back as far as 2008-9 in this study. I cannot use shots data from seasons previous to that date due to issues with an inferior data supply. A quick glance over the goals and minutes charts for the few seasons previous to 2008-09 tell me Cristiano Ronaldo - the 2007-08 vintage - has the best Goals Per 90 record of any player with 1.016 goals per 90. He was really something.

By Steve Fenn For those unaware of MLS' basic structure, the league is split into Eastern and Western conferences, with the top 5 in each conference qualifying for the playoffs.  The winner of the playoffs earns the league title, the MLS Cup. There is also a trophy, the Supporter's Shield, created & funded by fans around the league to mark the best record in the MLS regular season. The above are representations of minimum points per game needed over each club’s remaining matches to contend for the Supporters Shield or a top three spot in their conference. The second tab does the same for the fourth and fifth seeds, who will have to win 1 match at the 4th seed's stadium in order to meet the 1st seed in the conference semifinals. The point targets for these achievements come from Monte Carlo simulation run by Zach Slaton (@the_number_game) based on the weighting of Martin Eastwood’s (@penaltyblog) EI Index. Click the arrows if you think any of these targets are low or high & the graphs will adjust. Because of the methods for setting these targets, they tend to be slightly conservative. The third tab lays out tiebreakers, which are another MLS diversion from normal football standards. Wins are the first level, goals scored after that, and only if those are tied does goal difference matter.

Hello. Think of this article and it's center piece - the sortable table - as an introduction to what some of my work at statsbomb.com will look like. Of course there will be lots of words and some nice graphics, but I needed a home from my Bitter and Blue home where I could post tables like the one featured below. Sortable tables. Tables that will feature crazy player numbers, individual team numbers and, of course, league wide numbers. I have a ton of information that I collect on each Premier league season (and probably la Liga from 2013/14) that sits in my database lonely and bored that could otherwise see the light of day. At StatsBomb we have ability to post Tableau and sortable tables , which means we can show lots of weird and wonderful information side by side and allow the reader to sift through the table and sort the teams by the strength, or weakness, in any given statistic. Oh, before I forget. The numbers I will post in some of these tables will not look like your average Premier League table. We will feature new statistics, adjusted numbers, Close Game State and lots of other alternative numbers. Want an example? Here is the 2012/13 Premier league table: Sortable!!!!!! [table id=3 /] As you may have seen, this doesn't look like a normal table. Points  is in a funny place, there is no draws column (width purposes) and then there are four statistics that aren't usually featured on any table. PDO A teams scoring% + Save%. League Average is 100 Shot Dom Shot Dominance, or TK ratio in my db. Shots for/Shots against. Ted's link is here SoT Dom The same thing but just using shots on target instead. For the purposes of this table I included Difference to show which teams had a better SoT Dom than they did Shot Dom. Not many clubs manage the feat of having a better SOT Dom than Shots Dom. Man City and Man United skew the entire league. Everton, Liverpool scrape a positive number, West Brom look decent too. The team with the  biggest negative Difference? Southampton with a -0.23.

Final Word

This table isn't too in-depth or radical, and it's not meant to be. This post is a hello of sorts and an example of what I hope to be running here over the course of next season. Premier League tables with PDO, and shots statistics. Tables with save% and scoring%, Game State info. Player per 90 numbers. A whole lot of numbers, a whole lot of alternative table, all running on a week by week basis.

There'll be lot's of other interesting stuff from me in the next two weeks or so. Think of historic player per 90's, an in-depth look at age and decline by position, plenty of team focus articles. Probably a little (big) article on corners, too.

Hello, welcome and we hope you like the site.

We’re halfway through July and basically halfway through what was expected to be an extremely busy transfer summer. And thus far it has been… for seemingly every big club in Europe outside the Premier League. Monaco made a huge splash, Dortmund restocked and reloaded their arsenal up front, and even Napoli have been frisky thus far, presumably knowing they will be selling Edinson Cavani before the window closes, and grabbing some excellent targets to replace his output ahead of time.

But teams in the Premier League – the world’s richest football league – have yet to take the leap. Maybe all the manager swapping made them extra deliberate this year, or maybe after years of binging, they have finally learned some frugality when it comes to buying football players. That said, some teams have made purchases, and as the self-appointed arbiter of transfer value, it is my job to judge them for it. Good buy? Bad buy? Incredibly stupid and overpriced buy? That’s what we’re here to find out!

Teams will be graded on three major principles

1)      Did they get value for money?

2)      Does the player’s age match the price paid?

3)      Did they fill a need?

Principle one is fairly obvious to statheads, but seems to get glossed over in the general media. I absolutely do not care about whether a player’s name gets fans excited. I absolutely do care about their contributions on the pitch, both good and bad, across multiple seasons. Example: Luis Suarez’s headline numbers of goals and assists from last year look really good. However, when combined with his performance a year earlier and crossed with all the rest of his stats, Suarez adds up to a great big pile of statistical meh.

Principle two is also a very important one. Players tend to enjoy their peak years from age 24-28, and then experience a decline into their early 30s, at which point performance can fall off substantially. Signing younger players with good stats profiles has two major benefits. The first benefit is that they will likely continue to improve into their peak years. The second benefit is that they will keep a high resale value if they choose to leave or the team chooses to sell a few years down the road.

Thus if your team is paying big money for a guy who turns 28 this season and signs him to a 4- or 5-year contract, they are paying for one or two years of peak time and then two or three more years of decline, at which point the player will likely have zero sales value. This is “a big deal.”

On the other hand, paying big money for young players without the stats to back it up is a serious gamble, and something that should generally be avoided.  To quote Moneyball, “A young player is not what he looks like, or what he might become, but what he has done.”

Principle number 3 is a tougher evaluation, especially at this point in the window, since there are six more weeks of shopping to be done. For now just know that if a club is spending big money on a position they already had good production at, they are either selling someone, crazy rich, or slightly confused.

Grades run from A+ at the high end down to F at the low end.

Note: The numbers are pulled from the reddit update thread on Euro transfers.

Arsenal – Current Grade: D+

Needs: World Class Forward. Error-free center back. Defensive Midfielder. Pacey left wide forward.

In: Yaya Sanogo (Free)

Out: Denilson (Free), Arshavin (Free), Mannone (£2M), Squillaci (Free), Djourou (Loan)

The rhythm of an Arsenal transfer window typically goes like this:

Leak star target name during season ticket renewals.

Dither, dither, ……… whiff.

Arsene commentates a World Cup/Euros.

Dither. Sign a child………………………………….

Dither, dither, whiff. BuysomeoneFrench thenpanicbuysomeonecheap.

That’s supposed to be different now. Arsenal fans have been told point blank their club will spend money this summer.  It’s mid-July and they obviously have yet to do so. The more things change…

Additionally, Arsenal have a massive amount of dead wood on bloated wages that still need to be trimmed from the payroll, including Bendtner and Chamakh. That’s why right now, they barely receive a passing grade.

If Arsenal buy Higuain (despite their annoyance with Madrid moving the price on them), then this grade immediately jumps to a B+. Higgy is the most efficient forward left on the market, and he’s right in the prime of his career. There are no other targets that would deliver anything remotely close to what Higuain does without radically impacting the rest of the squad. However, if they miss out on Higgy and end up scrambling to find quality up front (Suarez ain’t it, Gomez is already gone), then who knows what this grade will be when the window closes?

Judging by his performance at the U20 World Cup, Sanogo could be a diamond in the rough, but Arsenal’s fans are ready for genuine class that can play now. Along with Higuain, they should be in for guys like Ljajic, Jovetic, Pizzi, Gonalons, and a center back who is both fast enough to make up for Mertesacker’s slowness, and error-free in judgment (unlike say, every other CB they have signed since Campbell and Toure left the club). Whether or not that will happen is another question, especially since Arsenal try to conduct their transfer dealings in complete silence. (This might explain why they rarely manage to complete a deal, though I hear Gazidis is an expert mime.)

There is a very real sense that this team could contend for a title this season if they buy well. There is also a very strong feeling that they lucked into the Champions’ League two seasons in a row, and won’t be able to do that a third time if they don’t open up the checkbook. If Arsenal fail to spend big money on quality targets again this summer, they will see another nine months of miserable fan response at The Emirates, and more protests from supporter groups. Then again, they might get the exact same thing if they sign Luis Suarez. Damned if you do…

Manchester United – Current Grade: ???

Needs: Central midfielders. A happy Wayne Rooney. Fergie’s magic offense. In: Varela (1M) Out: Scholes (Ret)

With the retirement of Paul Scholes, United are perilously short on quality bodies in the center of the pitch. Fergie somehow managed to win his final league title with 34 starts by Carrick, 18 from Cleverly, 17 from Kagawa, and a crazy midfield by committee/whoever was healthy this week. It was a masterpiece of management, but not something that could be repeated a second time, and especially not with The Special One back on the prowl. Therefore United need quality passing, like the kind Thiago Alcantara would have provided if Pep hadn’t snatched him off to Bayern Munich. There are other strong young players out there who are good enough to play in Manchester, but given how deliberate Moyes has been while shopping in the past, there’s an open question of whether we’ll see anyone new in a red devil kit until right before the deadline.

If Rooney leaves, United will also need to get involved in the forward market. If he stays, they need him happy and motivated to stay in shape and employ his amazing talents. I think he’ll stay - United would have been more active in the forward market by now if he were leaving - but I have no idea whether he will be happy.

The other area of concern for United is how their offense will perform without Sir Alex around. United are one of the very few teams in Europe (Barcelona was the other) that employed an offensive system that was truly adept at shooting a bit less than you might expect from an elite team, but creating higher value chances in the process. David Moyes, on the other hand, has never had a team that analysts would consider offensively adept. His defences have often been great, but the best Moyes has managed on the offensive side of the ball is “average.” If Moyes learns from the master and United attack like they did in years past, everything will be fine. But if United attack more like Everton, except with better players… let’s just say it will be a very interesting season at Old Trafford.

Manchester City – Current Grade: C-

Needs: Wide Forwards. Midfield and Center Back youth and depth.

In: Navas (15M), Fernandinho (30M)

Out: Kolo Toure (Free), Bridge (Free), Tevez (10M), Santa Cruz (Free), Kolarov (???)

City were always going to be in a weird spot this summer, with or without Mancini around. Most of their initial purchases after the takeover have either run their contracts out, or are getting too old to fully contribute (Gareth Barry). That meant they would have a ton of wage money cleared from the books, but also need to fill in the gaps with new talent. City owners were also going to have to decide what to do with Tevez, and presumably replace Balotelli and potentially Dzeko as well.

Then Mancini was fired and everything was suddenly up in the air.

City’s owners are aggressively moving to give Manuel Pellegrini what he wants, but they are doing it in a way that is neither cheap nor efficient. I like Navas and his gorgeous eyes just fine, but he’ll be 28 in November, and as a roadrunner-type player, I’m concerned about his age. Players get slower as they get older, and he might lose some of his effectiveness in a year or two. I’m probably more concerned about the fact that they overpaid by potentially £5M for a player almost no one else was in for. He’ll be good, but they should have gotten a better deal.

My concern about the Navas deal pales next to my alarm at the £30M City spent on Fernandinho. He’s already 28 years old, and while he’s likely an excellent player, he blows right through every price to age parameter I could construct that doesn’t involve Zinedine Zidane.  I just can’t see how other, younger targets with similar skill sets weren’t available for that amount of money or who City was competing against that they felt they needed to place a bid in that range for a midfielder. Oh, and the last time he played more than 24 matches in a league season was six years ago, so uh… he’s got that going for him?

Stevan Jovetic is rumoured to be signing with City as early as this week, and he’d be an excellent signing that fills one of their forward needs with an exciting young player. If they get him for £23M or less, that would represent the first decent value-for-money transfer City have made this summer. Beyond that though, City seem to be shopping mostly off their own menu, the contents of which are known only to club insiders and extremely expensive, so who knows what else they will do? (Note: if they buy Negredo instead, I’m marking them down again.)

Luckily for City fans, the one thing they seem to have more of than anyone is money and a willingness to spend it. Even if they are spending too much, they are at least signing good players and aren’t in any danger of going bankrupt.

Everton – Current Grade: B

Needs: To get younger. Scoring punch.  A fairly substantial rebuild.

In: Kone (£6M), Alcaraz (Free), Robles (Und), Delofeu (Loan)

Out: P Neville (Retired)

What happens when Roberto Martinez takes his passing and possession system and plugs in better players, but also ones that might not be wholly suited to it? We’re going to find out.

This is a talented team, but one whose parts are finally starting to show their age. Phil Neville retired and is now on the coaching staff at Manchester United. Distin turns 36 this season, Howard 35, Pienaar 32, Osman 33, Hibbert 33, and Heitinga 30. As we all know, Everton don’t have a ton of cash to spunk on a Martinez-lead rebuild. Therefore further buying will likely happen because

A) Everton sold Leighton Baines and/or Fellaini to a bigger club or

B) there is no B.

If it’s me, I sell both for the best price I can get, and then use the 35M or so on picking up talented younger players that will help build the next generation at Goodison. Political realities being what they are (fans will mutiny if Baines gets sold for anything less than a crazy sum), I’m not sure Martinez will have that freedom.

Regarding the transfers they have made so far, I’m on record as liking Kone, but disliking that price. Kone may help in attack, but he just makes them older and will have no resale value in a year. Since they presumably cost next to nothing, Alcaraz and Robles are fine, but getting Delofeu for a year is a real coup. Delofeu destroyed Segunda last season, and is definitely ready to test himself against better competition (and he’s one of the best silver players in FIFA13 Ultimate Team, a clear analytics indicator of future talent*). Meanwhile, Everton need more speed and creativity in an attack that largely focused around the talents of Leighton Baines last year. Kone and Delofeu should help them play it on the deck more, which may also allow Jelavic to return to the form of two seasons ago.

In short, I think Everton are getting better, but I’d rather see them sell their stars at peak value and reinvest than try to make Martinez’s tactics work with Moyes’ players.

* I am totally lying. Well, about the analytics part.

Swansea – Current Grade: A-

Needs: Friends for Michu. No Capitol One Cup hangover.

In: Bony (13M), Canas (Free), Amat (2.5M), Pozuelo (425K), De Guzman (Loan again)

Out: Gower (Free), Ashley Williams (Maybe?)

After being extremely frugal their first two seasons in the Premier League, Swansea seem to be comfortable now and splashed out big for Wilfried Bony (presumably at Laudrup’s behest).  This makes a lot of sense to me, as it’s hard to see Swansea as relegation candidates anymore – they have done a good job securing their place in The Show. Because of that, they should likely be willing to spend a bit more on good players, especially at the offensive end of the pitch. Look at it this way – for most of last season, Danny Graham was Swansea’s 2^nd best forward after Michu. Now Swans fans can look forward to another season of De Guzman (who was important to them last season) setting up Michu (presumably as a second striker now) and Wilfried Bony. That’s a big upgrade, and one that will help mitigate any regression from Michu’s second season in the Prem. (I am saying this knowing full well Bony will not replicate his goal rate from the Eredivisie in England. Er, and Wales.)

The rest of Swansea’s buys so far have bolstered the defense. Part of this is cover for if/when Williams moves, and part is likely an acknowledgment that they just weren’t good enough when Chico was injured last season. They allow too many shots per game and especially too many shots on target for a team whose system allegedly involves a lot of resting on the ball.

I really like what they have done so far. I’m not fully convinced Laudrup is anything more than an average manager, but he definitely has an eye for talent, and Swansea should once again end up safely midtable at the end of this year.

Wuzzat?

They signed Jonjo Shelvey from Liverpool?

FOR SIX MILLION POUNDS?!? That’s three Michus!

Sigh.

Revised Current Grade: B-

Tomorrow: Chelsea, Liverpool, Spurs and more.

A new sports stats site with a silly name. Why is it here, and why should you care? In November of last year, I was diagnosed with cancer. A year prior, I had a traumatic incident with the family jewels and hadn’t felt quite the same since. One of my kids had a cold, so I took them to the doctor’s office and casually mentioned my issue to the practitioner. Forget all the crap about not liking doctors and such, my son was starting school in the next year, and I figured paying attention to my health was the responsible thing to do.  Anyway, the nurse checked me out and then recommended me for an ultrasound at the local hospital. Two weeks later, I was on the operating table, having literally the first major surgery in my life. Thankfully, I had a good cancer. Good cancers are the types where you have your surgery, take your medicine, go through the excruciating repair and recovery process, and then you have a very strong chance of getting better. Bad cancers are the ones where you do all of that, but your chances of getting better are a coinflip or worse. Let me tell you, going from being a fit, active male to having your hip locked out, and not being able to walk properly for a month sucks. But it’s a helluva lot better than not getting the chance to see your kids grow up, so having a positive attitude about all of this was easy. The first book I read after my surgery was Nate Silver’s The Signal and the Noise. I’ve worked in sports betting for over seven years now and been a data geek for far longer than that, so Nate’s stuff is inherently enjoyable for me. Reading his section on PECOTA made me wonder if the football analytics community ever got off the ground, and once I got out of the hospital I began bebopping around the internet, looking for information. What I found was that there was now good stuff all over the place, but it was scattered to the four winds. Most of it existed on personal blogs and there was no central hub for the analytics community (though credit to James Grayson for trying).  In a way, it wasn’t that big of a deal because even in the moneyball era, football stats and analytics is pretty niche, right? The other thing I realized when I was sick was that I really missed writing. In a past life, I not only edited a fairly large gaming site, but I also used to get paid for traveling around the world and writing about a card game. The whole process of working through problems, and communicating thoughts and ideas suddenly seemed really important to me. So I did what every person who has ever been on the internet does, and started writing on my blog. Given my output in recent months, I apparently had a lot to say. Fast forward eight months and both the use of analytics and the community itself have exploded. Data is more readily available than ever, there are a ton of interested, clever parties writing about it, and teams and companies are certainly snapping up major contributors with interesting job offers. There’s even an actual, exists-in-paper book about analytics and football. But there’s still no central hub for publishing, everyone puts  most of their material on their personal blogs, and the community exists almost exclusively on Twitter. What do you do if you’re a new person that wants to get your stats writing out there? As a fan, how do you possibly keep up with all the excellent writing these days? Obviously, I wasn’t the only one noticing or experiencing these issues. The same thing was happening at the same time to the incredibly prolific Ben Pugsley. His home base was on the Manchester City fan page BitterandBlue.com, but what about when he wanted to write about all the other clubs? Or leagues? At that point, he either had to shop his stuff to specific fan pages, or plonk it on his own blog where it would be read in drips and drabs. Mutually frustrated, we put our heads together and decided to create our own site. What is StatsBomb? For starters, it’s going to be a place for analysts to publish their work on a website with a bigger, more regular traffic footprint than their personal blog. It’s also going to be a place where fans who are interested in this stuff can come every day for useful writing, info, modified league and player tables, etc. At some point in the coming months, we’re going to add a forum to the site, and hopefully that will become a bit of community hub in the process. The organization of some of the site will also change pretty regularly as we go along and discover better ways to display and update the information readers are interested in across the various leagues. We’re going to start with football, but I don’t see us as limited to that sport. Cricket will be a natural fit, as will basketball and hockey. And I don’t see us limited to just stats either. If the site grows and fans respond, I see StatsBomb growing into a place you can come to for good, thoughtful writing about sports and/or numbers. At the very least, it’s a home for at least Ben Pugsley and I to dump all the stuff we’ve been working on that doesn’t live anywhere else. This includes my player analytics work, and Ben’s extremely useful stats transforms. At the very best, it could become something special. If you are interested in being involved with the site at some level, get in touch, whether it’s involves hosting your work,  creating graphics, updating tables from your favourite leagues… whatever. If you’re a reader and there are stats or features you want to see, get in touch and we’ll add those to the list of potential topics. Anyway, welcome. Poke around, see what you like and don’t like, and don’t be afraid to give us feedback. Cheers, Ted Knutson - @mixedknuts Ben Pugsley - @benjaminpugsley StatsBomb Editorial - @statsbomb statsbomb@gmail.com Special thanks to Donnie Noland for creating the site logo.