Guest post: A prediction update from Steve Lawrence

The Scoreboard Journalism challenge for points and place predictions from prominent media, stats modellers, fans and online publishers for both the Eredivisie and the Premier League has attracted considerable interest and the original posts can be found at:
https://scoreboardjournalism.wordpress.com/2014/08/22/48-forecasts-the-eredivisie-predicted/

&

https://scoreboardjournalism.wordpress.com/2014/09/11/predicting-the-premier-league/.

Throughout the season, fellow blogger and analyst Steve Lawrence has been putting the predictions under the statistical microscope and predicting what might happen. You can find his last excellent post here: https://scoreboardjournalism.wordpress.com/2014/12/05/guest-post-steve-lawrence-assesses-the-premier-league-predictions/. Steve now looks at how the busy Christmas and New Year period have affected things.

The results of matches in the congested Christmas period, from Boxing Day to New Year’s Day, had a normalising effect on the points distribution in the English Premier league 2014/15 with no team taking a maximum nine points and only one team taking no points (mid-table Everton).

ChristmasresultsPrior to the Christmas period it looked as if the points range (top to bottom) in the Premier League was heading towards the high end of expectations with the potential for a final table range of over 70. Chelsea’s poor form over the four Christmas matches though combined with all the bottom clubs picking up points has seen the current points range regress to 32 whilst the standard deviation stabilised at just under 10. These will of course increase as the graph below shows but it is likely that the final range and standard deviation will fall within the lattice pattern shown on the graphic. Very few of the predictors had a range and standard deviation within their points which puts them in the same place.

SteveXmas1

The half-way point in the competition (Round 19) also occurred during this year’s Christmas/New Year period with clubs having played each other once. There is usually a slight correction evident around this time, perhaps due to the congestion of matches. It now looks as if the range will be around 64 at season’s end with a standard deviation of just under 19.

In the last blog I said that  ‘it looks unlikely that the coefficient of variation will dip to 0.3 unless QPR, Hull, Burnley and Leicester City all experience some kind of Christmas epiphany and Chelsea sack Mourinho’. In fact QPR, Hull City, Burnley and Leicester City picked up 14 points between them whilst Chelsea managed only 4 and that has had the effect of stabilising the coefficient of variation at 0.35. It’s hard to see it change much from this at the end of the season and that leaves many of the Scoreboard Journalism forecasts out in the cold. This can be seen by those below 0.35 and particularly below the line of 0.3 on the following graphic. For details of which forecasters these are, refer to my previous post (https://scoreboardjournalism.wordpress.com/2014/12/05/guest-post-steve-lawrence-assesses-the-premier-league-predictions) as a table appears there.

SteveXmas2

At this point of the season thoughts begin to turn to survival at the bottom of the league and the last 10 seasons show a relationship between the standard deviation and survival points (albeit not a strong one) where a higher standard deviation means that fewer points will be needed to survive. A standard deviation of just under 19 tends to suggest a minimum points accumulation for survival in the league of around 36 or 37, so perhaps a little less than the commonly quoted 40 points. If Leicester City, currently bottom with 14 points, continue their Christmas form for the rest of the season they’ll survive.

SteveXmas3

 

SteveXmas4

There is an argument that under pressure squads at the top of the Premiership have a tough time over Christmas with an intense calendar immediately after Champions League and Europa Cup fixtures whilst lower ranking clubs have a slightly easier time of it. The results this Christmas bear that out with the result that the Premier League just became a lot more interesting with no foregone conclusions at the top or the bottom.

Steve Lawrence researches age effects and their relationship with performance outcomes in sport. He uses the twitter handle @SteveLawrence_ and is a very occasional bloggerhttp://theanalyticslab.tumblr.com

Steve is the developer of the miTeamsheet app http://www.miteamsheet.com which allows both a rapid calculation of the ‘average team age’ (ATA) of any competing sports team along with an index of ‘relative age bias’ (RAEi) within the team.

Steve maintains an interest in International Sports Law, EU State Aid Law in respect of Sports Facilities, Corruption in Sport and Sports Betting.

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

4 Responses to Guest post: A prediction update from Steve Lawrence

  1. Alphax says:

    Steve, I do a very similar dataset to your Seasonal Comparative Results, but we have a different result in Tottenham Hotspur’s results in your graph published on 1/11: http://public.tableausoftware.com/profile/scoreboardjournalism#!/vizhome/SCOREPremierLeague20142015/Dashboard1

    You have them on 69 points, and I have them on 72 points.
    Up until this most recent graph, our results have been identical.

    The only 3 point swing I can see since your previous chart for Matchday 17 is Spurs 1-2 victory over Leicester City on Matchday 18 which was a reversal of their 1-0 loss to Norwich City last season.

    They’re newly promoted/ promoted sides, so I suspect that may have something to do with it.

    In my methodology, for the newly promoted teams, their equivalents for the previous season are the 3 relegated teams. Championship 1 for Relegated 17, Championship 2 for Relegated 18, Championship 3 (playoff winner) for Relegated 19, so

    Leicester City for Norwich City
    Burnley for Fulham
    QPR for Cardiff City

  2. Pingback: Prediction Analysis by Steve Lawrence | Scoreboard Journalism

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