Original story here
Specifically, you keep moving the goalposts, in that your variables measured don't necessarily reflect what you're trying to defend/explain away. Here are some better ways to measure them.
For example, #1: That Damn Discipline Problem!
You measure it in terms of personal fouls given to Lions players, which is an odd choice. It's odd because the discipline issue refers to the number of penalties committed by the team. The personal fouls are an off-shoot of the team not being disciplined, but the issue is a lack of discipline leading to penalties.
Then you have to analyze which penalties should count. Some are a lack of skill at a position -- a CB committing PI rather than give up the TD he was beat for, for example. So in that case, it might be better to measure, say, pre-snap penalties as a measure of discipline.
Undisciplined teams would have more pre-snap penalties than disciplined ones, correct? Well, the Lions were tied for 9th of 32 with pre-snap penalties. The average was 39.5 with a standard deviation of ~9, which puts 23 of the 32 teams roughly in the same category, penalty-wise. Five teams exceeded that deviation and in my view would be considered more undisciplined than the others. We were not one of them.
Your verdict: NOT undisciplined. (in terms of personal fouls)
My verdict: NOT undisciplined (in terms of pre-snap, discipline penalties)
math is my own. god only knows if 4 years of grad school research stuck.
#2: The Lions have a shoddy secondary that will end up losing them games in 2013.
There is no real way to measure how they will do in 2013, primarily because of the variables involved in passing. As you noted, teams that are losing will get run on more often than not, and teams we are beating will pass more. If we're up by a lot, teams will pass for even more "empty yards" as I like to call them --stat padders that occur while we go into a prevent defense. By the same logic, though, measuring total yards could give us the same dilemma of garbage time results.
Thankfully, we have advanced stats. In this case, I chose DVOA. It is explained here but is basically a sum of the total plays, taking out the effect of who we played and factoring in down and distance. It's a way to neutralize stats so they reflect the quality of play, rather than just the raw number.
In this analysis, the Lions were 24th of 32 teams in defense. They were 21st in pass defense and 24th in rush defense. In other words, the Lions collective performance was in the bottom quarter of the league's defenses. Now, did they upgrade? The data does not exist to measure this, as changes in players (including rookies who have 0 stat history), schemes, roles and injuries make it near impossible to predict.
What positive we can take from this is how consistent the Lions were -- though it was consistently bad. The Lions were first with the least amount of variance from week to week, suggesting that PERHAPS, if they can improve, they might keep those gains throughout the season. That suggests that IF the players gained are better and the playcalling is good, the Lions SHOULD have an improved defense -- including pass defense -- in 2013.
Your verdict: Lions secondary will probably be improved.
My verdict: Lions secondary will probably be improved. (I think your conviction in this is stronger)
#3: The Lions have the best D-line in football!
We can measure that, too!
The Lions were 5th overall against the run in defensive line play last year, again factoring in down and distance. The D-line was 29th against the pass last year. Stats against the run are measure by where the run was executed. (So LT plays are run toward our RDE, KVB if I recall)
LT: 2 Middle (C/G): 5
As you can see, runs up the middle and inside toward Avril were no successful for opposing offenses, whereas runs out wide were near league-average for success. Runs right at KVB were almost the easiest plays in the league. Even with those struggles outside, we still finished #5 for run defense. This suggests that our interior line play is superb (2 and 5 are excellent rankings), and a change in personnel outside + a tweak in the scheme SHOULD give the team dominant DL again against the run. But without more pressure, you cannot call this a dominant DL.
Your verdict: DL was not dominant
My verdict: It depends. DL was very dominant against the run (particularly runs inside and not at KVB) and very bad against the pass. Even average play from the DEs this year would make the team dominant at DL, as the DTs are very above average.
#4: Stafford is nothing without Calvin Johnson.
As you suggest, this is kind of bullshit. A look at the individual stats for reach of the receivers Stafford had bears this out. Let's hypothesize the following. We'll use EPA/P: Expected points added per play, a pretty self-explanatory stat. If Stafford is a good QB even without Calvin, then he would have some receivers who have good results other than Calvin.
Here's what you get.
By that measure, our top two starting WRs were all roughly equal, and Burleson was pretty good as well, meaning Stafford was successful throwing to all three of them. The two starting TEs cost us points. Robo had only 6 catches, so small sample size there.
This all suggests Stafford, like almost all QBs, is a good QB when he has decent receivers to throw to AND those receivers don't drop the ball.
Your verdict: Stafford is good with or without Calvin
My verdict: Stafford is good with or without Calvin
I hope you're not offended by this, and in all likelihood, there are even better ways to (a) ask the questions we want to measure and then (b) measure them with better variable. This is just a quick attempt to provide some better context for some good thoughts, and perhaps we can do more of this in the future.