True, but SSS.
I just did a bootstrap resampling of their defensive drives in tonight's game, and the 1-sigma confidence intervals for their points allowed per drive tonight is 1.56 - 3.78 (so 14.0 - 34.0 points, total). The current league average is 1.87 points per drive. Tonight the NEP gave up 2.67 pts/drive.
24 points on 9 drives isn't good, but the sample is so small that it's in the "one play goes different and it's so much worse/better" category.
edit: bootstrapping might not be the best method for putting appropriate numbers on the confidence intervals in this circumstance, as it doesn't account for things like opponent, starting field positions, etc., but those things only increase the uncertainty so I think the numbers above are a lower--limit on the "one-sigma" uncertainty.