Wednesday, May 28, 2008

The Promroiship 07/08: A Review

Manchester United completed another treble this season, winning the Premier League, the Champions League and the inaugural Promroiship. Alex Ferguson shows no signs of retiring; I think he might be after another Promroiship title.

The point of the Promroiship was, of course, to try and decide whether there was any detectable bias in the Match of the Day Running Order. So back in January I started the Promroiship (the PRObrum Motd (match of the day) Running Order Index with a -ship on the end) which ran with the following rules:

All teams appearing on Match of the Day will be awarded points based on the position of their game in the running order, the number of goals scored and the number of matches appearing on MOTD that weekend. Only Premiership games will count and MOTD2 is not included. Midweek MOTD will also count towards the total. The rules are as follows:

1) Teams appearing in the first match will be awarded 10 points, the next 9 and so on all the way down to 1 if necessary. As there are 20 teams in the premiership, a maximum of 10 matches can be played on any day involving all teams once.

2) Because matches with more goals should in theory come first, 1 point will be taken away for each goal scored in the match. I deliberated a lot about this rule, because a few high scoring matches before Christmas made this potentially unfair, but I think it is the easiest way to factor in "entertainment" at an objective level. In any case, it would not be football if their was not potential for an argument over the fairness of some rule or other.

3) Thanks to security considerations, Sky and European involvement, sometimes there can be very few matches on Saturday and so we need to factor in the number of matches shown. I have decided to multiply the result of the first two rules by the number of matches appearing on MOTD. Obviously the fact that some teams do not play on Saturday adds in some additional unfairness, but I hope that over half a season there will not be too much bias, and we should pick it up week by week if there is.
We ended up with the league table here which was very similar to the Premier League final table, in that Manchester United won and Derby were bottom.

As the season progressed, however, those who were following the competition would have noticed that Blackburn and Aston Villa seemed to play fewer matches on Saturday, which put them at a disadvantage, as foreseen by the third point above. So, for the purposes of analysis, I produced the following League Table based on average Promroiship points per appearance on MOTD - in other words, I divided the total number of Promroiship points by the number of appearances on Match of the Day each team had over the same period. This gave the following revised league table:

The Average Promroiship Points per Appearance 078

Average...Apps..Prom Points
Manchester United ...48.08......13......625
Chelsea .............47.50......10......475
Bolton Wanderers ....41.42......12......497
Everton .............37.38.......8......299
Fulham ..............36.81......16......589
Arsenal .............36.08......13......469
Reading .............34.94......17......594
Portsmouth ..........33.17......12......398
Liverpool ...........32.70......10......327
Birmingham City......32.27......15......484
Wigan Athletic ......30.69......13......399
Tottenham Hotspur....30.29......14......424
Newcastle United.....29.75......12......357
Sunderland ..........28.00......14......392
West Ham United......26.20......15......393
Manchester City .....25.91......11......285
Aston Villa .........25.08......12......301
Blackburn Rovers.....23.58......12......283
Middlesbrough .......21.43......14......300
Derby County.........18.50......14......259

This has a much more familiar feel than the old Promroiship table, with Manchester United and Chelsea convincingly taking the top two spots, disproving my assertion in the last article that:

The Beeb may be biased towards Manchester United. The other Premier League contenders finished lower than those involved in the relegation battle.

Those in the relegation battle finished lower using averages than total Promoiship points, indicating that they played more games on Saturday (we could say Sky is biased against covering relegation battles, but I believe the Sunday games are decided well in advance). Everton are the big movers, finishing in the top four rather than the bottom four. City rivals Liverpool also move up a bit, having only played in ten out of nineteen possible MOTD apearances (well maybe less, since for example Villa and Blackburn brought one of their games forward creating a "phantom" round).

The second part of the comment, namely that the Beeb were biased against Derby, still seems valid however.

I realise that most people prefer visual representations rather than raw figures. So I started producing some graphs. The most obvious assertion is that the Match of the Day Running Order is decided by promotion and relegation battles. Let's plot the average Promroiship points against the teams in order of Premier League finishing position, starting with the Champions and finishing with the wooden spoon (click on the graph to enlarge):

There are obvious spikes around the Premier League title rivals and the relegation battlers. Midddlesborough can feel slightly aggrieved, although winning 8-1 in a nothing game on the last day of the season did not help their Promroiship chances. Aston Villa and Blackburn seem to prove that the Beeb were not interested in the race for the UEFA cup, although Everton and Portsmouth strangely seemed not to be affected by this. Were the BBC overcompensating for Phil Neville's comment about Everton always being last on Match of the Day, a statement which helped kick this competition off?

I added in the trend line for effect, but it is probably skewed by Manchester United, Chelsea and Derby' Promroiship performances. The overall average Promroiship points per match is 31.99, so as a rule of thumb teams over the 30 point line in the graphs on this page got more than their fair share of coverage, based on goals scored in the games, and those underneath less.

To recap then, there seems to be some evidence for the theory that the Beeb basis its MOTD running order on title and relegation battles. Are they biased towards big clubs? The graph above shows that the Big Four do better than average (although Arsenal and Liverpool probably do less well than expected). What about other big clubs/ sleeping giants? Of course, deciding which clubs are big is a minefield, so I decided to use the size of the stadium as a proxy for size of the club and see if there were any trends.

The trend line seems to indicate there is very little correlation between size of a club's stadium and their average Promoiship points. Either our definition of big club is faulty (quite probably), or the Beeb believe that size does not matter. The fact that the spikes centre around the likes of Manchester United, Chelsea and the relegation battlers is probably more telling.

Finally, this would not be a Birmingham blog if I did not complain about London bias. Is MOTD biased towards the self-styled capital of the world? In the next graph I order the teams (broadly) in distance from Wembley stadium:

The trend line seems to pick something up, but again final Premier League positions seem to be more important when looking at the spikes. The North-East does seem to be discriminated against, but none of their three teams were probably ever serious relegation (or title) candidates.

In conclusion then, Match of the Day in the second half of the season concentrates on league positions rather than high scoring matches. Teams involved in relegation battles and title challenges are more likely to be featured before teams languishing in mid-table obscurity. As such, it would make more sense to run the Promroiship from the beginning of the season to December 31st than January 1st to the end. Such a season would be more likely to expose any potential MOTD bias, which, from this experiment, is hard to prove conclusively exists.

Another conclusion is that Phil Neville talks b*******s. Everton do not always appear last on Match of the Day, and are probably shown higher up the order than they should be when the number of goals in their matches are taken into account.

Match of the Day concentrates on title/ relegation battles. Phil Neville talks crap. Probably did not need to create the Promroiship to prove these two self-evident truths.

See you again (maybe) next season.

No comments: