Since points allowed/scored is the true measure in wins and losses, wouldn't points allowed be the stat that matters most for defense?
Not at all. Too many unknown variables in the points allowed per game will skew the determination. For instance, suppose the offense had 10 turnovers on its own 15 yard line and the defense gave up 0 yards, and had two sacks, but the other team kicked 10 field goals, scoring 30 points. In that circumstance, you would probably say we had a pretty stout defense, but the points per game stat does not reflect that fact. Obviously that is an extreme example, but it explains why points allowed is not a very good stat for judging the defense.
If you are keeping points off the board, something is working. Apply you're logic to that statement. Saturday night sucked, but if you want to take statistics into account, at least take into account stats that tell the whole story. USC only passed for 130 yards, does that mean that the pass defense shut down USC's receievers? Chew on that.
Sure, I'll apply my logic to that statement. Maybe on a lot of drives, the other team's offense started on their own one yard line, drove down to our 5 yard line, and then had a turnover on a bad snap, or even missed a field goal. The defense just gave up 94 yards on one drive, but the other team didn't score any points. That does not mean the defense did well, even though there were 0 points allowed.
You are right, that looking purely at passing yardage is not necessarilly indicative of shutting down the other team's receivers. I agree with you, and I would not try and argue against that. My post was merely responding to the assertion that points allowed per game was the primary stat you needed to look at to determine if the defense is any good.