I was messing around with some previews I’m thinking of doing for next year when I came across some of the awesome bubble plots that Derek Zona has been doing to compare large numbers of NHL players. It occurred to me I could quite easily adapt this to compare the teams in the Premiership.
I’m going to use these bubble plots kind of frequently I think so it’s probably worth giving it a proper explanation.
x-axis: PDO. Further right = more luck
y-axis: TSR. Higher = better team
Bubble size: Points. More points = bigger bubble
Bubble shade: Goal difference. Green = positive, red = negative. Lighter shade = further from 0
A few short summary points that might help with interpreting this:
Manchester City were good (high) and lucky (right). This translated into a lot of points (big bubble) and a high positive goal difference (light green).
United were less good (lower) and luckier (further right).
Liverpool couldn’t buy a goal this year (darker green). If I’m a manager looking for a new post this is the perfect time to take over at Liverpool, they’re almost guaranteed to score more points next time around, even if they aren’t as good.
The league was kind of split into three tiers. Seven teams were a cut above the rest (from Everton upwards). Then there’s a group of about 10 teams that are pretty close in performance with Wolves and Stoke being a step further back.
If there’s one thing this plot shows us it’s that (with the possible exception of the Manchester teams) there is literally no correlation between TSR and PDO.
How do we like the look of these? is there anything I could do to improve the layout?