How will form influence the Premier League game between Chelsea and Southampton?

What does form mean? We all know what it is: recent results. They have already happened. What do they tell us about what might happen next?

The answer, I think, is: something, but not as much as we sometimes imagine.

Pundits will say a team are in form. They should say the team have been in form. Nobody knows if the team are still in form. We will only find out when they play their next game. And if they win the right terminology again will be that they have been in form.

There is a lot of statistical analysis out there. Some of it helps to explain what has happened. Hardly any of it considers what I feel should be a natural follow-on question: is this stat any use as a guide to what will happen next?

This is the missing part of nearly all statistical analysis. It is the most important part for betting. And it should be interesting even to football fans who do not bet. They are always talking about what might happen in the next game and the rest of the season.

Say a team won their last six games and another team lost their last six games. Now they are going to play each other. What is the chance that the first team will beat the second? It is not 100 per cent. There have been such games and the team with perfect six-match form did not always beat the team with the worst possible six-match form.

Premier League teams this season have played six games. There are 14 points between top and bottom of the table. The biggest gap between teams who will play each other this weekend is nine points: Chelsea have nine points more than Southampton.

Six-match form is a length that many bettors have consulted for many years. What should we infer from it about what might happen next?

I explored the relationship between results over the last six games and result in the next game in 26 Premier League seasons, 1995-96 to 2020-21. I started in each season with the seventh game, noting results from the first six games and result in the seventh game, then moved on. The simplest way to compare teams, it seemed to me, was through the difference in points over the previous six games.

I plotted what I found on the graph. The numbers along the bottom are differences in points between teams over the previous six games. The numbers up the side are win percentages in the next game when they played each other. The green bar is for teams who had gained most points in the previous six games, the red bar is for teams who had gained fewest points in the previous six games.

Of course there were some games between teams who had gained the same number of points. Each team won 37 per cent of those games with the other 26 per cent being drawn. When there was a difference in points from the previous six games the team with the higher total won more often than the team with the lower total. And as the difference in six-match points totals got bigger so did the difference in win percentages in that next game when they played each other.

Did it grow as fast as you had expected? Let us go through some examples.

When one team had gained three points more than another over the previous six games they beat them 43 per cent of the time, drew 27 per cent and lost 30 per cent. They won more often than they lost, but almost six times out of ten they did not win.

When one team had gained six points more than another from the previous half-dozen games they beat them 49 per cent of the time, drew 24 per cent and lost 27 per cent. Just over half the time they did not win.

What about teams who had gained nine points more than another over the previous six games? They beat them 57 per cent of the time, drew 23 per cent and lost 20 per cent. More than four times out of ten they did not win.

Or teams who had gained 12 points more than another across the previous six games? They beat them 70 per cent of the time, drew 13 per cent and lost 17 per cent. Three times out of ten they did not win.

Look back at the graph. You will see that the bars become erratic when the difference in points over the previous six games was greater than 12. This is because there were not many of those games. Small samples can give erratic readings. In only one game out of 100 was the difference in six-match form between teams greater than 12 points. There are ten games in most rounds of Premier League fixtures so it happened about once every ten weeks.

After six games this season Chelsea have nine points more than Southampton. They play on Saturday. You will remember that in earlier seasons teams who had gained nine points more than their opponents over the previous six games beat them 57 per cent of the time. They won more often at home and less often away – typically ten per cent more often at home and ten per cent less often away. Chelsea will be at home. So make that 67 per cent.

Many six-match points totals can be reached in different ways, and the results can come in different orders. Arsenal and Tottenham this season have nine points. Tottenham won their first three games then lost the next three. Arsenal lost their first three games then won the next three. Some would say that for this reason alone Arsenal are now better than Tottenham. I do not necessarily agree, but that is a discussion for another day.

Chelsea are 2-5 to beat Southampton. Those odds imply Chelsea have a 71 per cent chance of winning. That is a bit more than 67 per cent. Of course there could be other reasons for thinking Chelsea are that much better than Southampton, and this is sort of my point. Chelsea might have a 71 per cent chance of beating Southampton, but not because of their last six results. Form means something but it does not mean everything.

Not got a bet365 account? Sign up today and get up to £100 in bet credits

Up to £100 in Bet Credits for new customers at bet365. Min deposit £5. Bet Credits available for use upon settlement of bets to value of qualifying deposit. Min odds, bet and payment method exclusions apply. Returns exclude Bet Credits stake. Time limits and T&Cs apply.

CLAIM OFFER HERE

MORE FREE BETS

Today's top sports betting stories

Follow us on Twitter @racingpostsport