With the second week of fixtures in the Group Stages of the Champions League about to begin, we thought it would be interesting to look at how the structure of the competition affects the percentage chance each team has of winning the competition. What if the Champions League really was an, admittedly large, league of 32 teams?
Is the current format, including draws which are seeded and which force teams from the same nation to avoid each other until the Quarter-Finals a disadvantage for the smaller teams? Is this disadvantage more than offset by the nature of a knock-out competition allowing lesser teams to progress off the back of a couple of lucky performances?
There’s clearly more of an emphasis in sport on winning than there is on entertaining, but we can turn our football modelling techniques towards a different target: goals per game. What you really want is for your team to be the best, but if you think about things from a neutral’s perspective the total number of goals per game might be more interesting. So who, as a neutral, should you go to see?
Without goals football probably wouldn’t be that interesting but what is the right number of goals? If we had too many goals then the best team would win more of the time and the game might become too predictable (technically, if all of the teams had about the same quality then it could still be unpredictable and have lots of goals). On the other hand, too few goals would make the game boring. If you look over many leagues, and many seasons you tend to find about 2.6 goals per game on average. Since football is popular perhaps we can make a leap of faith and say that 2.6 is a good number.
Goals of course only come if you first have a shot. In the Premier League, over the last ten seasons, about 22% of on-target shots have become goals. This number does fluctuate a bit from season-to-season as does the number of on-target shots. The odd thing is that their fluctuations largely cancel each other out: the number of goals remains relatively constant over time. 2.6 is just special.
Arsenal have scored only three goals this season. Should they be worried about their inability to convert shots into goals?
Arsenal fan @Orbinho – mentor to @OptaJoe – tweeted that Man Utd have scored with 30% of their shots last season while Arsenal have only managed to score with 7% of theirs. The chart shows that they rank equal 14th, alongside West Brom and Norwich, in terms of shot conversion this season.
Last season Arsenal ranked fourth with a shot-to-goal conversion rate of 14%. Man Utd topped the conversion charts with 16.4% of their shots finding the back of the net.
Another neat way of looking at the transfers is by division and by country – where do the new players come from and where do the old players go?
Across the football league structure, transfers are still quite parochial. Of 760 total deals, 144 were free agent signings and 356 were loans and transfers from one of the 92 English League clubs. A lot of that is due to the national rather than international nature of the lower divisions; but even of 105 deals with players signing for Premier League clubs, 39 were from another Premier League club, 20 were from the Championship and 13 were free agents. Only 33 signings came from overseas leagues.
Here’s another way to visualise transfers – by date. Again, thanks to the fantastic BBC transfers resource for the data.
According to the BBC within-association transfers can begin as soon as domestic competitions finish and they go on (as anyone who watches Sky Sports News will know) until 11pm on 31st August. I was interested to find out what type of transfers happen when and whether “transfer deadline day” really is as exciting as Jim White makes it sound.
It’s been a busy summer, with the usual frantic end to the transfer window. I’ve been looking at the excellent data provided by BBC Football, and playing around with different visualisations of transfers involving English and Scottish clubs.
I grabbed the transfers for the entire window and it’s such a rich data set that one post is not really enough space to visualise it all properly. For this first post I thought I’d just display the basic patterns that appear by club.