Premier League Projections, Week 13

Southampton had a huge match this week, home to Manchester City. For the first half, they dominated play, creating far more good chances than their moneyed opponents. A terrible mistake by Jose Fonte should have given City a penalty, but on the run of play you would have to give the edge to the Saints. I had the expected goals difference at about 1.0 for Southampton to 0.3 for Manchester City. Then at halftime Morgan Schneiderlin came off. Maya Yoshida come on as a center back with Toby Aldeweireld pushing into defensive midfield and Victor Wanyama taking a more attacking role. It quickly became clear that this new lineup was vastly inferior. An out-of-position Aldeweireld allowed Yaya Touré the chance that broke the tie, and terrible midfield errors freed Frank Lampard for the second goal to put the game out of reach. City dominated the xG in the second half 1.6 to 0.3 and coasted to three points.

Schneiderlin showed in the first half why Daniel Levy had been prepared to drop £25m on the midfielder, as he soundly outplayed Touré and Fernandinho in the center of the pitch.

But the second half has got me wondering whether something is wrong with my projections. The stats have been all-in on Southampton for a while, even after accounting for the Saints' easy schedule. My feeling was mostly that if beating Stoke City, Sunderland and Newcastle were so easy, perhaps someone should have told City, Chelsea and Liverpool. Nonetheless, there are a lot of skeptics. I was feeling pretty good about my position through that first half. Southampton were clearly playing at the right level. And then Schneiderlin went down and the second half happened.

And I started thinking, am I missing something really important about smaller clubs? I remain skeptical that Southampton lack the know-how-to-win and the history and whatever, but they really do lack depth. The dropoff from an elite midfielder like Morgan Schneiderlin to a cromulent one like Jack Cork is big. With Cork out injured, the drop goes all the way to an out-of-position center back. For a team like Southampton to compete at the highest level, they need 35-match seasons from a bunch of important players. If just one or two guys go down at the wrong time, in the wrong positions, suddenly the top four competitors look mid-table blah.

That's an issue that I have no direct method of accounting for in my projections. I do include team payroll as a factor, and it's helped keep Southampton's projections down to a more reasonable level. I wish I could build projections up from individual player quality, but I don't believe we'll have good player value stats until camera tracking data is available. Until then I have to make do with team statistics, and a hot run of form and health can end up misleading. We shall see. If Schneiderlin is out for a significant period of time and the second half is any kind of indication of his importance to Saints, Southampton are headed for trouble and it will take the stats a while to catch up with reality. I hate getting things wrong. I get things wrong a lot, because such is the nature of writing words, but I'm worried I've got this one wrong.

But This is a Spurs Blog, Right?

Hey, you're right! And you know who played real good football this weekend? Tottenham Hotspur. A solid across-the-board win for Spurs saw the North Londoners clinically finishing chances and playing a nearly spotless rearguard defense for the final half hour. Everton have been struggling in both the underlying stats and the table for several weeks. They looked every inch the capable mid-table side and possible Europa competitor that the projections have been calling them for a little while. It wasn't a dominant victory, but it was a clear one. Spurs move up in the projections, but there's still a lot of ballast from earlier poor performances that will have to be dropped before Tottenham can push toward a strong fourth place projection.

So on to the numbers.

Premier League Projections

• Full metholodogy, don't forget to click on "Methodology" for lots of math.
• EPL Advanced Statistics page, with many of the component stats for these projections.
• Because of rounding, the numbers may not all add up quite right.

• There was another huge result this weekend. Queens Park Rangers beat Leicester City in the match we all should have been watching while United were stomping Hull and Liverpool were scraping three points against Stoke City. It's not just that QPR came back from an early goal conceded to win 3-2, but that they created chances at will. The Foxes simply do not appear to have the defense to remain in the Premier League.
• Hull City's inevitable march down the table continued. Their 0.300 expected goals ratio is solidly in the Reading zone. I think this team is basically terrible.