My issue is he mixes pace adjusted with heavily pace dependent stats, which lessens their value. Take this chart:
The first half is some offensive efficiency, but super pace dependent. Of course more red zone trips equals more total points. You also gain that through more possessions and a faster pace (with the trade off being the other team also gets more possessions/points). The right half mixes two aspects. The first if yards per play is pace neutral efficiency. The second, there is a correlation, as this shows but he is leaving out the full statistics of how close. The data distribution is not the most defined line or he would have like made the data points a more narrow range to highlight the correlation. The reason is straightforward: run more offense get more points but give up more (think OK state this week).
But, if the Gators are going to continue to play at one of the slower paces (no indication that is ever changing - annual Robbie article aside) comparing these YPP stats to points per possession would give us a better sense of the offensive efficiency.
The third down conversion rate comparison is also a choice at a less than optimal stat. For instance, it ignores scoring on 1st or 2nd down. Success rate per offensive play inside the red zone and outside I suspect would have had a higher correlation (at least to point per possession, if not the pace weighted total points per game).
I’m not suggesting his stats in this particular article aren’t more information, and therefore helpful. They just aren’t as helpful as presented when you look more closely.
YAR is pace neutral and pretty helpful, and I don’t think it punishes non running qb’s so much since it only accounts for actual runs, not total yards (which would be more for running QBs), plus everyone gets sacked and those hurt the offense equally. The major weakness with the stat is it gives no value to the massive errors (worse than sacks) or a turnover.
TLDR- the mixing of pace neutral versus pace dependent stats makes his analysis less reliable. YAR is pretty telling but can’t give the huge negative value to turnovers.