Pressure Points - MMQB New D ranking

[richpjr]

I don't see why that's not clearly missing from those stats

Because they don't think it matters - hence the lack of them including it.

[homeless simpson]

they have stats for qbs on completions and yardage per down, distance and quarter, under pressure/blitz, stats based on the score (from behind or leading) etc. you would think they can easily compile stats for the pass rush....

[Den60]

You really have to wonder if this evaluation means much. We are in the bottom third of the league in 3rd down percentage on defense. That stat doesn't take into account the down and distance; we have given up a lot of 3rd and longs in our first couple of games and to me that has more of an impact than a team routinely defending 3rd and a yard or two. To me our inability to get off the field on third and long is directly related to our ability to pressure the QB. We are giving up 9.4 yards per pass and 370 passing per game, both worst in the league. If this evaluation was meaningful we should be in the middle portion of the league rather than at or near the bottom.

[Yubaking]

And yet, still better than just sacks...

The idea of factoring in the effect of hurries, hits and penalties drawn is conceptually a better measurement than just sacks, so I agree with your statement in theory, but I disagree as to the weighting assigned by the cited source in several respects.

The weighting system used in the cited source both undervalues sacks and overvalues hurries and QB hits (QB pressures).

Sacks are undervalued because their value is reduced for an "easy sack" whereas no such reduction is made for an easy pressure, which makes no sense because a sack is more difficult to obtain than a pressure when it comes from the same situation. Also, an easy sack is defined to include a regular sack that occurs in the last 4 minutes of a game when a team is trailing by 14+ (i.e., a known passing situation), but there are known passing situations throughout a game, so that distinction makes no sense and, again, there is no corresponding reduction for QB pressures under the same circumstances. Thus, the system is flawed in that it intentionally undervalues sacks.

But even worse than the undervaluation of sacks is the wildly overvalued QB pressure component (hurries and QB hits) in terms of its effect on the defensive result. While I have no exact numbers, to demonstrate the point from a conceptual basis, let's suppose that the average completion rate is 60%, but that it rises to 74% when there is no pressure and falls to 40% when there is pressure (hurries or hits). Further, without running the exact numbers, it appears that the average percentage of pass snaps that the QB is either hurried or hit is right around 40%. (We are at 43.8% and yet have yielded a 69.5% completion rate anyway.) Those numbers fit together with pressure on about 40% of snaps (74 times .6 plus 40 times .4 equals just over 60).

I submit that a QB pressure or hit that results in a completion adds no defensive value whatsoever and should be dropped as an element of effective defensive pressure. So, counting a sack as a standard 1.0, the first thing that needs to be done to ascertain the value of a random pressure is to remove the 40% that result in completions, leaving us for the moment with a standard value of .6. Then, we need to realize that of the 60% incompletions, some were going to be incompletions anyway. Using the estimated numbers referenced above, we can see that the completion percentage is reduced from 74% to 40% when there is pressure. The reduction in completion percentage caused by pressure is about 46%. So, taking the .6 representing the incomplete passes and multiplying by .46 representing the average causal effect that pressure is shown to have on incompletion percentage, we have an effect of pressure number of .276.

Of course, even this reduced .276 figure fails to take into consideration the value of yards lost by a sack, which means that the figure should be reduced even further, but the point should already be very clear--that by the time we are done with the analysis, the effect of a pressure is going to be less than a quarter of the effect of a sack.

When this approach is taken, while our team's overall ranking may not be adjusted that much, Freeney's individual ranking drops dramatically compared to the other pass rushers listed and provides a much more accurate description of what Freeney's overall impact has been or, perhaps more accurately, has not been.

[Beerman]

My only point of contention with the way they have it weighted is the hurries and the hits. I think the hits should be about .25 since they really don't alter the play, but it does punish the QB. Still it has no effect on the specific play.

The other weighting that I would try to change is the hurries. There are hurries that cause incompletions and ones that don't. The ones that do should be weighted higher and the ones that don't change the outcome of the play should be downgraded.

In any case, I do still think it provides a better overall picture than the conventional sack stat that was pretty all we had to grade a pass rush on.

I don't put too much stock into the down and distance. You either get to the QB or you don't. The 3rd downs allowed is fine on it's own to grade that specific circumstance.

[Yubaking]

My only point of contention with the way they have it weighted is the hurries and the hits. I think the hits should be about .25 since they really don't alter the play, but it does punish the QB. Still it has no effect on the specific play.

The other weighting that I would try to change is the hurries. There are hurries that cause incompletions and ones that don't. The ones that do should be weighted higher and the ones that don't change the outcome of the play should be downgraded.

In any case, I do still think it provides a better overall picture than the conventional sack stat that was pretty all we had to grade a pass rush on.

I don't put too much stock into the down and distance. You either get to the QB or you don't. The 3rd downs allowed is fine on it's own to grade that specific circumstance.

Conceptually, we are definitely not miles apart on this at all. We seem to agree that pressure that results in a completed pass does not have that much value and that pressure that results in an incompletion does have value. For this reason, I do not give Freeney as much credit. For all of the "pressures" he has had and our team has had, we are still yielding nearly a 70% completion percentage, which is not good.

I also don't put a whole lot into down and distance for the reason you cited.

[homeless simpson]

You either get to the QB or you don't

which is why sack stats still matter.

[Beerman]

which is why sack stats still matter.

Who says they don't? It's just not the only thing.