Expected-goals stats provide a crucial sense of proportion
Wise words from the Soccer Boffin
When Harry Redknapp managed Southampton he lost patience with a data analyst. “Tell you what,” he said, “next week why don’t we get your computer to play their computer and see who wins?”
Not everyone will think there is a use for a statistic such as expected goals, which has reached a mainstream audience now on Match of the Day. But I do.
I was surprised but encouraged to find expected goals on Match of the Day. Their most obvious use is to provide a sense of balance. It would be harder to savage Arsenal for losing 1-0 at Stoke if you had just been shown stats saying they should have won by a goal.
Guests on Match of the Day that night were Jermaine Jenas and Danny Murphy. Even without stats to prompt them they have always drawn on their experience as players to tell us things that are interesting and fair. But not every guest has.
One newspaper said viewers were confused by expected goals. I doubt they were. The question answered by expected goals is quite simple. With those shots from those positions how many goals would a team score on average?
In the Premier League about one shot out of every three requires a save, and about one out of every three shots that require a save does not get it. Which suggests that about one shot out of every nine reaches the back of the net (3 x 3 = 9). I rounded the first two ratios, however; the true goals to shots ratio is nearer one in ten.
Some shots are from more promising positions than others. Inside the penalty area about one shot out of every six is scored. Here, too, there will be a better chance from some places than others. Expected goals is a stat that summarises these variables.
Teams who score a high number of goals tend to have a lower expected goals figure, and teams who score a low number of goals tend to have a higher expected goals figure.
That is what I mean about the stat introducing a sense of proportion. It illustrates in numbers my favourite Graham Taylor words – that when you win you are seldom as good as people say you were and when you lose you are seldom as bad.
Over the first three weekends of the Premier League season the total number of goals scored corresponded closely with the total of expected goals.
We can adapt Eric Morecambe’s famous riposte to the conductor Andre Previn that he was playing all the right notes but not necessarily in the right order. The Premier League produces the right number of goals, but not always in the right games.
Expected goals are derived from actual shots, so they tell a similar story, but they seem to get it across even better. Among other things they remind us of how many shots do not become goals.
Commentators tell us we would have put our house on a player scoring. Whether we bet should depend on the odds, and we should never risk everything because even a value-for-money bet can lose. Often, though, the chance of the player scoring was much smaller than the commentator’s screams convey. Reading expected goal stats after matches educates us about such things.
Little variation in top-flight figures across Europe
There was a time when football differed noticeably from country to country. Now the best players and coaches from around the world come together where they can earn most money. And in those places there is little difference in how and from where goals are scored.
The Premier League, La Liga, Bundesliga, Serie A, Ligue 1 – in all of those competitions one goal is scored for every six attempts from inside the box.
Outside the box one goal is scored for every 28 attempts, or some number close to 28. One goal is scored from a free kick for every 16 attempts from a free kick, or some number that differs only a little from 16.
More to catenaccio than good defending
Ruud Gullit played at Milan for Arrigo Sacchi, a coach who was the antidote to defensive Italian football. So he can be excused for describing Chelsea’s tactics in their Premier League win over Tottenham as catenaccio. They were not.
Catenaccio means chain, as in chain on a door. An earlier Swiss version of the system was called verrou, which means bolt, as in bolt on a door.
An even better image was conjured by Gipo Viani, the first coach to employ catenaccio regularly in Italy. He coached Salernitana after the Second World War. Viani described fishermen hauling their catch out of the sea using two nets. Some fish slipped through the first net but they were caught by a second safety net.
Catenaccio was a spare man – the bolt or the chain or the safety net – behind man-markers.
It was applied most successfully by Helenio Herrera’s Inter, who in the 1960s won Serie A three times and the European Cup twice. With them it also became associated with bad sportsmanship and worse.
Chelsea’s play against Tottenham was not catenaccio in the tactical sense, and Gullit never implied that it was in the other sense.
Chelsea manager Antonio Conte, who is Italian, cleverly arranged his limited options into a shape that maximised their chances of beating Tottenham. And they did win.
They spent a lot of time defending, but mainly because Tottenham are a strong team who played well. Good defending? Yes. Catenaccio? No.
More articles by Kevin Pullein
Why the lion's share of possession is no guarantee of victory