**SITE NEWS:**
We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all PFR content, so you can quickly and easily find the content you want.

Also, our existing PFR blog rss feed will be redirected to the new site's feed.

Pro-Football-Reference.com ยป Sports Reference

For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.

## Rackers > Cribbs?

This post will do for field goal kicking essentially what Chase's recent series of posts did for return games. It will also introduce some new data that's soon to hit the main part of p-f-r.

Let's start by looking at Tony Franklin 1979, Kevin Butler 1985, and Matt Stover 2006.

FGA FGM PCT ====================================== Matt Stover 2006 30 28 93.3% Kevin Butler 1985 37 31 83.8% Tony Franklin 1979 31 23 74.2%

I'm now going to make a case that Franklin's season was the best of the three. You never saw that coming, did you?

The obvious observation to make here is that field goal kicking percentages have been rising consistently for a long time. And that's right. Here are the league-wide field goal percentages since the posts were moved to the back of the end zone in 1974:

1974 0.606 1975 0.642 1976 0.595 1977 0.583 1978 0.631 1979 0.631 1980 0.634 1981 0.659 1982 0.689 1983 0.715 1984 0.717 1985 0.722 1986 0.686 1987 0.705 1988 0.717 1989 0.725 1990 0.744 1991 0.735 1992 0.726 1993 0.774 1994 0.790 1995 0.774 1996 0.801 1997 0.781 1998 0.796 1999 0.777 2000 0.796 2001 0.763 2002 0.776 2003 0.792 2004 0.808 2005 0.810 2006 0.814 2007 0.830

So we might compute as follows: in 1979, an average kicker would have made 19.6 out of 31 field goals. Franklin made 23, so he was about 3.4 field goals above average. Using similar computations on Stover and Butler, we get this:

FG above average ============================ Butler 4.30 Stover 3.57 Franklin 3.45

Franklin's still at the bottom, so you know there's another twist coming. This is where I show off the new data: field goals made and attempted by distance!

0--19 20--29 30--39 40--49 50+ ================================================================== K Butler 1985 | 2 2 | 13 13 | 14 13 | 6 3 | 2 0 | M Stover 2006 | 0 0 | 13 12 | 9 9 | 7 6 | 1 1 | T Franklin 1979 | 1 1 | 5 5 | 8 7 | 10 7 | 7 3 |

More than half of Franklin's attempts (17 of 31) came from 40 yards away or more. Only eight of Butler's 37 attempts and eight of Stover's 30 were from 40 yards or more. This is quite a difference. In fact, if Franklin had made one more of his 30--39 attempts, or if Butler had missed one more of his, we would have a fantastic Simpson's paradox example.

Anyway, we'll adjust for this difference. Essentially, we do the same thing as we did above --- compute field goals above league average --- but instead of doing it to the totals, we'll do it for each individual distance and add them up.

Here are Franklin's numbers. I'll explain afterward:

0--19 20--29 30--39 40--49 50+ ====================================================================================== T Franklin 1979 | 1 1 0.7 | 5 5 4.4 | 8 7 5.3 | 10 7 4.5 | 7 3 2.3 |

Take a look at the 50+ numbers: "7, 3, 2.3". That means that Franklin attempted 7, he made 3, and a 1979 league average kicker would have made 2.3. So Franklin gets credit for 0.7 field goals above average (FGAA). Likewise, he gets 2.5 FGAA for his 40--49 yard efforts, and so on. Once you add it up, you get that Franklin was about 5.7 field goals above average. Here are Stover's and Butler's numbers:

0--19 20--29 30--39 40--49 50+ FGAA ============================================================================================= T Franklin 1979 | 1 1 0.7 | 5 5 4.4 | 8 7 5.3 | 10 7 4.5 | 7 3 2.3 | 5.69 K Butler 1985 | 2 2 1.9 | 13 13 11.4 | 14 13 11.1 | 6 3 3.6 | 2 0 0.8 | 2.32 M Stover 2006 | 0 0 0.0 | 13 12 12.4 | 9 9 7.7 | 7 6 5.1 | 1 1 0.5 | 2.26

Let's now engage in some moderate craziness and see if we can translate these numbers into points or, better yet, wins. I'll admit right now that in what follows I'm going to make some assumptions that aren't completely without justification, but that might rightly be called sloppy. I'm just having fun here. Consider this post to be the back of a big envelope.

So how many points do 5.69 field goals above average translate to? 5.69 times 3? No, there's more to the story.

What's the difference between a made 50-yarder and a missed 50-yarder? (I should say, between a sterilized hypothetical average 50-yarder and a sterilized hypothetical average missed 50-yarder).

1. Three points.

2. About 15 yards of field position on the opposing team's next possession. If you miss the field goal, the other team takes over at the 40. If you make it, then you kick off, and I'll assume that the average starting field position after a sterilized hypothetical average kickoff is about the 25 yard line.

Using the rule that (somewhere around) 20 yards is worth a point, we see that one 50-yard field goal above average is worth almost 4 points above average. On the other hand, a missed 20-yard field goal costs you three points, but it actually helps you a little bit (around 5ish yards) in terms of field position on the opposing team's next possession.

Now, if we apply these assumptions to Franklin's, Stover's, and Butler's numbers, we can get an estimate for how many points each of those kickers added beyond what an average kicker would have.

PTs above avg ================================== T Franklin 1979 | 17.1 M Stover 2006 | 7.8 K Butler 1985 | 5.6

Now I do this for every kicker-season since 1974, and I express it in terms of "points above average per 16 scheduled games" (i.e. I'm pro-rating the short seasons: pre-1978, 1982, and 1987). Here are the top field goal kicking seasons of the period:

PTs above avg ================================== N Rackers 2005 | 25.4 F Steinfort 1980 | 23.3 M Moseley 1982 | 23.0 G Anderson 1998 | 22.0 M Andersen 1985 | 21.6 R Allegre 1983 | 20.7 M Vanderjagt 2003 | 20.3 P Stoyanovich 1997 | 20.1 D Biasucci 1987 | 20.0 C Blanchard 1996 | 19.0 N Lowery 1985 | 17.8 T Fritsch 1979 | 17.6 T Franklin 1979 | 17.1 M Moseley 1979 | 16.6 J Wilkins 2003 | 16.6 E Murray 1989 | 16.6 R Ruzek 1987 | 16.2 N Lowery 1982 | 15.7 M Andersen 1987 | 15.1 N Lowery 1990 | 15.1 A Del Greco 1995 | 15.0 N Lowery 1980 | 14.8 J Hanson 2003 | 14.8 M Andersen 1986 | 14.6 J Stenerud 1981 | 14.5 M Moseley 1977 | 14.4 N Lowery 1988 | 14.4 P McFadden 1984 | 14.4 N Johnson 1993 | 14.3 M Andersen 1992 | 14.2 J Nedney 2005 | 14.0 A Haji-Sheikh 1983 | 14.0 D Biasucci 1988 | 14.0 G Anderson 1987 | 13.6 R Longwell 2000 | 13.6 M Vanderjagt 1998 | 13.6 A Del Greco 1998 | 13.5 R Szaro 1976 | 13.5 M Andersen 1995 | 13.4 M Zendejas 1987 | 13.4 T Zendejas 1991 | 13.4 M Vanderjagt 2001 | 13.2 G Yepremian 1978 | 13.2 T Dempsey 1975 | 13.0 M Hollis 2000 | 12.8 M Stover 2000 | 12.8 N Lowery 1983 | 12.7 J Elam 2001 | 12.7

What are 25 points worth? Obviously it depends a lot on when you get the points, but 25 "average" points are worth quite a lot. That's like starting every game with a one-and-a-half point lead. That's not insignificant. I'd guess it's worth between 1 and 2 wins over the course of an average season.

So, was Neil Rackers' 2005 worth as much as Josh Cribbs' 2007?

In the above-linked post, my colleague Chase Stuart argues that Cribbs was about 600 return yards better than average. Using 20 yards per point, that's about 30 points, which is a little more than Rackers' 25. However, consider the following:

- Cribbs' 30 points were virtually all what you might call generic theoretical points, the kind of "points" that you accrue over time by starting lots of drives at the 40 instead of the 30. But kicker points seem somehow more real, because you only attempt field goals when it makes strategical sense to do so. A field goal attempt on the last play of the game when you're down by one is worth a lot more than 3 theoretical points. A field goal attempt on the last play of the game when you're down four is worth nothing. That's why you don't see any of the latter. In short, I think it's likely that a greater proportion of the "points" Cribbs added (or any other non-kicker's; I'm not knocking Cribbs) were wasted.
- As Chase pointed out in the previous post, many of the "points" we are attributing to Cribbs actually belong to the 10 guys blocking for him. The same is true of Rackers, of course, but I think to a much, much lesser extent.

Adding it all up, I do think that Arizona's 2005 field goal kicking unit was at least as valuable as Cleveland's 2007 return units, and that Rackers' 2005 was more valuable than Cribbs' 2007.

As Chase points out, though, the Cleveland return game is probably (<---- unfounded assertion as far as I know) more likely to regress next year than similarly valuable passing or running attacks or defensive units. The same is probably true of kickers, as Rackers himself demonstrated by posting a 2006 season that was roughly 5 points below average.

I had hoped to talk about Mark Moseley's 1982 MVP award in this post, but it looks like that will have to wait for a future post.

This entry was posted on Wednesday, June 18th, 2008 at 1:31 pm and is filed under General. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

A really important distinction between the kickers and Cribbs, though, is that kicker performance varies wildly from year to year. While Neal Rackers' 2005 was highly valuable, he's not likely to keep it up.

Return man performance also varies a lot from year to year, but not as much as field goal kicking performance.

Also, kickers accrue some value from above-average kickoffs. When all is said and done, a typical stud kicker might be better than a typical stud return man.

On the other hand, Hester and Cribbs are crazy, and may very well be the two most valuable special teams players ever.

Not sure how you'd say that returners are more consistent than FG kickers. On the surface, one would expect the opposite. FG kicking is uniquely dependent on a single person with a single talent level, while kick returning is a chaotic event dependent on 10 blockers and 11 coverage players who change from week to week.

And don't forget Rackers' above average kick off depth. Think how many points he prevented with his touchbacks and deep kick offs.

I'm guessing the situational nature of field goals is going to make this difficult. Many 50+ yard field goal attempts are for the win at the end of a game, or to close out the half, etc. So then yardage penalty/bonus is negated. So a high percentage there may be less valuable in terms of virtual points than the numbers would suggest. On the other hand, I imagine a greater percentage of 50+ yard kicks are for the tie or win, so maybe they're actually undervalued, I don't know.

Also, when extrapolating to get a 16 game average, wouldn't one expect some amount of regression towards the mean? Chipper may be hitting .400 now, but the odds of him being there at the end of the season are much smaller... This would particularly affect Moseley...

So I was thinking it might be possible to make a system based on the assumption that a) the coach always makes the best decision and b) the coach assumes the kicker is average. Then we could say the rewards of making it must outweigh the risk of missing it, taking their likelyhood of making the FG based on distance and year into account. But I haven't been able to figure out such a system. :-/ It seems one would have to come up with the value for a turnover and the values of going for it or punting at that distance too... Bleh!

I'm disappointed by this post -- not because it was poorly done, but because I was working on something similar. I was afraid that y'all would come out with something like this first. : )

That said, I'll second Yaguar's comment: I think that when it comes to kickers, judging single-season values might not be all that useful. After all, single-season field goal percentages vary considerably from year to year, so many of those "points added" (or "points subtracted") over (/under) average may just be due to variance, and not the kicker's innate abilities. This problem is probably magnified when you start adjusting for distance, since the sample size for each kicker is so tiny. So while this work might be interesting for judging how much of a factor kicking may have been in a teams' wins, I don't think it's the best for judging the "true" worth or ability of a given kicker. However, if you turned this into a career stat (something like points above average per season), you might get large enough sample sizes to make more meaningful judgments about the players themselves.

Finally a question: where on earth did you get the distance data? I could only find distances listed from the NFL beginning in 1991. (Ironically, I planned to send an email your way offering the data I'm compiling once I was finished; it looks like that's no longer necessary.)

Simpson's paradox creeps me out.

But, I ran the percentages on Tony Franklin, and I don't think it applies. Can you help me out?

If Franklin had made one more 30+ yarder, his percentages would be:

0-19, 1.000

20-29, 1.000

30-39, 1.000

40-49, .700

50+, .429

Overall, .774

Wouldn't Simpson's Paradox mean that his overall percentage be higher (or lower) than each of the component percentages?

Hey cathartic,

Good point. This post (and the Josh Cribbs one) is much more valuable in retrodictive terms than predictive one. Most precisely, these studies show which players had the most impressive seasons rather than which players are the strongest in ability.

Richie,

If Franklin had made one more from 30--39, then the paradox would have been:

Franklin had a better (or equal) percentage than Butler from every distance, but Butler had a higher percentage overall.

DPAR suggests that 40 would be a better estimate for Cleveland's return value. Rackers' field goals and extra points (which for practical purposes is almost entirely determined by FG performance) were worth 20 - and here I think it's likely that the discrepancy is the result of altitude adjustment for the game in Mexico city, where Rackers kicked 6 field goals, if I remember correctly. Hence the Cleveland return unit may have been worth roughly twice as much as the Cards FG/XP team.

What is very hard to estimate is the value of Rackers' kickoffs. FO's numbers say that Arizona had a horrendous kickoff unit overall (-20.7, worst in the league), but given Rackers' strong leg and good kick-off performance in the rest of his career, that is likely the fault of the coverage unit, and the number would have been even worse without his touchbacks.