New stat for punters?

[QuikSand]

Tom Rouen and his dreadful Superbowl performance got me thinking a bit this morning about punters and ways to measure their performance. FOF, like the NFL, generally uses a couple of stats:

-gross yardage
-number of punts inside the 20 yard line

...and then derivatives of the same, like net yardage and % inside the 20.

(edit -- FOF doesn't even calculate net punting average, though the NFL does)


To me, punting stats have always left me wanting -- I have read (I think from Dr. Z of Sports Ilustrated) high praise for the "hang time" stat, but I always wonder why that shouldn't just work itself out in something like net yardage, if all else is equal with the coverage cast.

Regardless -- what I'm thinking about at the moment is measuring a punter's ability to actually effectively drop the other team into a tight spot. So, I have another idea for a stat:

Within Twenty Average

The WTA would be the average line of scrimmage gained by the receiving team any time that line is on or within its 20 yard line. This stat should be pretty easy to calculate, I'd think.

Ideally, what you're talking about are all the punting situations where the punter has the ability to pin the other team back... and in those situations you consider a touchback to be, essentially, a failure. On the other hand, the lower the resulting yard line, the more effective the pin-back was. (Yes, this will mean that the coverage team has some meaningful say in this, though this too ought to wash out as I don't truly believe this is a special football skill... and I atribute more o the result to the punter, who puts the ball into position to be downed)

So, here are two punters who each have 5 opportunities to "pin back" the other team:

Punter A: TB, 5, TB, 18, TB
Punter B: 8, 5, 15, TB, 1

Fairly extreme differences, admittedly, but here we'd have a way to assess how effective each guy was in the job:

Punter A's WTA: 14.6
Punter B's WTA: 9.8

The lower number is clearly superior, and in this case it's a pretty easy spread.



Strengths: I like that this, conceivably, takes some of the bias out of the current measures of "Inside the 20" frequency calculations. When you look at the whole range of punts, including those from the punter's own 20 yard line, and talk about "% inside the 20" you are adding a variable that doesn't belong -- and is probably out of the punter's control. Here, my goal is to focus just on the punts where the punter has the ability to get inside the 20... evidenced by him actually doing so.

I also like that this is a fairly easily calculated number, which ought to be pretty directly comparable from punter to punter.


Weaknesses: I suppose that a very crappy punter who punts from the 40 and shanks a 15-yard effort to the 25 will get a free pass here -- though this isn't meant to be the sibgle rating to best assess punter's overall skill. That 15-yard punt ought to hurt him in other ways.

Another potential weakness -- theoretically, a punter will be penalized for an extraordinary punt that happens to go to the endzone for a touchback. A 70-yard touchback is a play that teams woudl be delighted with, of course, but would be unflattering by this metric. Again, this is meant to go beside the other punters' stats, so he would be appropriately flattered in his yardage averages.

I also suppose that a punt to inside the 20 that gets returned to otside the 20 is up for grabs -- I'm inclined not to include it here (as my main focus is strategic punting) but that might skew things somehow (I haven't thought through what the effect might be one way or the other).

- - - - -


Yours to chew on...

[Bee]

I think the better solution is just to remove punts from the game. Force every team to go for it on 4th down every time. Frees up a roster spot for a real football player and would add more scoring. Everyone wins!

:D

[MIJB#19]

I've been thinking about this too recently, really.

To me, there are two types of punts, which should be measured differently:
- Punts to get the ball as far away as possible (from around the own 30-yard line)
- Punts to pin the opponent back (from around the opponents' 40-yard line)

The only thing I could come up with so far is penalizing punters for the 20-yards loss by the touchback rule. Although, then you'll be unfair to the cases where a touchback and no return is better than punting it to the Opp15 and see the returner take it to the Opp25. Net yardage usually is punt yards minus punt return yardage, right? The 20-yard loss of the punt could be factored in to reward punters who can put it in the 20.

Other than that, I really like your idea, QS, it's bascially what I was looking for to measure my punters in FOF, but couldn't come up with. It should work well in the NFL I think.

[Bonegavel]

How about a range of stats for this.

first range is midfield to opponents 1 yard line (-1 meaning opponents 1 yard line).

50 to -1
40 to 49
30 to 39
1 to 29

you could then assign your number system to each category depending on where the line of scrimmage.

I thinks it is safe to say that a touchback, while good in your example, is still not preferred to the ball going out of bounds on the 1 yard line every punt.

Is punting more situational than other aspects of the game? Sometimes the punter may want to coffin corner the ball because the returner is a Dante Hall. Or maybe the returner has a tendancy to bobble high kicks. Teams seem to be getting pretty good at downing the ball at the one by diving into the endzone and batting it out.

Punting is a hard position to pin down. Field goal kickers are simple; they make the FG or they miss the FG. Is there a way to categorize punts to determine if the punter can perform the most advantageous result? E.g.s of bad punts: a coffin corner attempt goes in for a TB; a CC punt is shanked out at the 30; E.g.s of good punts: punter in his endzone kicks it to the opponents 30 yard line; punter on his own 20 kicks it out of bounds at the opponents 25.

Those last 2 examples are good punts but not without knowing why.

just kicking in some thoughts here.

[QuikSand]

I thinks it is safe to say that a touchback, while good in your example, is still not preferred to the ball going out of bounds on the 1 yard line every punt.

I don't follow this argument.

What I'm suggesting above is that a touchback counts as a 20, and out at the 1 counts as a 1. The lower your WTA, the better your performance, so the low number helps. Lots of touchbacks is what you don't want from a punter.

[QuikSand]

On further reading, I don't think you and I are on the same page here, Bonegavel. I'm not sure how your criticisms apply to the model I'm suggesting above.

[Ben E Lou]

The only thing I could come up with so far is penalizing punters for the 20-yards loss by the touchback rule.I'd like to see an "adjusted net punting yards" number, with the 20-yard loss for touchback factored in, along with return yardage, because ultimately, the job of the punter (and the coverage unit) is to impact field position. Let's say in Quik's example, that all 10 punts were attempted from the 50. You'd have:

Punter A: TB, 5, TB, 18, TB (30+45+30+32+30)/5=33.4
Punter B: 8, 5, 15, TB, 1 (42+45+35+30+49)/5=40.2

This is a simpler calculation and takes into account more situations. The guy who got the 70-yard touchback still gets 50 net yards. The guy who shanks the punt and advoids the touchback still gets short net yardage. The guy who has poor hang time gets docked big-time for the three punts returned for TD's over the course of the season. Y'all get the idea.

[QuikSand]

I thought net punting yards already included the 20 yards from a touchback... but if not, then I certainly agree with you, Ben.

However, I think that stat still underestimates the value of a punter's performance from within TB range. Those yards between the 20 and the goal line are, in my view, more meaningful than the yards between, say, the 40 and the 20. And there is a separable skill involved with the control punting, which is what i seek to find with this metric.

[Ben E Lou]

I thought net punting yards already included the 20 yards from a touchback...I thought it didn't, but to be honest, I've never checked it. Either way, it's a pretty big assumption that I'd like to check on. Running a test now...

[Ben E Lou]

Ummm....I'm not seeing where FOF even TRACKS net punting yardage. Am I missing something?

[QuikSand]

Ummm....I'm not seeing where FOF even TRACKS net punting yardage. Am I missing something?

No FOF doesn't... I was refering to the NFL's stat for "net punting average" which I thought already included what you mentioned above. Sorry, I wasn't clear.

[Ben E Lou]

No FOF doesn't... I was refering to the NFL's stat for "net punting average" which I thought already included what you mentioned above. Sorry, I wasn't clear.You were correct, then.

<table border="0" cellpadding="0" cellspacing="0" width="100%"> <tbody><tr class="cnnSmall cnnNoBold cnnHeader"><td class="cnnLeft">Date</td> <td>Opp</td> <td>Punts</td> <td>Yds</td> <td>Avg</td> <td>Lg</td> <td>In20</td> <td>In10</td> <td>FC</td> <td>TB</td> <td>Blk</td> <td>Ret</td> <td>Yds</td> <td>TD</td> <td>Net</td> </tr> <tr class="cnnData1 cnnNoWrap"> <td>Jan. 14</td> <td>WAS</td> <td>5</td> <td>212</td> <td>42.4</td> <td>53</td> <td>4</td> <td>4</td> <td>1</td> <td>0</td> <td>0</td> <td>1</td> <td>3</td> <td>0</td> <td>41.8</td> </tr> <tr class="cnnData2 cnnNoWrap"> <td>Jan. 22</td> <td>CAR</td> <td>4</td> <td>169</td> <td>42.3</td> <td>45</td> <td>1</td> <td>1</td> <td>0</td> <td>2</td> <td>0</td> <td>1</td> <td>59</td> <td>0</td> <td>17.5</td> </tr> <tr class="cnnData1 cnnNoWrap"> <td>Feb. 05</td> <td>PIT</td> <td>6</td> <td>301</td> <td>50.2</td> <td>57</td> <td>0</td> <td>0</td> <td>0</td> <td>4</td> <td>0</td> <td>2</td> <td>32</td> <td>0</td> <td>31.5</td></tr></tbody> </table>

[MIJB#19]

You just had to beat me by posting how net punting yardage works, didn't you, SD? ;)

[Samdari]

I think the better solution is just to remove punts from the game. Force every team to go for it on 4th down every time. Frees up a roster spot for a real football player and would add more scoring. Everyone wins!

:D

I'm sold.

I think that another great rule would be to require players to play 5 (10? 15?) plays from scrimmage before being eligible to touch the ball with their foot on a scrimmage play(let's reduce those pesky FGs as well).

Let guys punt or attempt FGs, but let football players do it.

[DanGarion]

I think the better solution is just to remove punts from the game. Force every team to go for it on 4th down every time. Frees up a roster spot for a real football player and would add more scoring. Everyone wins!

:D
Ray Guy is going to kick your ass!

[Bonegavel]

I don't follow this argument.

What I'm suggesting above is that a touchback counts as a 20, and out at the 1 counts as a 1. The lower your WTA, the better your performance, so the low number helps. Lots of touchbacks is what you don't want from a punter.You have an example how a 70 yard TB would be delightful, which I do not doubt. (delightful meaning: good outcome, punter did his job, attaboy, etc).

What I want to see from a punter stat POV is not necessarily raw meaningless numbers but I want to see if he did the right thing in the right situation.

[Bonegavel]

On further reading, I don't think you and I are on the same page here, Bonegavel. I'm not sure how your criticisms apply to the model I'm suggesting above.Not a criticism, just attempting (not well, apparently) to add some more food for thought.

[bhlloy]

Quik - great idea and definitely a stat I would want to see in play.

Couple of thoughts - assuming a good placement punt is 45-50 yards tops in most situations, couldn't you just exclude all punts from say midfield (maybe the punters 45 yard line?) back from your WTA stat? Then the guy who hits a 60-70 boomer into the end zone doesn't have it against him. Of course then if he gets it out at the 1 yard line it doesn't count positively either, but I'd say that is a different skill than you are trying to measure (basically what MIJB said). Of course, if you are selectively choosing which punts to measure, maybe it's not so easily measurable anymore :)

To me, the biggest problem is what to do with the return yards. Another example - a guy who hits a horrible line drive punt from the 45 that bounces on the 20, rolls right into the returners arms at the 10 setting up a good return before the coverage team can get downfield. Technically that punt would be fielded at the 10 and would be a "good" WTA number, but if the returner returns it to the 30 it's worse than a touchback. I'm afraid I have no idea what to do with this one. I'm tempted to say that it's only the job of the punter to get the ball inside the 20 and then it's down to the coverage team, but in the example above that might not be the case.

EDIT - for my bad maths

[bhlloy]

DOLA - I just re-read your post and I think I agree with what you were saying about excluding punts fielded inside the 20 that are returned. I don't think the numbers will be so great that it will skew the stats much.

Another approach would be to say that in the situation we are looking at, the punters job is to avoid a return at all costs. Either get it out of bounds, or hang it up there so there is no return. I can definitely see a case for penalising each returned punt the same way you would penalise a touchback (+20), even if the return doesn't make it to the twenty yard line. That way it wouldn't be too hard to measure. A weakness of that might be that it doesn't legislate for a dumb punt returner who shouldn't field a high punt on his five yard line, but does anyway and is wrapped up for a gain of two yards.

[QuikSand]

Yes, on a little more reflection... I guess you could add to this by trying to sort things out a bit more based on the starting point. It detracts from the simplicity a bit, but it would reduce the noise you'd get from good long efforts that resulted in touchbacks.

I'm not sure how you work that into a formula quite yet... but trying to better recognize the long punt that just happens to get into the WTA zone would make the rating more fair, I reckon.

[Huckleberry]

Off the wall idea:

Percentage of Available Yards Gained

I ran this stat for Southern Cal's and Texas' offense and defense in previewing the Rose Bowl. Basically the percentage of available yards that the offense gains. If they start at the 20 after a touchback and score a touchdown then they get a 100% grade for that drive. Limitations for this stat are that drives that start near the opponent's goal line are highly volatile because of the smaller denominators. Losing 5 yards when you start on their 20 results in a -25% score for that drive.

Anyway, I wouldn't list it as a superstat, just one more piece of data.

For a punter perhaps his percentage of available yards gained on each punt could be tracked (maximum of 55 or 60 available yards on each punt). For example, QS' example above:

Punter A: TB, 5, TB, 18, TB
Punter B: 8, 5, 15, TB, 1

Let's say that all of those punt plays are snapped from the opponent's 40 yard line.

Punter A: 50%, 87.5%, 50%, 55%, 50% - Average 58.5%
Punter B: 80%, 87.5%, 62.5%, 50%, 97.5% - Average 75.5%

[QuikSand]

Percentage of Available Yards Gained

I ran this stat for Southern Cal's and Texas' offense and defense in previewing the Rose Bowl. Basically the percentage of available yards that the offense gains. If they start at the 20 after a touchback and score a touchdown then they get a 100% grade for that drive. Limitations for this stat are that drives that start near the opponent's goal line are highly volatile because of the smaller denominators. Losing 5 yards when you start on their 20 results in a -25% score for that drive.

I'm a little put off by the notion that weighing each possession the same leads to some weird situations. If you start a drive off a turnover at the opponent's 10 yard line... there ends up being a HUGE difference (in this metric) between kicking an 18 yard FG and a 27 yard FG, when in practical effect there is basiclaly no difference there at all.

For a punter perhaps his percentage of available yards gained on each punt could be tracked (maximum of 55 or 60 available yards on each punt). For example, QS' example above:

Punter A: TB, 5, TB, 18, TB
Punter B: 8, 5, 15, TB, 1

Let's say that all of those punt plays are snapped from the opponent's 40 yard line.

Punter A: 50%, 87.5%, 50%, 55%, 50% - Average 58.5%
Punter B: 80%, 87.5%, 62.5%, 50%, 97.5% - Average 75.5%

I think this essentially just converts to the same calculation as my so-called WTA rating... just with a negative multiplier and a different null point. A perfect correlation, where PYG = 100 - (2.5 * WTA).

[Huckleberry]

I think this essentially just converts to the same calculation as my so-called WTA rating... just with a negative multiplier and a different null point. A perfect correlation, where PYG = 100 - (2.5 * WTA).
Agreed but this can be applied to any punt from any position so the ad hoc decisions of when a punter has a chance to pin them don't have to be made.

A 45-yard punt from your own 20 could be graded against a punt from their 40.

[JeffR]

A 45-yard punt from your own 20 could be graded against a punt from their 40.

So, using 60 as the maximum available, that 45-yarder from the 20 is a 75% punt; to get 75% from the 40 the punt would have to be downed at the 10. That seems to be in the right ballpark to me, assuming we're talking about net yards. And if we are, how would you handle punt returns past the original line of scrimmage? 0, or a negative percentage?