Tottenham and Newcastle show that Premier League shooting is hit-and-miss

There is a cartoon on Tangotiger, a website about baseball and ice hockey that I found through an article by James Willoughby. The cartoon has two sports commentators. One says: “A weighted random number generator just produced a new batch of numbers.” The other replies: “Let’s use them to build narratives!”

Commentators, pundits, journalists, fans – we are all the same. We all come up with other explanations for things that really just happened by chance. The other explanations sound better but they are wrong.

Much in football depends on which shots go in. There is little rhyme or reason to which ones do.

Premier League teams average one goal every ten shots. Tottenham had 20 shots in their last two Premier League games and scored two goals, bang on the average. The shots that went in were the first two, both at Manchester City. None of the next 18 went in – one more at City then 17 at home to Newcastle.

That is why Tottenham drew 2-2 with City (though they would have lost 3-2 if VAR had got the new handball rule right) then lost 1-0 to Newcastle. The first two shots might not have gone in. The third shot could have gone in. Any of the last 17 shots could have gone in.

Newcastle faced 40 shots in their first three Premier League games and conceded four goals, bang on the average. None of the goals came in the last 17 shots, which were all those they faced at Tottenham. That is why Newcastle won 1-0 at Tottenham.

Okay, some other things also influence which shots go in. The skill of the kicker, the position of the ball, the skill, number and position of defenders. Compilers of expected-goals stats look into these things and they agreed that Tottenham should have scored two goals in their last two games while Newcastle should have conceded four goals in their first three games. Still nobody knows which shots will go in.

Goals in football are a randomly distributed function of attempts. Dull, I grant you. The truth can be.

Imagine a middle-of-the-range Premier League team who average one goal every ten shots. They have ten shots in each of three games, and after those three games they have scored three goals.

There is only a 25 per cent chance they will have scored one goal in each game. There is a 66 per cent chance they will have scored two goals in one game, one goal in another game and no goals in the other game. Think of the stories that would be told – how the team played well when they scored twice, ordinarily when they scored once and badly when they did not score.

There is a nine per cent chance the team will have scored all three goals in one game. The proper explanation for whatever permutation occurred is that it occurred haphazardly – in the same way that on each spin a coin can land on heads or tails and on each roll a dice can show any of one, two, three, four, five or six.

There is not a lot of time in football for goals to be scored

Something I have observed over the years is that less happens in football matches than most people seem to think. Looking through some stats from the past five Premier League seasons I came across one of the reasons why. The seasons were 2014-15 through to 2018-19.

About 40 per cent of the time that a match was in progress the ball was dead. So no goals could be scored. Another 27 per cent of the time the ball was in play but in the middle third of the pitch, from where it is rare for a goal to be scored.

Only for about 33 per cent of the time that a match was in progress was the ball in play in one of the attacking thirds. For only about a third of the duration of a match was there any practical hope of a goal being scored.

Another point I have argued down the years is that in most Premier League seasons there is little difference in the abilities of teams outside the top six. Action zone stats from the last five seasons have strengthened my conviction.

Knowing how long the ball was in different parts of the pitch gives you an idea of what a match was like. Below the top six there was little difference between positions in how long during matches the ball was in the attacking, middle and defending thirds.

This suggests to me that the matches of those teams were similar, which in turn suggests to me that the abilities of those teams were also similar, and that the sometimes large differences in results were attributable mostly to good or bad fortune.

This is more likely than the alternative, which is that the teams were similar in their ability to get the ball from the defending third through the middle third into the attacking third, but then strikingly different in their ability to get the ball from the attacking third into the opposition’s goal, or to stop the ball going from the defending third into their own goal.

Follow us on Twitter

Like us on Facebook RacingPostSport