Yups, exactly what I was picking up from it. QPR and Reading gone and Wigan still in it. Their fixtures look quite good going forward and Wigan-Villa on the last day of the season promises a much needed end of season nail-biter.
Re-imagining Storeytime
- Thread starter Eboue
- Start date
Yups, exactly what I was picking up from it. QPR and Reading gone and Wigan still in it. Their fixtures look quite good going forward and Wigan-Villa on the last day of the season promises a much needed end of season nail-biter.
Yep, you're right. The issue was with the numbers at the top. I edited the graph manually but didn't take those out. I've fixed it in the first posts and I'll repost it here. This is updated through today's matches assuming Chelsea hold on to a 2-0 at home to the Fightin' Fat Sams.
Uhmm... What the graph suggests is that Sunderland is the team thats better off out of the relegation candidates. Moreso than what the real table suggests.
Its an interesting way to look at the table but the problem is that because there are so few games during a season its very unlikely that this kind of table will reveal anything that the normal table doesnt already show us. It would take a real anomaly of a season for this kind of table to correctly and logically predict that one team is likely to overtake another in the future.
If we had a league with 100 teams this kind of table would be much more relevant.
Well I think what they were saying is that most people didn't consider Sunderland to even be a relegation candidate. As you point out, they are better positioned than most but the surprise was that they are even considered.
I think the best time for this graph is through matchdays 15-35 or so. At the beginning there won't be enough data to be helpful and at the end, the fixtures have mostly caught up to each other.
Yea but if people didnt consider Sunderland relegation candidates before, it makes no sense to consider them now because of your table. Your table shows Sunderland better off than the real table.
I agree this kind of table could be very useful for the second half of the season. Can you use your data to calculate backwards and see what the table looked like after 19 rounds or is that too much work?
Yeah I can do that easily. Give me a few minutes.
Surprised how "wrong" it is tbh. Expected better from maths!
Differences from OEP to actual table:
-OEP liked Chelsea more
-OEP liked West Brom less
-OEP liked Sunderland more
-OEP liked Wigan more
Keep in mind though, this isn't designed to be a predictor as much as it shows the true table. This will match up to the league table at the end of the year, whereas a predictor wouldn't. If you are interested in a predictor, here are the latest Pythagorean Expectation results.
