Premier League wins in round one may have been better than performances deserved

Don’t get carried away. After the first weekend of the Premier League season some people might have to work hard to restrain themselves. Supporters of Burnley or Southampton, for instance, and supporters of Brighton or Watford.

Burnley beat Southampton 3-0. Brighton beat Watford 3-0. Some hopes for the season will have risen, others will have drooped. Those reactions might turn out to be right, but we should not be confident that they will.

This season, I suspect, Watford will fare less well and Brighton better. I feel that last season Watford slightly over-achieved. And I have come to realise that when a club make a left-field managerial change, as Brighton did between seasons when they replaced Chris Hughton with Graham Potter, the natural impulse is to scream “what a ridiculous mistake” but the natural impulse might be wrong.

I am one of the football writers my colleague Bruce Millington referred to a few weeks ago, those who thought Southampton had gone mad when they sacked Nigel Adkins to bring in Mauricio Pochettino. Bruce may not remember that I was one of them but I do.

Still, do not take any score at face value. Not until you have scratched at the surface anyway. When Tony Pulis managed West Brom in the Premier League they won 4-2 at home to West Ham a week after they had lost 1-0 at Bournemouth. Pulis said: “We played well at Bournemouth and most probably created better opportunities and better chances than we have today.”

Odds for Burnley v Southampton and Watford v Brighton suggested that in each match the most likely score was 1-1. And 1-1 might have been the fairest outcome according to the chances created.

Expected goals stats estimate how many goals would be scored on average with the same number of shots from the same positions. FiveThirtyEight’s expected goal stats, similar to others I have seen, were 0.7 for Watford and 0.7 for Brighton, 1.0 for Burnley and 1.2 for Southampton.

Ordinary Premier League teams score one goal for roughly every ten shots. Burnley had ten shots, Southampton 11, Watford 11. Brighton had only five shots so theirs must have been from better positions.

In these games one team scored three goals more than the other but analysis of chances created gives the impression that a fairer outcome would have been for both teams to have scored the same number of goals.

Generally speaking, though, the best indication of what a team deserved from chances created is given by dividing a winning or losing margin by two.

I studied Premier League games from the last two seasons and compared the actual goal difference between teams with the expected goal difference – one team’s expected goals minus the other team’s expected goals.

And this is what I found. When a team won by six goals there was on average an expected winning margin of three goals. When a team won by four goals there was on average an expected winning margin of two goals. When a team won by two goals there was on average an expected winning margin of one goal.

As I say, those expectations are averages. Some winners deserved more, others less. Treat what I have just told you as a rule of thumb, a starting point from which you can dig deeper if you begin to wonder what you should make of a score.

Goals do not tell us everything about what happened in a match. Nor do shots.

FiveThirtyEight’s expected goals stats are the best I have seen. There might be better ones but I have not seen them. I wrote earlier that expected goals stats are derived from shots. Nearly always they are. FiveThirtyEight, though, give two sets of expected goals stats for each team in a match – the sort that I quoted before, derived from shots, and another sort derived from everything else.

An attempted pass between defenders into the penalty area might be cut out, in which case there will be no shot. But sometimes that sort of pass will get through, and when it does sometimes the recipient will get off a shot. At the moment the ball leaves the passer’s foot the team’s chance of scoring rises.

When FiveThirtyEight rate performances they give as much weight to the passes and dribbles as they do to the shots. When I rate performances I also take into account things that did not lead to a shot let alone a goal.

Looking at everything that happened in a match, rather than simply the goals, gives a broader and better perspective. The score on its own can over-excite the senses. When Pulis managed Middlesbrough in the Championship he said: “People get too carried away when you win and too carried away when you lose.”

Today's top sports betting stories

Follow us on Twitter @racingpostsport

Like us on Facebook RacingPostSport