I have been running projections for the Premier League all season. Each time I list week-to-week changes, but that doesn't give a longer-term picture. I'm probably already overwhelming the reader with numbers, but I can at least regularly post updates on the long-term race. I've been logging the numbers every week, and I've made some graphs.
As noted in my open letter above, I have had some problems distinguishing all of the lines. There are fiveclubs in the EPL who all have the same red as their primary shirt color. (Liverpool, Manchester United, Stoke City, Southampton and Sunderland). On top of that, there are three clubs whose primary colors are white and/or black (Fulham, Newcastle United, and Swansea City). I can only assume that this has been an oversight at the league level which can be easily corrected. I call the Premier League's attention to baseball's Miami Marlins and basketball's New Orleans Pelicans for some useful ideas for updating these user-unfriendly coloring decisions.
Let's Let That Go, Shall We... Now: The Title Race
So I made graphs of all the races over the season. First, here's the title race.
This has been, for the most part, a story of Manchester City's end-to-end dominance. I rated them as the best side preseason, and they have not disappointed. You can also see the cost of Arsenal's inability to get points off their top competitors in December, as the loss to City and draw with Chelsea took the Gunners from nearly tied to clearly second best. I also note the impressively static Chelsea line. I rated the Blues a little over 10% to win the title, and they have remained between 5% and 20% for basically the entire season. At some point they'll have to move up or down.
My stats gave up on Manchester United as title contenders very early. The dark red line that isn't Liverpool drops under 5% by the eighth week of the season and has now entirely disappeared behind late-comers to futility Everton and Tottenham Hotspur.
The less said about that dark blue line the better.
The Top Four Race
I think it's interesting how those dropped points in December harmed Arsenal's title hopes so significantly but show up as little more than a blip here. Arsenal have been good enough and successful enough that it will take a lot more than four dropped points to the other top two title contenders to impact their chances for reaching the Champions League again.
You can also see how, although the White Hart Massacree made a big difference in Liverpool's drive for the top four, it was continuing a trend from the previous weeks. Liverpool's combined 12-3 shellackings of Norwich City, West Ham United and then Cardiff City added significantly to the club's projections. The Reds took 25 shots from the danger zone in those matches while allowing just 11. And very few of Liverpool's close / central shots were assisted by crosses, while their opponents were forced into crossing into the box from wide.
The lack of impact of Tim Sherwood's results on the projections is also notable. As I laid out in my latest projections, Spurs have mostly been taking points without impressing on the shot chart, and this has kept the dark blue line from making any real movement back upward. The win at Swansea is hopefully portends continued improvement, but we will have to wait and see.
The Relegation Race
This is certainly the most difficult graph to follow, visually, due to the number of sides still closely packed together in the relegation race, as well as the coloring problem noted above. (I have a modest proposal that both Norwich and Hull be granted extra points by the league for having appropriately distinct colors. Thank you, Canaries and Tigers.)
Early on, I thought we were going to have a pretty boring race for the bottom as Crystal Palace and Sunderland fell asleep in the starting gates. But with their improved play under new managers, and concurrent collapses by West Ham and Cardiff City, the relegation race looks like it could go excitingly down to the wire. Stoke City, Norwich and Hull remain on the fringes as well, so today's five-team pile-up could even increase in the next couple months.
My choice to use Championship goal difference as a small part of my projections for promoted sides is not looking good right now. Hull have been perfectly ok despite a goal dfiference under 10, and Palace have not been that bad. Cardiff City, by contrast, look like the most likely losers from the group. I will have to go back to the drawing board on projections for promoted sides. The evidence for using goal difference wasn't that good anyway.
I think the last time I put up graphs I said I'd do them roughly every five weeks, but I got annoyed with the Tableau software and didn't follow up. This time I've got a better system in place and I can post an update in a month. I might, if time and interest allows, attempt to create motion gifs of the races. But don't count on that.