Swansea City lost 4-1 to Manchester United at home last Saturday in an underwhelming first national match for NBC Sports. It's an interesting one for the numbers because United didn't actually produce that many more good chances than Swansea. They just converted almost all of them. The Red Devils had 6 SiBoT (that's shots on target from inside the box), 1 SoBoT (shot on target from outside the box) and two big chances. That's a strong performance from which you'd normally expected two or three goals, but United bagged four. Swansea had near identical underlying numbers here with 5 SiBoT, 1 SoBoT and two BC, but they only converted one. The takeaway from most observers was that the 4-1 scoreline flattered United, and the stats back up that analysis. Swansea took the game to United quite well, and on another day might have pulled out three points or at least an exciting 2-2 draw.
So this is a darn good football club that we're facing. Despite that, the projected win numbers are a bit gaudy. There is a very large home/road effect in football, and Spurs should be a better than even bet to pick up another three points this weekend playing at White Hart Lane.
Match Projections and Odds
That's good! Hope that's right. These odds are only slightly worse than what were projected for Tottenham traveling to Palace last weekend. Basically, my projection engine thinks that the difference between a very poor team and a roughly league average side is only a little bit larger than the difference between playing at home or on the road.
The season projections based on different game outcomes are pretty similar too. Just with a little bump up.
As I noted last week, the projected points have a little more separation than you might expect. In theory, if Spurs are the same team, then they should project to take the same number of points from the last 36 matches following Swansea. This is because of the randomization component I use in team quality. My team quality estimates are randomized around a projected mean, so in scenarios where Spurs aren't quite as good as we hope, Spurs are more likely to lose to Swansea and to drop a few more points over the rest of the season.
This mirrors the experience of watching the weekend's game as a fan, I think. When you watch Spurs, you're not just watching hoping they win, you're watching hoping they play like the quality side that you can imagine putting in a run for a top three finish.
The Bloomberg Table
Bloomberg Sports is running a projected EPL table through the season, and I'm imagining this as a little friendly competition. So far, the only differences between our table projections are the ones visible from the preseason. As I discussed, Bloomberg appears to be relying heavily on goal difference from 2012-2013 in their projections, which leads them to rate United, Chelsea and Arsenal much more highly than I do, and City, Tottenham and Liverpool not quite as well. They had United ,Chelsea and City basically tied for the top three positions, far ahead of Tottenham and Liverpool, and a step ahead of Arsenal in preseason. In my projections, I have a much closer race overall, but with City as solid favorites and the next five not too far separated.
There appears to be little change in the underlying team quality ratings in the Bloomberg table, which makes sense. You don't want to radically re-think such things after just a week of games. In their table as in mine, the big loser in probability is of course Arsenal. The extent of the loss is quite different, however. The Gunners are still projected four points ahead of Spurs with a 53% chance at a top four finish. I have Arsenal five points behind Tottenham and just 33% for top four. Now, neither Arsenal pulling out another top four finish or Arsenal falling short would prove anything one way or the other about my projections and Bloomberg's. But in our little competition, it appears that Arsenal will be one of the important teams to watch.