Tottenham have had games of two halves – will they have a season of two halves?

I keep hearing that Tottenham are a team who get better as a game goes on. In five games in all competitions – three in the Premier League, two in the Champions League – they did better in the second half than the first.

In the three Premier League games they conceded twice before scoring themselves – they came from behind to beat Bournemouth 3-2, although before that they lost 2-1 to Newcastle and after that they lost 2-1 to Liverpool.

Tottenham started badly and ended well.

Did they do that on purpose? Why would they? Was it something unconscious in the character of the players? If so, why was it not there earlier this season? The players were the same. In their previous 15 games in all competitions Tottenham did better in the first half more often than they did better in the second half.

Or was it something that just happened in a few games, something that started for no discernible reason and will end for no discernible reason?

Perhaps it has ended already. On Wednesday, against Nottingham Forest in the League Cup, Tottenham did worse in the second half than the first.

Are there any teams who persistently do well in one part of matches and badly in another? I decided to find out.

I noted the first and second-half score in every Premier League game in the last 25 completed seasons, 1997-98 to 2021-22. Then I split each season into two – the first 19 rounds of games and the second 19 rounds of games. Were differences between first and second-half scores in one part of the season repeated in the other? The answer was an emphatic no.

Typically, over a period of time, a team got the same scores in both halves of matches. If a team did better during the first 45 minutes in the opening 19 games of a season, did they do so again in the closing 19 games? No. In the closing 19 games, on average, there was no difference between first and second-half scores.

If a team did better during the second 45 minutes in the opening 19 games of a season, did they do so again in the closing 19 games? Also no. In the closing 19 games, on average, they did equally well or badly in both halves.

There was no correlation between differences in first and second-half scores in one part of the season and the other. No matter how great the discrepancy in the first half of the season, in the second half, on average, it was zero.

After 19 games some teams would have been spoken of as first-half specialists and others would have been spoken of as second-half specialists. They had recorded much better results in one period of games than in the other. They did not carry on doing so. Whatever it was that had caused the disparity between first and second-half scores in the opening 19 games, it was not anything persistent in their tactics or peculiar to their character.

I am sometimes asked if it is worth knowing that a stat tells you nothing about what will happen next. My answer is yes. Especially if a lot of people are talking about the stat as if it does tell us what will happen next.

There are two types of stat: interesting and dull. Interesting is preferable, surely? It depends whether you want entertainment or enlightenment. Enlightenment is what you need for betting.

A rule of thumb: the more sensational a stat sounds the more likely it is to be telling you about something that was only a freak occurrence – and the less likely it is to be telling you about something that will continue.

I have five grandchildren. They are all grandsons. I mention it only because others find it strange. To me, every grandchild is a person. Out of every 32 sets of parents who have five grandchildren, one will have five grandsons. Another will have five granddaughters. Next month I will have a sixth grandchild. It is as likely to be a granddaughter as a grandson.

There are so many people out there competing for attention, in the media and on social platforms. If they are looking for stats they are looking for stats that will make you say “wow”. “Wow” can turn to “ow” if you bet on something happening again.

Only twice in the opening 19 games of last season did Brentford lose the second half. In the closing 19 games they lost the second half eight times. They lost the first half nine times in the opening 19 games and seven times in the closing 19 games. What happened early last season in second halves was unusual for them, and it was not repeated.

Chelsea two seasons ago did much better in second halves than in first halves over the opening 19 games. Over the closing 19 games there was little difference between halves. The same was true of Leicester three seasons ago, Arsenal four seasons ago and Manchester City five seasons ago.

Five seasons ago, over the opening 19 games, City did much better in second halves than in first halves. Two seasons ago it was they other way round: over the opening 19 games they did much better in first halves than in second halves.

Tottenham two seasons ago also did much better in first halves than in second halves over the opening 19 games. This season for a while it was the other way round. There is no reason to think such patterns will be replicated.

CLICK HERE FOR YOUR CHANCE TO WIN A 65" TV

Sign up to emails from Racing Post Sport and get all the latest news and tips

Today's top sports betting stories

Follow us on Twitter @racingpostsport