Math Formulas (via trindade.joao)
Since the last edition of the crunchy power rankings I've made some significant modifications. First, based on some work I did looking into extreme differences in strike rates between New York and Toronto, I've decided to switch from net shots in the box to net shots on goal. Based on last season's stats, they're actually equally correlated with PPG, so it's a fine substitution. It remains to be seen whether New York's extremely high (and Toronto's extremely low) rates of goals per shot can be maintained long term, but for now I'm going to help take them into account by shifting away from shots and into shots on goal. Note that this still doesn't fully account for New York's results and we sill rank them low (13th, in fact), in part because they're also converting an abnormally high number of shots on goal into goals (meaning they aren't getting saved or blocked). Again, it's possible that this is largely luck and it's possible that their shots are especially unsaveable. I'll take this step for now and we'll see how it goes long term. It's worth noting that in their one MLS match since I looked at the stats, New York had a pedestrian strike rate of 6%. Small sample, but maybe the start of a regression.
Second, I've reweighted the stats in the final score. The previous even weighting was based on an error in my original calculation. In fact, net box shots (and now net shots on goal) should be weighted as much as offensive passing and duels combined, with passing weighted a little higher than duels.
In other news, I'm sure there will be suspicions of homerism, because I have the Seattle Sounders ranked at #2. That's a climb from 3rd in the first rankings, despite the fact that Seattle's gone on their worst run of the season. But has the run really been that bad? By points per game, yes. But consider that they generally dominate every game they're in in chances. In the Columbus game, the Crew had exactly 2 shots on goal, and they both went in, and that was the winning margin. That's not very likely. And in the crunchy power rankings, we don't rely on the unlikely. We just look at how the team is playing and how that correlates to long term success. Also, the team ranked 1st is probably a big surprise. Full table and some more analysis below the jump:
|Team||Net Duels||Net Shots
|Seattle Sounders FC||2.71||1.49||7.32||90|
|San Jose Earthquakes||-1.37||1.91||1.03||80|
|Sporting Kansas City||4.6||0.62||6.24||78|
|Real Salt Lake||0.6||0.66||3.41||69|
|New England Revolution||0.96||1.11||-2.21||65|
|Los Angeles Galaxy||-0.41||0.63||-2.08||57|
|New York Red Bulls||-1.54||-0.65||-2.29||38|
Chicago are the big beneficiaries of shots being weighted higher than the others. They currently lead the league in net shots on goal, which probably comes as a surprise to most people who aren't watching the stats. Their worst stat is net duels, but again that's been de-weighted a bit. Note that Houston is ranked 4th in a system with no home/road adjustment despite playing an almost entirely road schedule so far. So they may be due to break out if they can play even better with a heavy home schedule for the rest of the season. Toronto remains much more highly ranked in the stats than they are in reality, but they really are fine statistically, even when switching to shots on goal. They're really suffering in the conversion rate — the rate at which shots on goal actually go in. Still luck, or are they doing something wrong? More of the Ryan Johnson effect? Time will tell.