There is actually a way to beat the Pats but I don't think any coach has thought of

[Ring 6]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

This has been shown to be clearly true by lots of statistical analysis. Teams punt much, much too often. Risk aversion is rampant in the NFL. Even before this scorched-earth season, BB was one of the few coaches who took the academic analyses seriously and started leaving the punter on the shelf. It's not just running up the score, it's sound football.

What Ive seen statistically 'proven' goes like this:
The average play in the NFL gains 5 yards (or 50% gain 5 or more, 65% 3 or more etc) so if you go for a 4th and 5 its a 50% chance. Getting 50% of them, and keeping the ball, rather than gaining the average field position of a punt, adds up to slightly more points in your favor over the long haul.

Obviously, this is significantly flawed. 4th and 5 is not like the 'normal play'.

 

[PatsFaninAZ]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

What Ive seen statistically 'proven' goes like this:
The average play in the NFL gains 5 yards (or 50% gain 5 or more, 65% 3 or more etc) so if you go for a 4th and 5 its a 50% chance. Getting 50% of them, and keeping the ball, rather than gaining the average field position of a punt, adds up to slightly more points in your favor over the long haul.

Obviously, this is significantly flawed. 4th and 5 is not like the 'normal play'.

Right. Both third and fourth downs are different statistically from 1st and 2d, for a lot of reasons. Typically, the focus of the offense the mindset of the defense is much more about defending the line of gain. Also, most stats you see on 4th down conversions are completely stilted -- there are so many very short yardage plays, which is why the team goes for it in the first place.

My problem with all the statitical analyses I've seen of going for it on fourth down look at the reward side of the equation, but not the risk side (other than nonconversion risks.) (For example, teams turn the ball over on offensive plays -- as much as 1 of every 12 to 20 plays -- than on field goals.) Field position matters. Distance matters. Score, situation, and time outs matter.

There surely are some situations where teams do not go for it on a regular basis where they should. The most common is when down by 10 points late in the game, with manageable fourth down yardage, inside field goal range. The Titans were the most recent team, on Monday night, to kick a field goal in this situation.

It's rarely the correct play, unless your defense is playing lights out or you're at home in a loud stadium that has forced false starts and other mistakes. Since the kick off was moved to the 30 yardline, the ability of a decent offensive team to get into field goal range against a tired defense is, in my opinion, usually far greater than the chance you'll be stopped on fourth down in the above scenario.

We almost had this situation against the Colts. Welker converted the third and goal for a TD down by 10, but if he had not, the correct play there, in my opinion, is to go for it on fourth down. I'm not sure whether Belichick would have done so. The conventional wisdom always is "to make it a one score game." When you start taking into account the possibility that you could miss the field goal anyway, and the difference in field position between turning it over on the five or making the field goal and kicking from your 30, I think it's rarely the right play percentagewise.

Here's another one: You're down by 3 with 2:00 to play, say at home. You get to the 30 yard line and have a 4th and 1. Your kicker is about 80 percent on 47 yarders. Kick?

If the question here is what gives you the higher percentage of winning the game, it's a complicated question and depends on many factors. But I think the answer is "go for it" much more than one would think, and nobody EVER goes for it in that situation.

 

[Ring 6]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

Someone (Madden maybe?) said that what the Patriots are doing to other teams is making them change what they do before they even get off the bus. They don't stick with the gameplans, strategies, coverages, play calls that have worked for them, because they aren't confident they will work against the Pats.
While, I think thats true, I think we would be destroying teams even worse if they tried to do what they do against everyone else vs us.

If I was coaching against the Pats, my approach would be:
-I have 2 outcomes win or lose.
-Losing by 70 is no worse than losing by 1
-Winning by 1 is no worse than winning by 70
-If my team goes out and does what it does, I have very little chance of winning
-I must do the unexpected.
-I would open the playbook on offense. I would break out everything that contradicts what I have done. I'd run the option. Id reverse off it, throw off it. I'd have eveery trick play in the world in my gameplan. I would play WRs at FB to get a mismatch in coverage. I would go for 4th downs whenever reasonable. I'd fake FGs, run reverses on kick returns, onside kick, etc.
-Defensively, my only goal would be confunsion. I would play coverages that are uncommon in the NFL and/or uncommon to my system. I'd have a defender on each pass play 'play possum' showing coverage on one receiver, then breaking off to undercut another one, to create an Int by doing something the QB never expected. I would disguise every coverage making Brady read one thing, but stem into something else. I would mix up the rush, coming from different places. Early in the game I would overload one side with the pass rush. If I line up a DT and DE on the weak T and G, then blitz the MLB and OLB to that side, there are 2 guys for one back to pick up. This would ideally force the Pats to run the ball. My chances are better if they run that if they throw. Of course, once I have confused them with a pass defending style Ive never used before, I come up and try to stop the run (one zero or negative play on a drive and I get down and distance and go back to wacky pass d schemes).
-Basically, I approach it as I can't win with my system. So, I come out and get extremely risky, and break every trend they are anticipating, and hop[e to trick them.

 

[makewayhomer]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

What Ive seen statistically 'proven' goes like this:
The average play in the NFL gains 5 yards (or 50% gain 5 or more, 65% 3 or more etc) so if you go for a 4th and 5 its a 50% chance. Getting 50% of them, and keeping the ball, rather than gaining the average field position of a punt, adds up to slightly more points in your favor over the long haul.

Obviously, this is significantly flawed. 4th and 5 is not like the 'normal play'.

that sounds like a dumb TMQ analysis.

sound analysis done in The Hidden Game of Football did something like the above with actual realtistic data, and did find that teams should be going for it on 4th and short much more often than they do.

here is a sound study

subsequent analysis revealed that field position didn't matter as much as previously thought. turnovers help or hurt you nearly the same amount no matter where they occur

visually, see here

and read here

 

[LiveShot]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

That is exactly what John Madden told a Philadelphia audience yesterday.

 

[PatsFanSince74]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

as others have observed above, statistically, going for it on fourth and short, no matter where you are on the field, leaves a team better off OVER TIME (one of the first contrarian points concluded by Bob Carroll in 'The Hidden Game of Football' years ago).

the problem is that an NFL coach who goes for it on fourth and two on his own 25 and then loses the game by the touchdown the opponent scores as a result on that particular sunday will be skewered in the media. so, coaches choose to punt.

 

[RayClay]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

So you make a 4th down on your own thirty. If you don't make the next first down in three, you hand the ball over in your own territory.

Are there studies with success rates by giving the other team good field position every possession?

Also, teams rarely go for it on 4th in their own territory, so this must be a small sample size and all the results are therefore invalid.

 

[makewayhomer]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

so this must be a small sample size and all the results are therefore invalid.

oh the irony

but to speak to your point, the sample size is actually large. the trick is that you don't measure just the times a team went on 4th down from it's own 30, but you measure all the times a team turned the ball over AT ALL on their own 30. when you do this, you can measure the value of field position and "how bad" a turnover is at any point. this is the graph I showed above.

conventional wisdom says that there is nothing worse than turning the ball over at your own 1 yard line, but this is actually just as bad as turning the ball over on your opponents 1 yard line. ditto for your 20 vs their 20, etc.

 

[RayClay]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

oh the irony

It's a joke homer.

I'm even goingto go back and insert the smilie for the sarcasm impaired.

Adding "totally giving up field position" to the other problems a team faces is not going to help.

Having good yardage on the first three and playing defense might.

 

[RayClay]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

oh the irony

but to speak to your point, the sample size is actually large. the trick is that you don't measure just the times a team went on 4th down from it's own 30, but you measure all the times a team turned the ball over AT ALL on their own 30. when you do this, you can measure the value of field position and "how bad" a turnover is at any point. this is the graph I showed above.

conventional wisdom says that there is nothing worse than turning the ball over at your own 1 yard line, but this is actually just as bad as turning the ball over on your opponents 1 yard line. ditto for your 20 vs their 20, etc.

So it's OK to turn the ball over on your own 1 yard line. Who knew?

 

[makewayhomer]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

So it's OK to turn the ball over on your own 1 yard line. Who knew?

no. read my post again - it's equally as bad on your 1 yard line as theirs. the application of this is basically that if you are going to go on 4th and 2 on your opponents 40, you should go on your 40 as well.

 

[RayClay]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

What Ive seen statistically 'proven' goes like this:
The average play in the NFL gains 5 yards (or 50% gain 5 or more, 65% 3 or more etc) so if you go for a 4th and 5 its a 50% chance. Getting 50% of them, and keeping the ball, rather than gaining the average field position of a punt, adds up to slightly more points in your favor over the long haul.

Obviously, this is significantly flawed. 4th and 5 is not like the 'normal play'.

Also, "statistically" you've gained 15 yards every time you needed 10, so why would you ever have fourth down?

Over time, you're averaging a first down every two downs.

It's amazing how statistics can be "interpreteted"

 

[makewayhomer]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

Also, "statistically" you've gained 15 yards every time you needed 10, so why would you ever have fourth down?

Over time, you're averaging a first down every two downs.

It's amazing how statistics can be "interpreteted"

well, yeah. the gigantic problem with whatever Andy references is that it just takes some basic average for all plays. you need to consider what the average play gains in THAT situation. ie the average gain on 3rd and 2 is much shorter than the average gain on 1st and 10, so it's silly to include 1st and 10 averages into your 3rd and 2 calculations

 

[Ring 6]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

that sounds like a dumb TMQ analysis.

sound analysis done in The Hidden Game of Football did something like the above with actual realtistic data, and did find that teams should be going for it on 4th and short much more often than they do.

here is a sound study

subsequent analysis revealed that field position didn't matter as much as previously thought. turnovers help or hurt you nearly the same amount no matter where they occur

visually, see here

and read here

I dont think you understand your own reference.
The curve HAS TO look like that. The value of a turnover on your own x yardline or the opponents x yardline HAS TO BE THE SAME, because of the method used.

It is NOT the likelihood of the turnover leading to points.
The study shows the POINT VALUE of the average drive starting at that field position.
The graph shows the NET VALUE.
In other words, a turnover on the 20 has a POSITIVE VALUE of the average points scored on drives starting from the 20, PLUS a NEGATIVE VALUE for points lost from the drive NOT starting on the 20.

Let me illiustrate. I'll use the 1s as the example.


Teams taking possession on the +1 average 5.8 points on those drives.
Teams starting on the -1 average 1.3 points on those drives.
Pats turn the ball over on their own 1 yard line.
The VALUE of that turnover (to the opponent) is 5.8 points gained and 1.3 points PREVENTED, for 7.1 points (which it this chart is then halved and would show as 3.55) GAINED BY THE TEAM GETTING THE TURNOVER.
The Pats turn the ball over on the other teams 1.
They lose 5.8 points, and 'give up' 1.3 on the subsequent drive.
The value of that turnover is EXACTLY THE SAME, because it has to be.

This study accounts for BOTH points lost by turning it over, and points gained from the subsequent drive after getting the turnover.

In essense, it is entirely useless. All it has proven is that AT ANY POINT ON THE FIELD there is a statistical value (in terms of points scored) to having the football, and a (negative) statistical value to not having the football.
The value on the +20, -20, or +33, -33 HAVE TO BE EQUAL so it therefore makes it look like going for 4th and down on your own 1 is the same as going for 4th down on the opponents 1, and nothing could be further from the truth.

In a tie game, you are on the opponents 1. Your choice is to risk 3 points to get 7 or 0, knowing that if you get zero, they are less likely to score.
At your own 1, you theoretically reduce their chances of scoring if you make the first down, but give them almost a guaranteed score if you do not.
Are you really telling me you would make the same decision in either case, in the 4th quarter of a tied game?

 

[RayClay]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

well, yeah. the gigantic problem with whatever Andy references is that it just takes some basic average for all plays. you need to consider what the average play gains in THAT situation. ie the average gain on 3rd and 2 is much shorter than the average gain on 1st and 10, so it's silly to include 1st and 10 averages into your 3rd and 2 calculations

It's no different than other statistical interpretations. Like comparing plays on your own one to those on the opponents one then making inferences about plays on each others forty. There are a lot more unaccountable variables there than in my admittedly spurious example.

 

[makewayhomer]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

In a tie game, you are on the opponents 1. Your choice is to risk 3 points to get 7 or 0, knowing that if you get zero, they are less likely to score.
At your own 1, you theoretically reduce their chances of scoring if you make the first down, but give them almost a guaranteed score if you do not.
Are you really telling me you would make the same decision in either case, in the 4th quarter of a tied game?

No. obviously with 1 second to go in a tie game the decisions you make need to be different than in a blowout in the 2nd quarter - the situation matters, and clearly a human element comes into play here. how the game has been going, how your line has been playing, etc. all of this is very important.

but the study provides a good reason why coaches should at least consider going on 4th much more often than they do (except maybe BB, he leads the league )

 

[Ring 6]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

No. obviously with 1 second to go in a tie game the decisions you make need to be different than in a blowout in the 2nd quarter - the situation matters, and clearly a human element comes into play here. how the game has been going, how your line has been playing, etc. all of this is very important.

but the study provides a good reason why coaches should at least consider going on 4th much more often than they do (except maybe BB, he leads the league )

I think the study is fatally flawed for the reasons I listed.
What it is doing though is factoring in that if you keep the ball, you retain a chance to score. That chance typically comes close to equalling the difference of how much you reduce the other teams chance to score by punting.

Example:
You are on your own 20 yard line. If you convert a first down, you will on average score, say 2.2 points on all drives from that spot.
If you turn the ball over at the 20, the other team on average will score 5.5 points.
If you punt, on average, you will gain 30 yards, and the other team will, on average score 3.3 points on drives starting at midfield.
Therefore its a wash. You try to get your 2.2 points at the risk of giving the other team 5.5 instead of 3.3.

Its a circular argument, and it will always work out that way, becuase it isn't based on going for 4th downs, its based on POSSESSION, and it happens to work out that the value of POSSESSION and of the average gain in field position from a punt are roughly equal.

It has little value when applied to real football. Football isn't about being in a position where on average you will score 2.2 points, or putting the other team in position where on average it will score 2.2 more.

If going for 4th downs was a smart move often, (and especially since this model is based on turnovers) how do you account for the fact that the team with turnover differential advantage wins 85% of the time?

 

[Ring 6]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

No. obviously with 1 second to go in a tie game the decisions you make need to be different than in a blowout in the 2nd quarter - the situation matters, and clearly a human element comes into play here. how the game has been going, how your line has been playing, etc. all of this is very important.

but the study provides a good reason why coaches should at least consider going on 4th much more often than they do (except maybe BB, he leads the league )

I didnt say with 1 second ot go.
Your model says on the first play of the 4th quarter its a toss up.
It says AT ANY POINT IN THE GAME you should flip a coin on 4th and 3 at your own 8 yard line. Clearly that is wrong.

 

[makewayhomer]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

Football isn't about being in a position where on average you will score 2.2 points, or putting the other team in position where on average it will score 2.2 more.

it actually is. until you understand this, it's useless to argue this with you

 

[makewayhomer]

Re: There is actually a way to beat the Pats but I don't think any coach has thought

It says AT ANY POINT IN THE GAME you should flip a coin on 4th and 3 at your own 8 yard line. Clearly that is wrong.

where does it say this?