Stats history says United for the title, with Chelsea, Arsenal, Liverpool next
By Brian Sears
29 March 2011
As soon as Fabio Capello's 'B' team have faced Ghana at Wembley, attention in English football will switch back to the enthralling Premier League races for the title, for European qualification, and to avoid the drop.
Just 82 Premier League games remain to be played this season, and it's great fun trying to predict the outcome. But on what basis can we try to do this with some credible, evidence-based data?
On a hunch? Of course not.
Tossing a coin? Nope. Favouritism? Niet.
So here is one statistical stab at it, based on previous Premier League results for the 82 games still to come.
Of course an exercise like this needs a proviso, and this is no exception. All manner of things from great management to poor management, injuries, form, luck and oodles of other factors will mean that the actual results don't pan out as we're suggesting.
Nonetheless, using statistical history as our guide, we have predicted the 82 remaining Premier League games using the following rules.
.
1: Where the away team has already gained more points in previous Premier League games than the home team, an away win is awarded. This weekend, for example, Liverpool play at West Brom, where they have already won all their previous four games there in the Premier League. We award Liverpool an away win in our calculations. Less clear cut, but still an away win in our reckoning, will be the outcome for Tottenham this weekend at Wigan, where in five Premier League meeting, Spurs lead 8-5 on points, or have 62 per cent of the points from that game. Of the 82 games remaining, on this basis, we reckon that 20 will be away wins overall, which is also in line with the League's long-term share of away wins.
.
2: Where home teams and away teams have historically shared points, or the away team has gained at least a third of the points on offer, but not enough for a win, then we've predicted the game will be drawn. This 30 per cent of points for an away team to get a draw is clearly an arbitrary figure, but in statistical modeling it does have the virtue of contributing to this process 22 drawn games out of the last 82. This is in line with the long-term share of results. Thus this Saturday, when Aston Villa travel to Goodison Park, we'll have a fixture where both clubs have so far gained 24 points from the 18 Premier League games they've played at the venue to date. So a draw it is. And of the six games Bolton have played at St Andrews to date, they have gained six points (a third of 18) so they just scrape in with a draw. Closer examination of that fixture shows four Birmingham wins and two Bolton wins from six Premier League games and no draws. Bolton have therefore proved they can win there, but so have Birmingham.
.
3: The other 40 games are declared home wins on the basis that the away team has not done enough to warrant a win or a draw, historically.
.
4: So what about Blackpool, with no Premier League history for their upcoming games? Here we take into account the games already played in the reverse fixtures this season. Thus as Blackpool only drew with Fulham in August at home, the result this coming Sunday is given as a win for Fulham at Fulham ... and so on. Yes, Blackpool fans, I know that life's not fair but as statisticians we work with the material we have. And in this reckoning, Blackpool's alarming recent dip in form is ignored to an extent, and actually the predictions see them finish 10th. That could be optimistic but it's what the stats tell us.
.
5: And what about West Brom, who have not yet played a Premier League game at Wolves? Going by the Blackpool method, as West Brom only managed a home draw when Wolves visited in February, a win is awarded to Wolves when they have home advantage. I know, Baggies fans, that's cruel. Indeed our final table, published in full elsewhere today, is very cruel for you, but it will surely be all the sweeter if you defy history and win more points that projected.
.
When all the above are considered, Manchester United run out as title winners, after gaining 20 more points from their remaining eight games for a total of 83 points. Those 20 points come from six wins and two draws from matches against West Ham, Fulham, Newcastle, Everton, Arsenal, Chelsea, Blackburn and Blackpool.
Chelsea are projected to find a storming finish and win 25 points from nine remaining games (eight wins and a draw) to finish in second, with 79 points.
Arsenal are projected to win 19 points from nine games remaining, to finish third on 77 points. Clearly a few results here and there could change the real order. We know that. But this is what one set of predictive rules suggest, no more.
Perhaps surprisingly (or maybe not), Liverpool are projected to finish with a bang, and take 20 points from their eight remaining games against West Brom, Manchester City, Arsenal, Birmingham, Newcastle, Fulham, Tottenham and Villa.
If all that transpires, then the 'big four' will finish as the big four again, with Tottenham in fifth and Manchester City in sixth.
The full projected table is elsewhere today.
Anyone wanting to predict the top four and bottom three can do so via Twitter, to @sportingintel, and use the hashtag #PLtop4bottom3. As and when they come in, we'll post them as comments on this article.
.
More on United / Chelsea / Arsenal / Liverpool / Tottenham / City