Exploring The Causes Of A Sack

Found this latest research from "Advanced NFL Stats" to be quite interesting.

Conclusions • A sack is attributed mainly to a quarterback taking too long to pass the ball and his offensive blockers getting beat physically by the defender. o The average time from the snap until a sack is initiated (4.3 sec) is far greater than the average drop and pass time of an NFL quarterback (2.7 sec). o Of the sacks that occurred in the 2011 season, 77% were in situations where an offensive blocker was beaten physically

• The offense was leveraged and offensive tackle was physically outmatched on 37% of sacks. This demonstrates that the offense, specifically the offensive tackle, was simply overpowered, compounding the negative effects of the quarterback taking too much time to pass the ball. This data suggests that perhaps the team should potentially look for ways to aid the offensive tackle in pass protection.
o One potential way to help in pass protection would be to devote more players to blocking. However, given the data, leveraged offenses are obviously no less apt to give up a sack (84% of sacks occurred when the offense was leveraged). Therefore, perhaps dedicating additional blockers to stop a defense is a waste of resources. It might be more beneficial for the offense to have more targets for the quarterback to pass the ball to. This larger number of potential targets would hopefully lower the time the quarterback holds onto the ball and negate poor offensive tackle play by increasing the probability that an offensive player will be open to pass to in a shorter amount of time, thus limiting the team’s sack count.

• Beyond reducing the time the quarterback holds onto the ball, the other option is to look for scenarios where the average time until a sack occurs is significantly higher. Plays that involve play action seem to offer the quarterback a little more time to pass the ball (on average about .3 extra seconds). However, given that there is not a significant difference in the average time until a sack occurs in each scenario observed in this study (the majority of sacks occurred around 4.3 sec), it would be interesting to find a scenario where the quarterback has more time until a sack is initiated.

• Unfortunately, the reason for the quarterback taking too long to pass the ball is unclear as the data does not show whether or not the extra amount of time taken was due to the superior coverage of the defense downfield or the indecisiveness of the quarterback (a subjective opinion).

• In situations where there is a non-traditional blocking scheme, there are almost twice as many sacks when facing a non-traditional rush than a traditional rush, regardless of whether the offense is leveraged, neutral, or non-leveraged (89% difference). It is unclear what causes this discrepancy, but it would be interesting to find out why so the team can plan accordingly

[url=http://www.advancednflstats.com/2013/05/exploring-causes-of-sack-pt-1.html]http://www.advancednflstats.com/2013/05 ... -pt-1.html[/url]

I thought I'd post this because I find it very interesting in general & especially when considering how many people comment upon Chevon Walker's ability as a blocker and their comments regarding the quality of OLine protection.

This study clearly shows that the majority of sacks are a combination of the quarterback taking too long and an offensive lineman getting beat one on one. So why are we debating Chevon's blocking ability? Is it it not better to have a tertiary option out of the backfield (check down) and have the QB make a good decision (check down dump off) if the pass rush is in his face?

Just thought I'd put it out there. Have at it everyone.

Interesting assessment - I might have to dig deeper into their statistics if it's available (I didn't open the link yet).

As for Walker, I'm on record for stating that I think he's a decent pass blocker, and provided stats last year of one game where people had criticized his blocking, showing that in fact, he had blocked quite well. Not perfect (he did get beat once or twice that game), but who is? Even Cobourne, known for his great blocking, gets beat once in a while. I also think that Walker is very good at releasing to provide an outlet for the QB - exactly what the article is calling for, providing more targets. In fact, this is where he excels, allowing him to then use his open-field running skills. The biggest problem I saw with Walker was an inability to get the extra yards after contact. Better than Cobb, who seemed to let up before making contact, but not as good as Cobourne, who seemed to be able to get an extra yard or two after being wrapped up. Hopefully he's been working on that's over the winter.

I think this article also applies to Stephenson. Is it better to have him as an extra blocker on the line, or as an extra receiver down field? The article seems to be making the case for more use of the tight end as a receiver, and Stephenson showed last year that he is capable of doing that, as well as being a good blocker.

I think you can have your cake and eat it too. While most sacks are going to be the result of the line getting beaten + the QB taking too long, a good blocking RB can help that by delivering a good chip-block or full-block at the line and then check-releasing into the flat to be precisely that safety outlet for the QB so he doesn't have to take the sack.

I do agree that a strong O-line is your best defense against sacks. If the QB needs to adjust protections at the line pre-snap, he should (and Burris is a veteran, so there's no reason he can't), but at the end of the day, the offensive line keeps your QB upright.

Something missing from the first post is the mobility of a QB. The traditional, NFL-style drop-back QB might be sacked more often (if he doesn't get a pass off, throws the ball out-of-bounds, etc), while a more mobile QB might be able to avoid a sack and/or gain positive yardage.

Examples: Danny Mac in 1999, when the team had the lowest sacks allowed - a combination of an amazing O-line, and a QB who had the fastest release on passes.

Burris - Veteran QB with good reads, but also having the ability of being a legitimate running threat.

In an effort towards full transparency ... I give you "The Value Of A Sack"

[url=http://www.advancednflstats.com/2008/11/value-of-sack.html]http://www.advancednflstats.com/2008/11 ... -sack.html[/url]

[b][i]"If we average the expected points of all situations in which there wasn't a sack, and compare it with the average expected points following plays that did result in a sack, we get a difference of 2.0 points. In effect, a sack swings the balance of the game by an average of 2 points in favor of the defense, either by forcing a punt or a longer FG try, or even just putting a team in a predictable passing situation. That's a big swing for a single play. A turnover is generally worth 4 points, so a sack could be thought of as half as good as a fumble or interception.

Two points seemed like a lot, so I dug a little deeper. One reason sacks are so valuable is that they often result in fumbles. In fact, even if a quarterback isn't tackled, if he's forced to fumble, even just by a hair, that's technically counted as a sack. So when we take fumbles out of the equation, sacks are worth an average of 1.7 expected points. It's the possibility of a fumble that makes up the extra 0.3 points.

One thing to note is how the effect changes on the cusp of field goal range. On 3rd down, sacks on plays that started between the 20 and 30 are especially costly to offenses, probably because this makes a field goal far more difficult. On 2nd down, the effect is slightly different. And on 1st down, there doesn't seem to be any special effect near FG range. Offenses still have 2 more downs to recover any yardage given up."[/i][/b]

With so much talk about the TiCats lack of pass-rush and lack of sacks this past Friday in Toronto ... I thought I'd re-open the discussion and pose the following;

Neither QB was sacked until the late stages of the 4th quarter, that being Henry of course.

The Argos Dlinemen pinning back their ears ... Chris Jones bringing pressure ... Knowing Hamilton has to score ... With less than 3 & a half minutes to play ...

Up until that point, no one sacked Henry ... No one sacked Ricky ... So is Hamilton's DlIne that horrible in comparison?

I think the likelihood & occurrence of the sack is declining in the CFL. These QB's get rid of the ball a lot quicker and the Oline's are a lot better across the league.

I feel the Cats D-line is a little bit weak, in the sense that they seemed to get held up too much at the line. The only time Hamilton really got any pressure on Ray was when they blitzed, and if you live by the blitz you die by the blitz.
It was also evident on stopping the run, they got manhandled quite a bit. However, in fairness a couple times the defensive play was called opposite to what the offence was running.
I noticed this a lot last year under Creehan, it seemed he would call for a defensive play in which the whole line would go left and the offence ran the play in the other direction, effectively you have taken your entire D-line out of the play. Creehan's play calling was not very good last year.

As for Walker's blocking he's not that bad, once again it depends on the play and his assignment. His primary role to run with the football which is does very well, being able to block is helpful but it's not what he does best.

I'm a firm believer that football is won in the trenches, whichever team controls the line of scrimmage will win the game. An average QB can look very good if given time to throw, whereas even if you had Peytong Manning, with a poor O-line you fail.
A strong D-line causes all kinds of problems because it requires the offence to use 5 or 6 guys to block only 4 players freeing up more defenders.
If you have the best offensive and defensive line in the league, chances are you will be one of the best teams in the league.
That's one of the primary reasons the Tiger Cats were so successful in the '60's, and early '80's.

I doubt Chris Jones is sitting around feeling good about his D's performance either. Bottom line is: neither defense was very good, but Jones's D did just a little bit more, enough to help Toronto win.

They got the equivalent of a sack, or would have had the officials made the right call. On the Argh's final drive, Ray was being chased and threw the ball to ... no one, and it fell four yards short of the line of scrimmage. CFL Live Play shows the pass being to Owens, but the only player in the area from what I saw was a O-lineman.

Ray is one of the most difficult QBs to bring down. To bring him down (or equivalent) once in a game isn't that bad.

The Cats were second last in sacks last year with the same players returning to play. I can't believe they cut Brandon Gugese
over Greg Peach. The D Line will need to improve so the DB's can cover. Lets hope this problem will e rectified quickly.

Glancing at the detailed report (https://docs.google.com/file/d/0B--fKJN ... sp=sharing), it looks like he compared the number of sacks under each of his formation conditions without normalizing for how frequently each condition occurred, whether there was a sack or not. Essentially, he didn't answer the question "for every hundred times a particular condition occurred, how often did that condition result in a sack?"

OT but still on stats: There's an interesting story about statistics and the analysis of damaged aircraft returning from missions in WWII. Mathematician Abraham Wald was tasked with developing a method for estimating the vulnerability of different parts of an aircraft to enemy fire, based on observations of the returning planes. Though Wald's actual methods are quite technical, the glib interpretation that makes the rounds nowadays is that the recommendation is to reinforce armor in regions where the returning planes were not hit. The point being that the damage on returning planes was insufficient to shoot them down, so the inference is that the planes that did get shot down were likely to have been hit somewhere else. The tricky part is that Wald wasn't able to make observations of the planes that couldn't return because they were shot down.

The story has been retold in some popular books and online. A re-print of Wald's original work can be found here: http://cna.org/sites/default/files/rese ... 320000.pdf

For the record, Wald concluded among other things that "The greatest probability of being destroyed is .534, and occurs when a plane is hit by a 20-mm cannon shell on the engine area. The next most vulnerable event is a hit by a 7.9-mm machine gun bullet on the cockpit."