Over the last couple of days, I've been doing some work on my databases for the last few years' worth of league play. In the process, I've come across some interesting patterns that I thought were worth sharing. One of them was simply the distribution of full time results in the Premier League last year.
For the league as a whole, the 2010/11 season saw more draws in the Premiership than elsewhere. 29.2% of all matches EPL teams played last year ended in a draw; that's several percentage points higher than in the other three leagues I have been writing about (Bundesliga, La Liga, and Serie A). Here's a graph of match outcomes at the level of teams (with wins and losses equal by definition).
To stats aficionados who know more than I do this may be old hat, but I thought this was a curious pattern - curious in part because we have seen in other analyses that the Premiership has become more imbalanced over the last few years. Of course, these do not have to be inconsistent - you can still have a few dominant teams that win all of their matches and everyone else producing mostly draws. At the same time, a more imbalanced league producing more draws is interesting (well, to me).
So which teams were particularly likely to produce draws last year? That is, who was most responsible for the high % of draws in the Premier League last season?
Turns out, it was Fulham at 42.1%, followed by Everton and Birmingham producing draws in 39.5% of their matches. In plain English, the Cottagers played their opponents to a draw in over 4 of 10 matches they played. This also means that Fulham produced 44% more draws than the average Premier League team last year. In contrast, Liverpool, Stoke, and Wolves left the pitch with a draw less than half as often as these clubs - only 7 or 18.4% of the matches they played ended in a draw.
Below are the full time results for all 20 teams for last year.
Few teams had results profiles like Fulham's, with draws dominating over wins or losses as the most common category of match outcomes. Everton and Wigan were the only other two clubs where draws outnumber wins or losses.
So how did the Cottagers do it? Comparing halftime and full time results, it turns out that Fulham were able to turn a good number of halftime losses into fulltime draws. They were behind at the half in 16 (or 42%) of their matches, but ended up losing only 8 of those (the only team that was behind in more matches was West Brom with 17). And when Fulham were tied at the half, which was the case in 12 (or 31.6%) other matches, they ended up losing only 3 of those. In contrast, the typical Premiership club was behind only 30% of matches at the half, but ended up losing 75% of those matches. In essence, Fulham were able to turn halftime deficits into draws (and even one win).
Interestingly, Fulham's was also a tale of two seasons, as most of the draws came in the first half of the campaign, coupled with most of the losses (and only three wins), as the next graph (courtesy of the FFC website) shows. In contrast, the second half of the season showed a steady upward trend in results.
So this year's award for coming from behind and salvaging a draw goes to Fulham. The feat of earning draws against the odds was accomplished by a very experienced squad. It is perhaps even more remarkable in light of the fact that half time scores are a better predictor of full time results in the Premier League than in the other three big leagues, as the statistical relationships between halftime and full time scores for the 2010/11 season reveal. When we regress full time results on halftime results, we find that halftime results account for more of the full time results in the Premiership than in the other leagues (as measured by the r-squared).
Bundesliga: 26%
EPL: 39%
Serie A: 31%
La Liga: 34%
I hope for the Whites that they will be able to avoid falling behind quite as much this coming season. And I hope for any supporter of teams in the EPL that their clubs won't be called upon to do what Fulham did - it's a hard thing to do, it turns out, harder than it may appear at first glance.