Relatively Speaking About the Denver Broncos (Part One)

     Tunesmith, in his article "The 2010 Roster - Approximate Value" took a look at the Broncos 2010 roster using the Approximate Value rating system developed by Doug Drinen of Pro-football-reference.com (PFR). Tunesmith took the scores generated by the AV system and divided the roster into twelve categories, ranging from "The Scrubs" to "The Pro Bowlers" to "The Tigger." I'd recommend that you take the time to read Tunesmith's article.

     After talking to both Tunesmith and a representative named Neil from PFR, and having just taken a look at the rosters from Denver's six Super Bowl appearances, I thought it might be fun to compare the Approximate Value ratings for the starting offense and defense from each of those six seasons. I also thought it might be fun to apply that system to the Broncos' starting rosters from 2005 (the last time Denver played in the AFC Championship game) to 2010 (that shipwreck of a season which was just completed), the 2005-10 rating will appear in part two of this series.

Take a jump with me and see what kinds of surprises might await us.

     Before we get into how the twenty-two Broncos starters fared in the seasons under consideration, it would serve us well to take a brief look at the Approximate Value system. If you want to see the complete explanation of the system, it may be found here -- please note, it is a multi-part explanation of the math which generates the rating.

     The AV system was developed by a mathematician by the name of Doug Drinen. He liked the approximate value system that had been developed as a metric for baseball players and wanted to create an equivalent rating system for football. Doug gives this brief explanation of the system:

"AV is not meant to be a be-all end-all metric. Football stat lines just do not come close to capturing all the contributions of a player the way they do in baseball and basketball. If one player is a 16 and another is a 14, we can't be very confident that the 16AV player actually had a better season than the 14AV player. But I am pretty confident that the collection of all players with 16AV played better, as an entire group, than the collection of all players with 14AV."

He goes on to say:

"Essentially, AV is a substitute for --- and a significant improvement upon, in my opinion --- metrics like 'number of seasons as a starter' or 'number of times making the pro bowl' or the like. You should think of it as being essentially like those two metrics, but with interpolation in between. That is, 'number of seasons as a starter' is a reasonable starting point if you're trying to measure, say, how good a particular draft class is, or what kind of player you can expect to get with the #13 pick in the draft. But obviously some starters are better than others. Starters on good teams are, as a group, better than starters on bad teams. Starting WRs who had lots of receiving yards are, as a group, better than starting WRs who did not have many receiving yards. Starters who made the pro bowl are, as a group, better than starters who didn't, and so on. And non-starters aren't worthless, so they get some points too."

     I was intrigued with the way that Tunesmith has divided up the ratings, so I went to the PFR website to see if they had created a similar categorical explanation. When I could not find one, I contacted them to see if I was simply missing where they had placed a categorical description, mentioning the types of categories created by Tunesmith. Neil at PFR responded with this comment:

"It varies a lot by position -- like you said, it ranks all players relative to each other -- so a good Approximate Value for a safety isn't the same as a good AV for a quarterback."

     He went on to say that, perhaps, the staff at PFR should look into creating a generalized guideline like the one I has asked about. I thanked him and turned my attention back to Tunesmith's categories. About an hour later I received an email from Neil offering this:

A very vague guide across all positions would be like:

20-25 = MVP
14-19 = All-Pro/Pro Bowl
8-13 = Starter
3-7 = Backup
0-2 = Scrub

I just made that up off the top of my head, so it's probably wanting in many ways. And again, the mileage varies greatly by position, so a starting QB is more of a 10-12 and a starting safety is a 7-8. It makes sense, because a QB is generally regarded as having more impact than a S.

     So, keeping in mind that the numbers are very relative -- players on good teams will score higher than players on bad teams, some positions will score higher than others, even though players at those positions may be roughly equivalent -- we will take a look at how the twenty-two players that were the primary starters for the offensive and defensive units fared in the rankings. First, I will present the basic ratings, then we'll take a look at how they compare. It should be noted that I rather arbitrarily chose to omit the kickers, punters and return men.

 

 

 

 

Looking Behind the Numbers

When we look at the way the starters break down into the categories suggested by PFR, we find:

Team	1977	1986	1987	1989	1997	1998
Margin of Win/Loss	-17	-19	-32	-45	+7	+15
MVP (20-25)	0	0	0	1	0	1
All Pro/Pro Bowl (14-19)	6	3	2	1	6	4
Starter (8-13)	9	6	3	12	7	11
Backup (3-7)	7	13	17	7	8	5
Scrub (0-2)	0	0	0	1	1	1
Team Average	10.2	7.6	6.8	8.8	10.1	10.1

     It is interesting to note that as the Broncos average AV rating declined, the margin of loss in the Super Bowl games increased. We did see a surge in the average in the 1989 team, but strangely enough we saw the largest margin of loss in that Super Bowl. It is also interesting to note that in our two Super Bowl wins, the team averaged an AV rating of 10 or better.

Offense	1977	1986	1987	1989	1997	1998
Margin of Win/Loss	-17	-19	-32	-45	+7	+15
MVP (20-25)	0	0	0	0	0	1
All Pro/Pro Bowl (14-19)	0	1	1	1	4	4
Starter (8-13)	4	4	2	4	3	4
Backup (3-7)	7	6	8	5	3	1
Scrub (0-2)	0	0	0	1	1	1
Offensive Unit Average	7.7	7.4	6.7	7.2	10.9	12.4

     This set of ratings was interesting to me. In the four Super Bowl losses, our offense averaged an AV rating that was in, or just barely out of the "backup" range while in the two wins, the offensive average was solidly in the "starter" range. It can be seen that our lowest offensive averages (6.7 and 7.2) correspond to the worst of the Super Bowl defeats, while the highest averages correspond to the wins.

Defense	1977	1986	1987	1989	1997	1998
Margin of Win/Loss	-17	-19	-32	-45	+7	+15
MVP (20-25)	0	0	0	1	0	0
All Pro/Pro Bowl (14-19)	6	2	1	0	2	0
Starter (8-13)	5	2	1	8	4	7
Backup (3-7)	0	7	9	2	5	4
Scrub (0-2)	0	0	0	0	0	0
Defensive Unit Average	12.6	7.9	6.9	10.4	9.4	7.8

     This set of ratings took me by surprise. There is the old football adage that offense sells tickets but defense wins championships (roughly paraphrased). It is fairly common to hear people state the second half of that adage: defense wins championships. Yet, in every case where the defense average higher than the offense, the Broncos lost, while in the two games where the offensive average was higher, the Broncos won. In four of the six games, including the two wins, the team averaged in the Starter range. Only in the two wins did the offensive unit average in the Starter range. The defensive unit averaged in the Starter range in two of the losses and one of the wins.

When all is said and done, and the relative nature of the AV ratings are kept in mind, we are left with a picture of the Broncos that suggests -- for them at least -- the path to winning the Super Bowl was to put together a stronger offense than defense -- but not too much out of balance.

In part two of this series we will look at the starters from 2005 to 2010 to look for any patterns that may have affected the Broncos' fortunes.