Welcome Guest [Log In] [Register]
Welcome to Soccer Futbol Forum. We hope you enjoy your visit.


You're currently viewing our forum as a guest. This means you are limited to certain areas of the board and there are some features you can't use. If you join our community, you'll be able to access member-only sections, and use many member-only features such as customizing your profile, sending personal messages, and voting in polls. Registration is simple, fast, and completely free.


Join our community!


If you're already a member please log in to your account to access all of our features:

Username:   Password:
Add Reply
Statistical analysis in soccer
Topic Started: Jun 3 2010, 02:01 AM (323 Views)
hobbes
Advanced Member
[ *  *  * ]
After I obsessed about the up-coming World Cup and spent half of my work day pouring over obtuse numbers from the National Hockey League scouting combine, I thought I’d throw one of my pet hobbies out to this board.

I’ve always been interested in the analysis of the sports I love, but not just in what is seen, but what also can be quantified. I have no formal math or statistics training, but I read a fair bit on different sports that have undergone a sort of stats-based revolution.

Granted sports like baseball and (American) football are infinitely easier to quantify while soccer and hockey are much harder because of their fluidity. Still I do think that sound analysis would debunk some myths and make for some interesting talking points. I find that there’s a lot of accepted wisdom in soccer than just doesn’t make a lot of sense.

To juggle between sports a bit, I find so much of the testing done by scouts in hockey to be almost completely irrelevant. The standing long jump? vertical jump? There’s no sprint testing and while I have no idea if there’s a correlation to speed on the ground to speed on the ice. Regardless shouldn’t skating speed be a factor? I know there’s a perception that Americans needs stats and numbers for everything, but I find it amazing that every half-decent high school football player in the U.S. can probably tell you his 40 time without thinking about it and no one has any clue how fast some of the greatest soccer players in the world are.

Speed isn’t everything, but I do see value in some of testing hockey does like the VO2 max test or the WIngate. And I find it interesting how those numbers stack up across world class athletes in different disciplines.

Because I know the mentality of the people who make these decisions better in hockey than I do in soccer there are some schools of thought that drive me insane. There are players who have a technically ugly skating stride who are in fact fast. I’m sure there are sprinters like this too — you break the stride down and they’re not maximizing their stride, their motion isn’t fluid, whatever. But because they don’t quantify speed and only use their eyes to gauge things they see the ugly stride and give it precedence over the effectiveness of it.

I don’t want to dehumanize the game and make numbers the be-all and end-all. You can’t go too far the other way. I think that’s ridiculous too. Some forms of genius aren’t quantifiable. I know that. How fast or fit was Ferenc Puskas? What would the team of doctors made of Garrincha’s potential with his curved legs? There will always be players who won’t fit a test and will be great.

However, reading Soccernomics I do think that the way people approach the game is out-dated and there’s a lot to gain from more detailed analysis. f

I know there is small cottage industry growing out of statistical analysis of soccer, but those are almost always sold rather than made public. When you see a player subbed out of a match, often you’ll see his distance covered and I’ve always been interested in seeing a lot of these numbers, but you never seem to. So the number is pretty arbitrary. How do you know if that’s good or bad or if it even means anything . . .

I would love to be able to play around with some of the ‘tracking’ software that is used to calculate these distances (because clearly you could use them to determine other measurables like speed), but alas I’m just a schmuck with a cheap computer and no access.

So after all that rambling, I was wondering if anyone has any sites they like that do any statistical analysis. What has the most value to a successful team (work rate? distance covered? possession? completed passes? attempted passes?). Soccernomics discusses how the transfer market has been successfully manipulated by some teams and how bad most are at it, how some clubs are able to make the most out of players who should be past it and a few other interesting topics.

I know we have some people who have coached or do coach here and were wondering if they see any value in the numbers or if they even analyse any statistics. How much is video used in match? I know in hockey video is relayed and viewed between periods and is all coded so you can watch goals, shifts for a specific player or line, special teams, it’s all broken down so you can watch a few minutes of tape from overhead and see if there are adjustments that need to be made.

Anyway sorry for the rambly disjointed post if anyone made it this far.

cheers,
hobbes
Saskatchewan for the CPL: multis e gentibus vires
Offline Profile Quote Post Goto Top
 
Simon
Advanced Member
[ *  *  * ]
I made it to the end hobbes, ha, very interesting it was too.

I've not seen Soccernomics, I take it it's one of the many books derived from Freakonomoics - it sounds like a book I'd like to read.

I agree that football doesn't happen entirely at random, and one could usefully get a better insight into it by studying statistics, but it's always struck me that there are great problems with quantifying football. All of this is a very recent phenomenon in English football, where the stock cliche is that 'there's only one statistic that matters'. But that doesn' stop Sky in particular compiling a huge wealth of stats on each game. But let's say Player A is a winger who puts 10 beautiful pinpoint accurate crosses onto a striker's forehead, only for that striker to miss with every one. Player A gets no 'assists' and yet Player B would get two assists for a couple of aimless balls he hoofs into the box that a better striker sticks in the net.

Likewise, I think statistics have a hard time with people like Romario or even Matthew Le Tissier, who would stand virtually stock still for 88 minutes of a match but would use the other two minutes to score a couple of goals. So did they have a good match or not?
Offline Profile Quote Post Goto Top
 
Onslow
Member Avatar
Advanced Member
[ *  *  * ]
This is an interesting topic and I would say I have been one of those who look at statistics and say, "OK, but who won?" Do the statistics really tell us the story of who won or do they just confirm that fact?

Opta are the stats that Sky Sports love to show during their matches as Simon referred to. Here are their season ending statistics:

http://www.skysports.com/story/0,19528,11096_2705370,00.html

It makes for good reading but does it really matter that much that say Barry Ferguson of Birmingham completed more passes than anybody else in the Premier League? What does that tell us, that he makes side to side passes or that he can make incisive passes forward? Bet it would surprise many that both John Terry and Lucas Leiva are in the league's top five in terms of passes completed. Which to me signifies tha players playing deep in their own end are likely to be near the top of this list as they complete numerous "safe" passes. Although some stats do confirm our beliefs, such as Kevin Davies and Javier Mascherano are at the top of the fols committed list!

Bottom line to me is statsitics make for a good read but football is still a difficult sport which to quantify, how else to explain a sport where one team can control passession 70-30% but the team with less possession won by converting a counter attack into a goal?

Offline Profile Quote Post Goto Top
 
shelsoccer
Advanced Member
[ *  *  * ]
There are a number of companies that do this kind of statistical tracking but, as Hobbes notes, most are commercial. I've run across probably a half dozen, most of which I can't recall. There is a Scottish company, TACTFoot, that I do recall.
Offline Profile Quote Post Goto Top
 
Manzanares
Advanced Member
[ *  *  * ]
Statistics, not just in sport but in many things can be used to support whatever position you want. I was thinking about that a month or so again when we discussed the statistic which was revealed that during one of the Barcelona-Inter semifinals, Barcelona goalie Victor Valdes ran more distance than Zlatan ibrahimovic. Sure Zlatan did not play the full 90 minutes but many were arguing that Ibrahimovic's lack of mobility really hurt Barcelona. But what if several crosses came his way and he put them away, there is somethng to positive to say about a penalty area predator.
Offline Profile Quote Post Goto Top
 
shelsoccer
Advanced Member
[ *  *  * ]
What was the old Mark Twain quote? Something along the lines of, "There are statistics, damn statistics and lies."
Offline Profile Quote Post Goto Top
 
Simon
Advanced Member
[ *  *  * ]
shelsoccer,Jun 3 2010
01:07 PM
What was the old Mark Twain quote? Something along the lines of, "There are statistics, damn statistics and lies."

"Lies, damn lies and statistics" I think, but yeah they can say pretty much anything you want them to.

I recall people having a go at Gerard Houllier for how boring his Liverpool side were, and he refuted that with a dossier of statistics about how many corners Liverpool had had, how many passes completed, how many shots on target etc to 'prove' how exciting his team were.

Needless to say he was the object of some ridicule after that
Offline Profile Quote Post Goto Top
 
shelsoccer
Advanced Member
[ *  *  * ]
Leave it to an Englishman to correct an American about a Mark Twain quote. Well done, Simon.

I know there were some professors at a Liverpool university (John Moores?) that were big into statistical analyses in the 1990s. Perhaps that was where Houllier was getting his stats.
Offline Profile Quote Post Goto Top
 
hobbes
Advanced Member
[ *  *  * ]
Simon> I believe in the UK they decided to go with the much more inflamatory “Why England Loses” as the title.

The reason there was such a change in statistical analysis in American sports is because there had always been statistics, but people started thinking about them and realized many of them were completely useless.

So I’m not looking for numbers for the sake of numbers, but numbers that actually mean something. The ultimate question is if there is a correlation between a certain quantifiable trait or attribute and success.

Simon mentioned a winger getting no credit for an assist after making a great cross to a forward who missed. Which helps illustrate why assists are an inexact statistic. Perhaps completed passes in the area is a much better stat. It would show which wingers or midfielders (primarily) are able to complete a pass into a dangerous area to a teammate. For me that would be interesting.

As Onslow points out Terry tops the EPL in passing percentage. This is incredibly obvious — he’s a central defender on a team that dominates posession because the back line isn’t put under a lot of pressure. I would guess whoever was the best central defender on the EPL champions would be at or near the top of that list every year.

A lot of people in hockey like plus-minus — it gives you a plus if you’re on the ice for a goal scored and a negative for a goal against while you’re on the ice provided you’re at even strength. The problem is that the worst plus-minus is almost invariably the best defenceman on one of the worst defensive teams in the league. Simply put they’re on the ice the most on a team that gives up a lot of goals. You could say that is a sign they’re not very good, but history has shown that if you take that same player and put him on a good team his numbers shift accordingly.

I do think there’s some value for passing percentages in soccer, but you have to compare like for like. You can only compare positionally and even then there’s flaws because not all passes are created equal. I’ve seen people break passing stats into forward passes, square or back passes and long passes (over 10 m). They’re harder to collect, but much more valuable.

The other OPTA stat I find somewhat interesting is take aways, especially in the midfield.

cheers,
hobbes
Saskatchewan for the CPL: multis e gentibus vires
Offline Profile Quote Post Goto Top
 
DealsFor.me - The best sales, coupons, and discounts for you
« Previous Topic · General Soccer Discussion · Next Topic »
Add Reply