enCYCLONEpedia A football fallacy
- Thread starter ChrisMWilliams
- Start date
So since 2009 nearly 40% of all our offensive drives are 3 plays or less and we score just over 7% of the time on these 3 play or less drives. Not good.
I'm not sure bend don't break is a fallacy.
I think it stands to reason that the more plays in a drive, the better the chances you'll score -- because you are getting closer and closer to the goal line.
That most TOs occurred during the shorter sets of downs also stands to reason, because a TO ends the drive.
I think it stands to reason that the more plays in a drive, the better the chances you'll score -- because you are getting closer and closer to the goal line.
That most TOs occurred during the shorter sets of downs also stands to reason, because a TO ends the drive.
"To investigate, I analyzed every drive by every FBS team from 2007-2013."
Holy crap!
Not good at all. It even looks historically for the offense like the only above average had been 4-6 plays to score points on (1.77 us to 1.70 points average). For our defense, 4-6 and 7-9 plays were average or above. The rest of the scenarios, not good.
Last edited:
Yes, I agree. Another point is that in the 16-18 play range it's less likely to score a touchdown in both defense overall and moreso for the ISU defense. That implies there is a certain length of drive that does stall out and prevent a touchdown. I think a major goal of the "bend but don't break" defense is to turn the opponents' touchdowns into field goals by getting tougher near the goal line. The data seems to imply that it does work at least to a certain extent.
I'm not sure that drive length vs. scoring is the best measure of a bend but don't break defense. I think the best measure is how well it can reduce big plays and turn touchdowns into field goals.
Something along the lines of Red Zone scoring percentages of TDs vs. FGs (as well as turnovers).
Well, there are two sides to this coin. On the one, a short drive suggests there wasn't success. On the other, the more plays there are in an offensive series, the more likely something wrong or bad will happen (penalty, turnover, etc.) that will halt a drive. That's why points per possession is a very helpful statistic when trying to analyze the efficiency and success of an offense (or defense). Field position is a complimentary stat.
To be fair, I didn't analyze every drive so much as just lump it into the pool of data.
I mean, you're not wrong, but I addressed the obviousness of it toward the end. Though your statement doesn't really consider how turnovers are more prevalent on shorter drives when forcing them is one of the premises of the "bend but don't break" defense.
Honestly, I don't know that is the best method either but it seems straight forward enough and logical enough to me. It is important to remember it isn't factoring in yards gained on the drives. That would be another way to look at it...change out plays on the x-axis for % of available yards gained or to where the drives reached on the field.
The goal of every defense is to stop the offense as soon as they can. Bend but don't break just says that we aren't going to give up the big play doing that. They don't want the defense to be out there for 12 plays. So this analysis says nothing about bend but don't break. It just says more plays = a higher chance of scoring which is common sense stuff.
I'm not sure I agree that number of plays is the way to identify bend-don't-break defenses. Isn't it more about no big plays? If an opponent scores a touchdown on a play >= 15 yards, then the defense broke. I don't care if that was the first play of a drive or if it was the 20th. Similarly, if the defense gives up a 20 yard pass play from the 34 to the 4, it broke. Well, I guess if they keep the opponent from scoring then they didn't fully break, huh?
I think of "giving up lots of yards without giving up touchdowns" as a working definition of "bend-don't-break". (Giving up yards between the 20s.) Or at least I'd say it's better than simple "number of plays". Perhaps any drive that *starts* inside the 50 (40? 30?) should be thrown out of your statistical analysis.
Furthermore, this is really about *comparing* a bend-don't-break philosophy against the alternative(s) (which is -- what -- aggressive risk taking?). Drives that are killed primarily by big defensive plays (turnover, sack, tackle for loss -- aggressive defense) compared to drives that seem to just sputter out (bend-don't-break).
Re-do all your stats after first splitting out the type of defense represented. But even then, when you're behind and the clock is ticking, bend-don't-break goes away. You can't say that ISU employed a bend-don't-break philosophy on every opponent's drive.
This is why I think football is nearly impossible to get meaningful stats from. Points-per-possession seems like a good offensive indicator (eliminates pace of play), but there's no adjustment for length of scoring drive. An offense that consistently starts it's drives from 75+ yards away from a touchdown because of a bad defense could still be a lot better than an offense with better points-per-possession because their drives consistently start 25 yards closer.
So that's a long winded way of saying I don't agree with your base premise that bend-don't-break can be thoroughly analyzed by number of plays in a drive.
Oh -- and finally, I think the fact that the ratio of field goals to touchdowns as the number of plays goes up lends support to bend-don't-break.
I think of "giving up lots of yards without giving up touchdowns" as a working definition of "bend-don't-break". (Giving up yards between the 20s.) Or at least I'd say it's better than simple "number of plays". Perhaps any drive that *starts* inside the 50 (40? 30?) should be thrown out of your statistical analysis.
Furthermore, this is really about *comparing* a bend-don't-break philosophy against the alternative(s) (which is -- what -- aggressive risk taking?). Drives that are killed primarily by big defensive plays (turnover, sack, tackle for loss -- aggressive defense) compared to drives that seem to just sputter out (bend-don't-break).
Re-do all your stats after first splitting out the type of defense represented. But even then, when you're behind and the clock is ticking, bend-don't-break goes away. You can't say that ISU employed a bend-don't-break philosophy on every opponent's drive.
This is why I think football is nearly impossible to get meaningful stats from. Points-per-possession seems like a good offensive indicator (eliminates pace of play), but there's no adjustment for length of scoring drive. An offense that consistently starts it's drives from 75+ yards away from a touchdown because of a bad defense could still be a lot better than an offense with better points-per-possession because their drives consistently start 25 yards closer.
So that's a long winded way of saying I don't agree with your base premise that bend-don't-break can be thoroughly analyzed by number of plays in a drive.
Oh -- and finally, I think the fact that the ratio of field goals to touchdowns as the number of plays goes up lends support to bend-don't-break.
Sure, but there is a cumulative effect not being accounted for. In other words, just because a drive was stopped after 4 plays without scoring DOESN'T mean it WASN'T from a bend-don't-break philosophy. By splitting the stats into specific ranges of # of plays, it is implying that short drives are NOT bend-don't-break, while long drives ARE bend-don't-break.
