Dan Hanner builds on an e-mail I wrote him discussing whether rating systems should consider 3-point percentage defense or free-throw percentage defense when they rate teams. In my e-mail to Dan, I focused on a specific example - North Dakota State. If you're interested in the details, here's the relevant portion of that e-mail excerpted below: In 2013, NDSU allowed 26.2 free throws per effective (non-turnover) possession, and that number ticked up slightly to 28.7 per 100 in 2014. Yet the difference in "free throw defense" came from the FT% of its opponents, which moved from 69.6% to 75.3%. While in 2013 NDSU gave up about the number of points you would expect from the line given its defensive capabilities, in 2014 it gave up 1.7 more points per 100 possession thanks solely to FT shooting by its opposition, completely out of its control. The same story with 3s, though that requires a belief in Pomeroy's theory on the topic. You'd have expected the opposition to score 11.9 points per 100 effective possessions on 3s against NDSU in 2014, but they scored 13.2. That's another 1.4 points outside their control (rounding). Add it up, and that's a 3.1 point swing in 2014 based on luck, not talent. Take 3.1 points off their defensive rating, and you get to a fair defensive rating of 100.0 points per 100 possessions, which would rate 74th in Pomeroy. If you did the same exercise for 2013 and strip out the Bisons' good fortune, you'll get a fair defensive rating of 70th. Basically the same as their fair rating in 2014.