As rain began to drizzle down on the Estadio Palogrande pitch, the tournament hopes of the Young Socceroos, who were trailing 1-0 to Ecuador in their opening game of the Under-20 World Cup, were starting to look as bleak as the clouds above them.
That’s when it happened.
Standing above a prone ball 30 meters from goal, FC Utrecht winger Tommy Oar took a deep breath before striking a swirling free kick that fooled Ecuador goalkeeper John Jaramillo and flew into the back of the net.
1-1 and Australia had snatched an unlikely draw.
Later that evening, as I sat in a mostly empty press box, I remember looking at the possession statistics and thinking to myself how little it told us about the game.
Australia had only just been shaded in the possession stakes (49 percent to Ecuador’s 51) but they’d been clearly outplayed – Ecuador controlled the game even when they didn’t control the ball.
It got me thinking about how useful the possession statistics, which are routinely broadcast around the world at half time and full time of games, actually are.
24 hours after Australia’s game I was in Medellin to watch England draw 0-0 with Argentina.
At half time the stadium announcer read out that Argentina had controlled 54 percent of possession compared to England’s 46. Yet it didn’t tell you a single informative thing about what had transpired over the preceding 45 minutes.
What was interesting is that England had ended up with 46 percent possession at half time when at the 20-minute mark they’d only seen 37 percent of the ball.
In the 30th minute midfielder Reece Brown had dropped into defence to replace the injured Nathan Baker and all of a sudden the amount of long balls the English defensive line were playing seemed to decrease.
Brown’s distribution into a rather open Argentine midfield proved vital in helping his side get back into the contest.
Yet while these two possession figures together highlight how and when England wrestled back control of the game, they don’t reveal the inability of Brian Eastick’s side to create clear goal-scoring chances.
This was a trend that continued throughout a game where England was arguably the better side without being good enough to win.
Which leads me to my main point – we are often hesitant to embrace statistics in football due to the fluidity of the game but I believe this is an over-reaction.
As an example let’s consider “shots on goal” and “shots on target” figures. In truth these numbers, at least when considered alone, really don’t give us a clear indication of how well a team is shooting.
This is because there’s a great difference between hitting a target when one on one with a goalkeeper and shooting from outside the area while an on rushing opponent is closing down your shooting angle.
By tracking more specific information on shots at goal you could find something out about the game that are impossible to gauge by simply aggregating them all into one lump “shots on goal” sum.
In fact, these statistics do exist, and most astute top class teams use them, but they aren’t the ones football fans tend to be fed.
So just because the statistics we consume are irrelevant doesn’t mean there isn’t a way to use numbers to further highlight and inform us on what has transpired in a game.
The true fallacy of football statistics is not that they are pointless, but that maybe we are yet to have discovered those that are truly revealing.
clayton
Guest
Anyone ever hear a story along the lines of ... SAF decided to move Jaap Stam on because he had an inkling that he was starting to decline, and this was seemingly backed up by statistics that showed that Stam was making fewer tackles than before. He moved on to Italy (I think) and went on to have several more very good seasons. At a guess - the reduction in tackles made was due to Stam reading the game and not needing to make so many tackles, a la Canavarro? And heat maps? KMs logged in game? I worry that they encourage headless chicken running around.
BigAl
Guest
But, but . . . what about Fantasy Football ?
Kasey
Guest
I'd like to see a stat that either proves or disproves the "2-0 is the most dangerous scoreline in football" line Sabermetrics was invented in the USA to either prove or disprove commonly held axioms in the baseball world....if you really are interested in statistical analysis of sports I suggest you read the book Moneyball....about the application of a new form of baseball stats to helkp the small market Oakland Atrhletics choose better players beforte they became really good and famous and could be bought by the big payroll Yankees. Saber = (Society for American Baseball Research SABR)
Kasey
Guest
North Americans have a seemingly unquenchable thirst to break down their sports into statistics, and I see them using Ice Hockey style 'assists' and 'points' in MLS to track forwards scoring prowess. In my experience of the major NA sports, Hockey is the closest to football in its rhythmic and consistent flow of action. One stat I could see as furthering the in-depth analysis of football comes from Hockey. Its the +/-(plus-minus) stat as it directly measures the effectiveness of the players on the ice(pitch) when the team scores a goal, everyone on the ice at that time gets a +1 in their +/- column. If the team concedes, they get a -1. Penalties do not count. To adapt it to football if a team is short handed and scores the players on the field might deserve +2 instead of +1and if a player gets sent off s/he gets -1(increasing to -2 if the opposition scores after s/he is sent off. this stat would help in directly comparing two midfielders and perhaps the midfield of two competing clubs prior to a game. If player A had a plus/minus of +15 over a season one could reasonably conclude that his/her team played better with them on the pitch and were more dangerous in front of goal, compared to a player with a +/- in the negative. just throwing it out there, can anyone see problems with it, besides the impost in some numbers guru to actually calculate it.
Fussball ist unser leben
Roar Guru
Interestingly, Sheek ... Contrary to popular misconception, a 0-0 scoreline is one of the less likely scores in Football. In fact, based on statistics from the past 3 season in the EPL the chance of a 0-0 score was around 8.6% - i.e. LESS than one nil-nil score every round in the EPL. So, if you're a betting man, stay away from prognosticating about scoreless results in football. :-) Source: http://www.soccerstats.com/scoring.asp?league=england_2011
sheek
Guest
Errrrr ..... May I boldly suggest we use final scores as an indication of who played the better football. Obviously difficult in football with so many 0-0 scorelines! Sport, as in life, is about opportunity & the choice you make, or don't make, when given that opportunity. Most of us get many opportunities which we fritter away for all sorts of reasons. Others need only one great opportunity in order to cash-in, because they're focused. And what Darwin Stubbie said.....
gawa
Guest
I recall a time when football was starting to be regularly televised live in the late 80's on UK tv and there was no mention or analysis of statistics. It is a simple device of the tv companies to help us think that we need these stats to understand what's happening in any given match. As stated already these statistics do not even start to grasp the tactics of some teams, you just have to watch a team that plays an effective counter attacking which is typical in European club competitions or International matches to see that these teams will happily concede possession for much of the game waiting for the right time to launch an attack. There aint much point in having 70% of the ball if it involves countless balls across the park back and forth while trying to unlock a well drilled defence. Player cam - now thats a really great invention of modern tv that is a great way for aspiring players to understand the game.
darwin stubby
Guest
Top tip - read Jonathan Wilson's truly excellent "Inverting The Pyramid" ... you'll get a far better understanding of the game unfolding in front of you than any A4 computer print out with columns of meaningless numbers ... the game is fluid - stats don't, and can't, tell the story
Roarchild
Guest
I disagree that the possession stat was meaningless. Before Oars goal judging by twitter I knew the young Socceroos were going to get a pasting by the public and felt it was unfairly. They certainly looked badly coached compared to Equador but we still played some football and the game was open as a result. A few players don't have a good enough first touch but most were ok. We were let down by our passing a lot but the game was at altitude wasn't it? People are talking up Equador and they certainly look more likely to do well because there well coached. This is youth football and setting up a compact defense and playing well on the counter with fast forwards not rocket science. That's why the possession stats were even. They didn't dominate the play of the game nearly as much as most fans seem to be making out. Australia had the ball in Equadors half plenty of times but were let down by their options/passing and Equador didn't seem overly creative to me. Yes Equador were better but wasn't as bad as people are making out and in this case stats don't lie. That said surely shots on target is a better measure of domination. -- Comment left via The Roar's iPhone app. Download The Roar's iPhone App in the App Store here.
nordozzz
Roar Guru
what about these tracking devices for generating stats? You can see it with metres covered stats they have trialled. They'd use GPS or something similar i imagine. An extension though could take it to the point of having a full game radar-type interface, similar to what u have on console football games. Made available to or even generated by broadcasters. So as a way of logging every action as it happens and somehow crunching those numbers live as a game progresses. Do they do this already as a post-mortem for some of this game analysis software? And to take it in a slightly different direction ... have thought a radar setup has potential for tracking offsides, perhaps accessed by a 4th official using a tablet device on the sidelines. All a bit too 'technology' for our rule governance overlords ... though given how widespread these devices are who knows how cheap it could be made, with IT partner support of course. An 'Offside App' wouldn't be a bad thing!
Football Fan
Guest
Football needs a Billy Beane style 'Moneyball' prophet who sees beyond the current situation and reads what are the truly important actions to be effective in a game, then tweaks teams and styles to suit. Fuss - I like your 2 but I humbly suggest we add % of crosses delivered to a teamate... Anyone else?
Midfielder
Guest
Davidde Agree but it helps lazy football writers seem like they know what they are talking about...
Fussball ist unser leben
Roar Guru
Nice article, Davidde. I find statistics to be totally meaningless in analysing a Football match - heck, very often, even the final score, is not a true reflection of what transpired during the 90' ... although, I'm not suggesting we stop keeping score! I reckon there are only two statistics I wouldn't mind seeing recorded, to give me an idea of how creative the teams were: a) number of passes received by an attacking player, inside the opponent's 18 yard box, from a team mate outside that 18 yard box. b) number of times an attacking player runs with the ball from outside the opponent's 18 yard box into the 18 yard box. Whilst these 2 statistics don't tell everything about the game, for me, they are the most important statistics when assessing creativity and attacking prospects.
nordozzz
Roar Guru
one different way to look at possession would be to break it down into phases in the game ... centred around scoring for example ... so % poss stats at 0-0, % after the first goal, equaliser, second etc. Also those 'last five mins' possession stats can be interesting.
Futbanous
Guest
Personally I find match statistics boring. Why bother with stats when Messi goes on a mazy dribble or indeed Tommy Oar scores a goal like the other day. For me what you see is what you get.
Birkish Delight
Guest
I think the "passing", "time in opposition half" and "balls into the opposition penalty box" would be better indicators, however iIt's always going to be hard to get a statistic to measure a qualitative aspect of any game.