Spurs lost, but this was a game we were pretty much expected to lose, by the numbers. It's definitely frustrating that outside of two defensive errors, Spurs were the better side, but the way that Liverpool overran the Spurs midfield for about 60 minutes of the game is exactly the sort of thing that can cause defensive howlers. Some credit here should also go to Liverpool, too.
The underlying numbers have Liverpool only slightly outplaying Spurs overall, even accounting for the big chances they allowed. This is another kind of frustrating, but it also gets at why we should not panic. Spurs played a good game against excellent opposition. There is little reason to worry, based on that result, that Spurs aren't the exact same top 3/4 contending side we and the numbers considered them to be before they traveled up to Anfield.
Based on discussions in last week's comments, I have renoobulated the power rankings a bit. First, they are now shown relative to 100 (a perfectly league average club would rate 100). Second, I have built an overall team quality estimate using the Monte Carlo method. I play out 10,000 games of each club, based on estimated quality of attack and defense, against a perfectly league average opponent. I compare their points achieved over 10000 games to the league average, and that produces the team quality estimate. I'm currently calling it "Team+" because I am not good at naming things.
|West Bromwich Albion||89||89||102|
|West Ham United||88||81||98|
|Queens Park Rangers||73||75||113|
A little nerdery: There is a bit of change in some of these ratings because I realized I was being inconsistent. My season projections are built by regressing G/SoT most of the way back to league average, adjusted for big chances and shots in the box. I had been building the power rankings entirely off of the SoT/BC/SiB numbers. So for consistency, I have moved the power rankings over to the "heavily regressed" methodology. This has helped, in particular, United, Chelsea, and Reading. That's how bad Reading are - this is what it looks like when I alter my methodology to their benefit.
The loss to Liverpool was already mostly baked in to the numbers, so its effects are relatively small for Spurs. Liverpool, though, needed the win to keep their possible dream run alive.
|Club||W||D||L||Pts||%Title||ΔTitle||%Top 4||ΔTop 4||%Rel||ΔRel|
|West Bromwich Albion||16||6||16||54||0%||0||.1%||0||0%||0|
|West Ham United||12||8||18||44||0%||0||0%||0||1%||0|
|Queens Park Rangers||7||13||18||33||0%||0||0%||0||75%||-7|
- Chance of Spurs taking 3rd or better dropped from 56% to 50%, chance of 2nd place dropped from 5.5% to 4.5%. (There were two seasons out of 10,000 where we beat out United for the title.)
- In other numbers that don't show up in the table, City did miss out on the Champions League 12 times, and Swansea qualified thrice.
- Chance of Aston Villa being relegated on goal difference actually increased slightly, from 5% to 5.5%, though their overall chances of remaining in the top flight were massively improved with the win over Reading.
- On Reading, I wonder if they wouldn't have been better off with worse fortune on G/SoT early in the season. Their conversion rates obscured their overall crapulence, and now it's most likely too late for a new manager to make a difference. Certainly it's too late to add the talent the club needed from the beginning.
- Both QPR and Villa taking three points on the weekend was enough to hurt Wigan, even though the Latics didn't play at all.
- As before, the numbers often do not add up properly, due to rounding. I prefer to publish whole numbers wherever possible, for readability and to avoid giving a false sense of precision. These numbers should not be considered anything more than rough estimates based on rough estimates, with obvious blind spots. But hopefully they're fun.
A little more nerdery: the +/- numbers don't line up with the numbers from last week's table. This is because of a programming bug that has been fixed. One of the first comments I got was from Lennon's Eyebrow, asking why Liverpool projected below Everton and why Spurs projected to drop so many points. I was all like, no man, the numbers, the numbers, but of course he was right and I was wrong. There was a programming bug which led to bad teams having too high a chance of taking a result from good teams. It's since been fixed, and the +/- numbers are based on the fixed percentages from last week.