By Phil Blundell
STATISTICS. They’re a recent phenomenon in the game of football, and I’ll be honest, they drive me up the wall.
Football is not a game suited, in general, by statistical analysis, and the new craze of associating any little statistic to back up some point is starting to get ridiculous.
American Football – the NFL, not where Robbie Keane has always dreamed of playing – is the perfect game for statistics; there are few variables and what the statistic says is generally what happens. When a quarterback throws a ball 50 yards and it’s caught, he’s advanced play 50 yards. The game stops. His job is to move ball up the pitch, away from his own end zone and towards the opposition’s, and he’s done so by 50 yards.
When a running back runs 40 yards and is tackled, the game stops at where he’s advanced the ball. His job is to advance with the ball as far as he can up the pitch, and he’s done so by 40 yards.
When a defender tackles someone, play stops. The defenders are there to either turn the ball over or prevent the attacking side achieving a first down. A pass is a pass, a tackle is a tackle, an interception is an interception.
It is a very simple game to realise when a player does and doesn’t perform his job, and a statistic is a strong gauge of how well a player is performing.
If a wide receiver fails to collect a catch, play is rendered incomplete and continues to the next down with the ball returning to where it started from – it’s easy to see he’s doing a bad job of being an outlet for the quarterback; on the contrary, it’s easy to see if he’s doing a job when he makes a lot of catches. If a corner back catches a lot of the opposition quarterback’s passes, it’ll be reflected with statistics and it’s easy for everyone to see how that player is doing.
Anyway, you get the picture and that’s enough of the American Football exemplars.
Football is far too fluid and dynamic for constant statistical use. The fluidity of football renders so many of the statistics quoted obsolete when you take them their literalness and explore them thoroughly.
Throughout his early time at the club, Lucas was defended by some on the grounds that he had an extremely high pass completion rate; while on paper this looks great, it only reflects a small portion of reality. It simply claims he received the ball and found a red shirt – but this is football and the game isn’t that simple.
We don’t stop the game when a completed pass is made and ready the side on the pitch for the next phase of play like the aforementioned NFL. It didn’t tell us what the passes actually achieved; it didn’t tell us what happened once the player got the ball off Lucas and how easy his job was made; it didn’t tell us if these passes actually went anyway to achieving anything at all. It simply said he gave the ball to a team mate.
Sometimes an incomplete pass is preferable to a completed pass. You’re more likely to score, after all, if you try a through ball to a striker, but obviously there’s less chance of success than passing it back to your goalkeeper.
Lucas could have played 100 hospital balls in 90 minutes with his team mates collecting them all. Would that be a good thing? No. Would he have had great statistics? Yes. He could have got the ball and played it back to the goalkeeper every time. What would that have achieved? A boss statistic. It’s a pointless statistic that tells us absolutely nothing whatsoever other than he simply gave the ball to a team mate. It means nothing.
Tackling is another area that can be easily misrepresented. What is a tackle? How do you correctly decipher when the ball is won and lost? If a player harasses an opposition attacker, puts his foot in, fails to get the ball but causes the opposition to give the ball away because of the pressure applied, then he’s failed to win a tackle – but he’s got the ball back for his team.
Statistically it reflects badly, but your eyes will tell you differently. His pressure has seen the attacker give his side the ball back. Where’s the problem with this? There isn’t one, it’s a good thing; yet if he did this three times in a game, he’d end up with terrible tackling statistics, which occur because he caused his side to regain possession.
One thing ignored is when a tackle is won, what actually happens next? If a tackle ends up as a through ball to the opposition striker who goes one-on-one with the keeper, the player gets credited with a successful tackle that has put his team in a worse position than they were before. Not all tackles are good tackles, nor are all intelligent ones.
Over the summer I saw one of the more ridiculous statistics: Liverpool’s most effective tackler last season was Raul Meireles. I say ridiculous, but the more you think about it, the easier it is to decipher why it may be the case.
Meireles was highly selective when it comes to tackling, only really attempting a tackle if it looked as if he’d come out
of it unscathed – there’s no 50/50’s in his statistics, as he ran the other way if there was a possibility of facing one. He had a higher percentage success rate than any other player, which some will regurgitate. But my eyes saw him not tackling when the odds aren’t in his favour, so to my eyes he wasn’t a good tackler – but statistics say he’s our best.
I’m happy to entertain the notion that I could be wrong and he’s a fantastic tackler, but letting the opposition run off with the ball isn’t criticised statistically, it’s ignored. Context, again, isn’t incorporated into this analysis of performance.
But nothing quite gets my goat like assists. Certain players, usually wingers and attacking midfielders, are in the team with one of their primary goals being to put the ball on a plate for a striker to score; this quite often gets represented with an assist statistic. ‘He’s got 10 assists this season, he’s been boss him lad.’
Football Manager, Fantasy Football. Football simplified to the nth degree. They rely solely on the striker and what he does with the ball. It might not sound it, but five assists can be better than 20; the usual assumption would be the person with 20 is creating more, but that’s not true. If you place a League Two striker in your side, he’s going to convert fewer chances than a world class striker. It’s only logical. If a winger puts 50 chances on a plate for his centre forward and he misses 48 of them, is that really worse than another winger putting five chances on a plate for his striker who manages to convert three of these? Of course it isn’t, but the player who has created five chances is statistically better, in assist terms, than the player who has created 50. Nonsense.
There’s some valid statistics, but they’re the ones that aren’t really seen or contemplated; they’re the statistics that go in to more depth and explore context. What we see in every day football analysis are goals, assists, passes completed, tackles won – statistics that are in very little context and focus on the simple facts.
Out of two strikers, the more clinical one isn’t necessarily the one who scores the most goals. The more creative player isn’t necessarily the one with more assists, the better passer doesn’t necessarily find his team mates with the greatest regularity, and the best tackler doesn’t necessarily win the most tackles.
“Statistics are just like mini-skirts; they give you good ideas but hide the most important thing”. Who are we to argue with Ron Atkinson and his lovely tan? Next time you try and prove a point using a statistic, have a think about what you’re trying to prove and what the statistic tells you.
And if you see or hear me trying to back up a point with a statistic, feel free to punch me in the face.
Ha, ha. Very good read and well written, nice one!!
Oh, just had a thought. You should tell this to Paul Tomkins!!
Allan, how very banal of you. Not only is this article aimed at the lowest common denominator but you sunk below it.
If you’d bothered to read any of Paul Tomkins articles fully you’d find he goes into much more detail than just goals scored or assists made. He does look into what more meaningful stats.
Maybe have a quick chat witty Comolli and Kenny as to how the players bought in the Summer were statistically what we needed whereas most supporters and tabloid media were questioning signings like Henderson.
Very good! Although I think some use of statistics can be very revealing. There is a lot of meaningless stuff out there these days. The stats that gave annoyed me are the ones that compare a player against another player but only using stats from last season. This was used to ‘prove’ Downing was better than Ashley Young. I always though Young was the better player. It was also used to prove Henderson and Adam were great signings. You see, they both were in the top 5 for ‘key passes’ last season. Interestingly though there were no Man Utd players in the top 10 despite them scoring the most goals last season.
Variables exist in all statistical analysis. As a lover of both football and American football, I can tell you that stats are equally as useful or as useless in both. Your American football examples are as simplistic as if I were to say a goal is a goal, an assist is an assist, a tackle is a tackle, etc. in arguing on behalf of the usefulness of pure stats in footy.
For example, how would you rate a QB in American football who throws 20 perfect passes, but 4 of them are dropped and 2 are intercepted after hitting the receiver in the hands but he couldn’t catch it and it deflected directly into the hands of a safety/defender? The stat sheet would show a horrid performance by that QB, but in reality it’s not the case.
Or how about a QB who completes 27 out of 30 passes for 325 yards and 4 TDs. MVP type performance, right? What if the majority of those were 2 yard, dump off passes to a running back in the backfield who made phenomenal plays to miss tackles and scurry down the field, all in spite of an inexperienced and indecisive QB?
There’s a million other examples of how you can’t rely purely on stats in American football, just like most other team sports. The bottom line is that in team sports, individual stats, to a large degree, rely upon not just the performance of the individual but on related performances of individual teammates as well as the team as a whole. I’d argue that stats are most useful in purely individual sports, e.g. golf, tennis, etc.
YNWA!
amen.
Couldn’t agree more with the statement about Rauls tackling, however no matter what way you break down the stats about Lucas he was fans player of the season last year. Thats not a stat its a fact and for me he is shaping up to be the best defence midfielder in the prem, and thats an opinion much like your own but it seems that the majority of fans that subscribe to the offical LFC site agree with mine. I don’t subscribe to the offical site but for you Alan i do happen to subscribe to the tomkins times
Jeremy explained very well why American football statistics are bad on the offensive side of the ball. The same is true defensively. If you have a great cover corner (eg Deon Sanders), QBs wouldn’t throw the ball his way. so he got very few tackles/interceptions, while the guy on the other side got loads. But he also got beat a bunch of times, while Deon was rarely beaten. Point is, you can’t take any individual statistic in isolation, but have to use all of them.
So MerryLes might have a great tackle success rate, but you can temper that by saying he achieved 5 successful tackles every season, while Lucas has one every 5 minutes. That illustrates the point you were making, using statistics. ;)
Just because it is a young discipline, and many people use them badly does not mean that statistics have no place in football. It just means that you need to understand the limits of the statistics and the bounds of what they can tell you…
Stats is an interesting subject and one which polarises opinion. There’s an interesting article in this months Esquire about an American Baseball player turned General Manager who sought to take emotion away from decisions in sport and look at statistical and scientific data when looking to acquire new players, one of his more high profile ‘disciples’ is a certain Damien Comolli. The full story is a book called Moneyball apparently soon to be made into a film.
As someone that is very much into American football and has a much more natural ability to analyze it and break it down than football/soccer, baseball would have been a much better analogy. American football is probably the game where players are most misjudged on statistics despite there being a stat for almost everything, since it’s a game where one particular player’s statistics are the result of a multitude of factors and is very much dependent on other players’ performances around them. But I get what you’re saying.
As you mentioned in your article, the buzz for statistical analysis in football has only picked up recently, and I think the fact that it’s not something traditionally associated with the game has people misunderstanding or overrating their value and their proper utilization by decision makers like Damien Comolli. Also, the popularity of fantasy sports has a hand in this as well I think, as you mention. There are players in both American football and football/soccer who are stat monkeys, they produce big numbers(say a running back who plays behind an elite offensive line and with a talented pass offense) or a lot of one relevant number(someone like Darren Bent who is not elite but if nothing else scores goals). These players can be seen through by simply watching them play most of the time or looking at more statistics that help better paint the picture.
Simply, statistics are nothing without context. They’re tools. They don’t tell the story, they just help tell the story. And if relevant trends are recognized, they’re useful. People like Comolli don’t simply just use numbers to paint the entire picture and let that tell everything. If a stat like chance creation(the hotly discussed one with Comolli and his transfer dealings) in the past has been a great indicator of on-the-pitch success, than it’s a stat worth looking at as a factor.
At the heart of it statistics are just numerical representations of occurrences on the field. It’s up to us to determine their value and what they mean. If trends say that “Stat A” is a relevant dictator of on-the-pitch success, it’s something worth factoring in. If some guy on an internet forum is arguing some point and can manipulate his argument by throwing in a statistic or two that can mislead, than you need to do some research yourself.
Example – if someone says a certain running back is the best in the league and then uses a statistic like he led the league in rushing with 1500 yards, that in itself could manipulate readers. But what if it’s brought up that he only averaged 3.8 yards per carry(lacked per-carry efficiency), played for an overall good and balanced offense, and was poor in goal line situations and only had 3 touchdowns. Certainly, he’s not better than the running back who accumulated 1350 yards but did so at a rate of 5.2 yards per carry and finished the season with 10 touchdowns.
The most important thing is to understand what the stats actually represent and how they could be accumulated. It all goes back to something simple whether you use statistics or not, you either understand the game or you don’t. Someone’s argument of midfielder A being better than midfielder B because his pass completion percentage is better doesn’t hold merit if you don’t get the types of players they are, their supporting cast, coaching scheme, etc. The thing is, if you actually understand the game, then you can understand what the numbers mean as well. Anyone who watches football/soccer knows that a “safe” passer of the ball like Lucas is going to have a higher pass completion percentage than Charlie Adam. Sure, there are going to be fools who will wave around numbers haphazardly to make some weak argument, but those can be seen through by more intelligent people. Statistics themselves aren’t the problem, it’s the people who don’t know how to use them for relevant analysis that are.
I personally have a keen eye on statistics myself, I regularly write posts on the subject matter. I find it to be a very useful tool to help understand a players strengths/weaknesses, but the key words there are help and tool.
There is nothing better than the naked eye to make a judgement on a player, but you need to have a trained eye to spot some weaknesses/strengths at times. Not always the case, but you certainly have to know what you are looking for.
The example used on Raul Meireles is a good one. Possession duels 50/50’s. Raul probably did win his fair share of 50/50’s and that is what the statistics say. The problem was, how many of them were 50/50s and how many times did he just back out of the challenge, meaning it was 0/100’s so to speak? So it wouldn’t be registered as a “statistic”.
Which brings me nicely onto what is a “statistic”. Well, using Opta as an example, clean sheets, Goals, Assists, are all fairly easy to measure. The difficult ones are crosses, chances created, possession duels. What is a duel, and what is a tackle? What is a cross and what is a long pass? They all have a definition, which is subjective of course. Where as a goal really isn’t. A goal, is a goal. And 99% of the time, it is pretty obvious who scored it.
The thing is, whether people like it or not, statistics in football are a growing trend. Not just because we have FSG (who are open about the statistical use in Baseball), or because we have Comolli (who is open about how looking that these sort of trends, have helped him identify certain transfer targets). But what good is all this if you don’t have context?
A little example.
Second top “chance creator” in the league last season was Kevin Davies. Amazing (in one way), as the rest of the top 10 are made up of players who take a lot of set plays corners/free kicks. Kevin Davies is a player who uses what ability & strengths he has, and uses them very well. to his credit. But is he really one of the most creative players in the league?
It is all about the way Bolton play. They play some neat stuff at times, but Davies is the focal point of their attack, no doubt. So naturally he is probably involved in more attacks than most other players. He is the target for crosses/corners/free kicks, and he will win his fair share of knock downs, creating chances for other players. He isn’t going to be breaking down a defence with a through ball, or delivering quality crosses into the box, as that is not his game. Davies can’t “create chances” on his own so to speak, he needs someone to set him up to create the chance. So, you need another “creative player” to make Davies creative. The stats say he is the second most creative player in the league. But in my opinion, he isn’t.
Another example is Nani and Modric. Both players had the same number of “created chances” last season (both had 64). Nani had 14 assists, Modric had 2. Why is that? Is Nani creating the better chances, and more are converting? Possibly, but without looking at every single one, it is difficult to say. More likely the case however is United’s 4 main strikers (Rooney, Berbatov, Owen and Hernandez) scored 46 goals. Spurs (Defoe, Crouch, Keane & Pavyluchenko) scored just 24.
There are multiple examples of this sort of thing. Just putting up a chart/table of players and their statistics only tells half the story. Or, in the example with Kevin Davies, it may tell you a misleading one. There are a lot more articles/statistic based posts around now (even Rafa has started doing them), but where some differ, is they try to add a picture to them. This post – http://www.eplindex.com/liverpool-stoke-acid-test/ is one of the best examples I can think of doing things properly, and well. It takes the scenario of how we struggle against the likes of Stoke, and have done for a few years. Looks at why, and then explains how we can “beat” them. The last two posts he has done have been very accurate (read back if you want – V Arsenal – http://www.eplindex.com/liverpool-arsenal-scouting-report/ V Bolton – http://www.eplindex.com/liverpool-bolton-preview/ )
In summary, I think this is something that has gone on for years in the game, bu it is only now that Opta have started to licence out the use of them to 3rd party websites (not just the Premier League, Sky, etc), that they are much more in the public domain. Now, more everyday people have got access to them.
Benitez is fairly open about his use of statistics in the past. Villas Boas has dossiers (with stats, and much much much more) on many of the players he has worked with (and others). As do many other managers, I am sure. But these are “experts” in their field. They use them, as well as their trained eye and instinct, to be some of the best in the business.
Give someone this book “Path Integrals in Quantum Mechanics, Statistics, and Polymer Physics, and Financial Markets”. Anyone can easily tell you stats and facts from within the book itself, but how many could actually use them in a way that makes sense? That actually “works”? And can explain them? Not everyone, and football will be no different.
Whilst some facts in football are misleading (Raul’s tackling being one of them) some statistics don’t lie because they just can’t. With this in mind I think that we were keen to pursue Stewart Downing over Ashley Young because of his better and more consistent appearance record at Aston Villa as much as his stats as one of the “top crossers in the Premier League”.
It’d be a very naive manager who would base his entire recruitment policy on stats alone but they can certainly flag up certain aspects of a player’s arsenal, for good or bad when they’d investigate and start watching individual players thoroughly – I remember hearing of a Steven Gerrard/Rafa Benitez discussion where Benitez had been saying Gerrard wasn’t shooting with his left foot enough, stats backed him up and lo and behold it lead to him scoring a winner with a left foot shot (might have been Man City at Anfield a few years back?).
Lies, damned lies…
Very poor article and read.
It is clear from the body of your post that you have been waiting for a glorious opportunity to tear apart statistics and their (valid) presense in the modern game and you found that in the Meireles stat about being our most effective tackler which we all agree to be nonsense. Other than that your article was totally one-eyed and biased. Your intent becomes clear when you start on Lucas and I think most sensible people will have spotted that.
Also your comments around assists are embarrassing – what universe would a striker who has had 50 chances and converted 2 still get a game? Or a winger that creates just 5 chances over the course of a season keep his place?
As has been said earlier this article is shamelessly aimed at the lowest common denominator with a massive chip on it’s shoulder.