Give-Away's/Take-Away's, how important?

[PATSNUTme]

All stats are not the same. Probably one of the most telling stat's in what a team's W/L record is- G/A's-T/A's.

Here are a list of teams that have a +3 or better in G/A - T/A's and their records

Dallas +7 4-0
Indy +7 4-0
Seattle +5 3-1
Tampa +4 3-1
Green Bay +3 4-0
Patriots +3 4-0
Steelers +3 3-1

Now here's a list of teams with -3 or worst

Saints -8 0-3
Bears -7 1-3
Miami -6 0-4
Rams -3 0-4
Ravens -3 2-2

This list comes from Patriots Football Weekly. It's one of the stat's that I check every week.

So when people ask if the Patriots can lose to Cleveland, the answer is yes if we turn the ball over more than they do.

 

[PATSNUTme]

OK, probably not that important.

 

[Fencer]

Overall, they're clearly important. Some are of course more important than others. E.g., an INT for no return 30 yards downfield on third down is only slightly better than a deflection. A forced fumble on first down with a TD runback is a whole different matter.

 

[363839]

OK, probably not that important.

I remember a time when this stat WAS important for the Patriots.
I mean now, we don't really need a huge turnover ratio to pound teams.
This team is so strong it could probably over come even a negative ratio.

 

[patsfan13]

If the Pats get a +1 to +2 TO ratio per game they may not lose, one of the few less that outstanding stats so far.

 

[MrClutch]

It really depends on the situation of the turnover. I mean, if you're up by like 28 points with 2 minutes to go and intercept a pass, it really doesn't mean much.

 

[Fencer]

It really depends on the situation of the turnover. I mean, if you're up by like 28 points with 2 minutes to go and intercept a pass, it really doesn't mean much.

That would be true for most other stats as well.

One thing I've been harping on -- and nobody has much cared about -- is that individual unit stats should be viewed in the context of field position. So TOs that affect field position are real modifiers to unit performance. TOs which are stops deep in enemy territory -- well, they probably stopped points, which is great, but to look at them next to the points-allowed figures is to double-count.

Similarly, if the offense turns over the ball deep in enemy territory, that's no worse than a drive that ended in a punt. But if they turn it over in their own territory, that's terrible.

 

[PATSNUTme]

It is just a coincidence that he teams with the best records have +3 or more? And the teams with the worst records have -3 or less?

Last year the Bears were noted for creating turnovers and that overcame lousy QB play.

 

[Oswlek]

That would be true for most other stats as well.

One thing I've been harping on -- and nobody has much cared about -- is that individual unit stats should be viewed in the context of field position. So TOs that affect field position are real modifiers to unit performance. TOs which are stops deep in enemy territory -- well, they probably stopped points, which is great, but to look at them next to the points-allowed figures is to double-count.

Similarly, if the offense turns over the ball deep in enemy territory, that's no worse than a drive that ended in a punt. But if they turn it over in their own territory, that's terrible.

The only way I agree with this is if it is 3rd and long and an int is thrown on a long pass attempt. Otherwise, the loss of a scoring opportunity is just as bad as the provided scoring opportunity.

 

[FlyingElvis75]

It is just a coincidence that he teams with the best records have +3 or more? And the teams with the worst records have -3 or less?

It could be that creating turnovers gives you a better chance of winning (turnovers cause wins). It could also be that winning makes you take fewer risks with the ball, and makes your opponent take more risks (winning causes turnovers). Or they could be completely unrelated, and both are caused by something else (good play causes both winning and turnovers), i.e. a spurious correlation. My money's on the latter, if there even is a correlation between turnovers and wins.

In this case, I'm not sure the correlation is even "established." Dallas, Buffalo, and Baltimore all had a +0 differential in their last wins; the Bears actually had a -1 in their only win. Just a few counterexamples...

 

[PatsFaninAZ]

It is just a coincidence that he teams with the best records have +3 or more? And the teams with the worst records have -3 or less?

Last year the Bears were noted for creating turnovers and that overcame lousy QB play.

Its a hugely important stat. I'm not sure there are many bigger predictors of result. As an experiment, I looked last year to see how often teams won despite being negative in this stat for the week. It was rare to see any week in which more than one team a week overcame a negative giveaway/takeaway to win the game. Overall, it was like 80 percent accurate at predicting the outcome of the game and higher on the road. And sometimes, when you dig a bit deeper in those games, you learn that actually the teams weren't negative in takeways -- it's just that some things don't count for turnovers that basically are. Like a blocked punt that dribbles a few yards across the line of scrimmage. That doesn't count as a takeaway -- it counts as a 3-yard punt. Or a turnover on downs. If you count things like these as turnovers, and disregard the over-40-yard-from line of scrimmage interceptions (which are basically punts) you almost never see a team overcome a true turnover differential and lose.

The Patriots, though, seem to be a bit resistant to it. I think they had 4 games where they won despite a negative in this category. Steelers too, for same reason. Not sure why. I think it has to do with coaching philosophy and risk taking, making the situation of the turnover less costly.

Note that there's also something of a chicken-or-egg quality to this stat. Teams that are losing tend to turn the ball over more as they press.

 

[FlyingElvis75]

And sometimes, when you dig a bit deeper in those games, you learn that actually the teams weren't negative in takeways -- it's just that some things don't count for turnovers that basically are. Like a blocked punt that dribbles a few yards across the line of scrimmage. That doesn't count as a takeaway -- it counts as a 3-yard punt. Or a turnover on downs. If you count things like these as turnovers, and disregard the over-40-yard-from line of scrimmage interceptions (which are basically punts) you almost never see a team overcome a true turnover differential and lose.

The Patriots, though, seem to be a bit resistant to it. I think they had 4 games where they won despite a negative in this category. Steelers too, for same reason. Not sure why. I think it has to do with coaching philosophy and risk taking, making the situation of the turnover less costly.

That would seem to support the idea that good play creates turnovers AND wins, not that turnovers create wins. Correlation <> causation. Turnover differential may be an indicator of success, but I haven't seen evidence that it's a predictor.

 

[Oswlek]

I'm going to go out on a limb and say that tunrover differential is pretty improtant.

 

[FlyingElvis75]

I'm going to go out on a limb and say that tunrover differential is pretty improtant.

Not as important as points differential!