Hot streaks likely to push points higher than model forecasts?

Alexej Behnisch and Constantinos Chappas have both produced excellent articles over the last few days outlining Premier League predictions at approximately the halfway stage of the season. The six models contained in Alexej’s analysis have their Premier League champions this season finishing on between 75 and 78 points. Constantinos’s 15 modellers estimated the points total of their leading teams between 72 and 81 points. A look at the history of the second halves of Premier League seasons provides good reason to doubt that these team estimates are high enough.

Expecting a total of between 72 and 81 points to be enough to win the Premier League title from this stage fails to take account of something which happens almost every season in the division and which could push that total somewhat higher. In almost every one of the past 20 Premier League seasons, at least one of the top clubs has produced a hot streak of results which, if repeated this season, means that well over 80 points will be necessary to become champions.


In 18 of the 20 previous Premier League seasons in the 20 club era, at least one team has managed 42 points or more from their last 19 matches. In 13 of those seasons, at least one club has taken 44 or more from their last 57 possible points. The only seasons in which that did not happen were 1996/1997 and 2010/2011. Manchester United were the best in each with 41 and 39 points respectively, and that was enough to win the title with room to spare on both occasions.

Five current Premier League clubs – Arsenal, Chelsea, Liverpool, Manchester City and Manchester United – have taken 42 points from their last 19 fixtures over the last 20 years, with the since relegated Leeds United also managing it in 2000/2001. Fifteen of the 20 champions over that period have amassed at least 42 points from a possible 57 in the second half of the season.

Points predictions in these types of models are usually based on summing up the expected points in each match and therefore every team is given an estimated total based on performing exactly as expected given their current relative strengths. Over a relatively short space of time, there is a very good chance though that at least some of those at the top end of the quality span will be over-performing, perhaps by quite a lot.

Even in seasons similar to this one, with relatively low points at the halfway stage for the leader, one of the top teams has usually had a more than decent second half of the season as shown below.

Secondhalves

In each of the four seasons in which the halfway leaders had under 41 points, the eventual champions recorded 41 or more points in the second half of the season. Arsenal’s 51 points in the second half of the 2001/2002 season represented a 41% improvement on their 36 from the first 19, and improvements of at least a third in the second half of the season are the norm for the teams in the table producing these runs.

Assuming that one of this season’s strongest teams has a run of this magnitude, the likelihood is that at least 80 points will be necessary to win the title. If there is an improvement of anything like a third by Arsenal or Manchester City on their first 19 matches, the winners of this season’s Premier League could even finish a fair bit above that 80 point mark.  That is, unless the patterns of English football that we have observed over the past 20 years have really changed.

Advertisements
This entry was posted in Uncategorized and tagged , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s