My name is Azrakal (or Az, if you hate typing), and I'm here to give you guys a statistical look at our future Hall Of Famer: Demarcus Ware. A little bit about me first: I'm originally from the DC area (where I've given my brother fits for being a Cowboys fan mwahaha), but am currently studying Computer Science at the University of Illinois. I've followed the Cowboys for about 4 years, and this blog for about 2. I'm also a HUGE Romo and Garrett homer, but probably because I'm just a naive college student. Random fact about me: I was also a pretty big Terrence Newman fan back in the day because he always got the interceptions when I played Madden.
Ok, enough about me, let's get down to some stats, shall we?
I'll admit to you, I'm no stats genius or anything, but I do LOVE me some stats. I say this because stats don't lie and they can tell us incredible things about our world. Of course, people can get very, very misled by statistics, but only because those people see/use incorrect statistics. However, when you look at the right statistics in the right way, your data will tell you things you never thought you could know.
Football is a statistics enigma. The metrics used to measure players are not really fully developed yet. An example of this is the PFF player gradings that are so hotly debated on this site. These measurements are valuable in their own right, but since the formula PFF uses is not public to my knowledge (and since some of their measurements are a bit wacky), we can't accept this as a universal solution to player grading. This will be a huge hindrance to what I'm about to do, and one reason why a lot of people/journalists don't use statistics in football analysis. However, I'm going to eat my words here, because there is really no alternative for me, and run an analysis using PFR's Approximate Value ratings since I have no money for PFF. (Also, because PFR actually gives us some of their stats equations)
AV is calculated as follows:
individual_points = [(games played) + 5*(games started) + sacks + 4*(fumble recoveries) + 4*(interceptions) + 5*(defensive TDs) + (tkl_constant)*(tackles)] + (all_pro_bonus),
approx_value = [ (individual_points) / (sum of individual_points for all front-seven players on the team) ] * team_points_for_front_7
Basically, the formula gives us a way to "weigh" stats and account for the differences in the way football was played from say 1970 to 2007. Also, a player's performance is weighed for their performance with respect to their position group on the team, which I believe is an awesome idea. If you want to read up on it yourself, here you go.
So what I did was take 10 Hall of Fame DE/OLB players, average their AVs and create a kind of trend line from which we can extrapolate an approximate career path of our beloved DE/OLB.
A little explanation here: The first data points are the averages of the players' AV values through ages 24-28. I started at 24 because I wanted to give the player a few years after coming out of college so his "peak" AV values could shine through. Obviously, that didn't happen with a lot of the players, so I marked the ones that were 24 within a year of entering the NFL with stars next to their names. I also assigned an AV of 0 to the players after they retired. I did this to take those retirements into account for this analysis, since I think Ware's retirement could be reality in the near future. Y-axis is Approximate Value; x-axis is Age. Pretty self-explanatory beyond that.
You can see the players I made this average line from, all of whom are Hall of Famers, played in the NFL during the 1980s to the early 2000s. I believe AV's fudging factors (see: link above) account for this enough, but I suppose that's up for debate.
Notice the trend of the line is downwards (obviously), but the real interesting thing we see here is that the Hall of Famers' average really begins to drop off around age 32-33. One could attribute that to the retirement of Derrick Thomas, but if you look at the data only Rickey Jackson and Bruce Smith (whose AV is ridiculous) improve their AV during this time, so I think this it's reasonable to say that generally our Hall of Famers begin to drop off around that age.
So let's compare this average line to Demarcus
Ware's AV looks to be a bit above the trend line, up until he hits 30. I think he looks enough like he's going to follow this trend line pretty well, but I don't really know how to quantify this opinion. If anyone knows a statistical method to do this, please tell me! Otherwise, I'm going to draw the conclusion that Ware is going to follow this trend, and will produce at his current level until he dings 33.
One final comparison:
Oh boy, this looks even less encouraging for Ware. I included this graph with Randle because it follows a similar trend to Ware in the way that they both peak at 29 and then steeply decline after 31. From this graph, it becomes apparent that Ware's AV value could decline as soon as age 32 (which is next year).
I've got to say, I did not expect this outcome. Like our beloved owner Jerry Jones, I hate letting go of superstars, but in this case I think it may be safer to bite the bullet on Ware. Statistics is not about sentimentality, and the statistics in this case suggest that he either needs to take a pay-cut or be released because judging from past players, he will start declining pretty soon. Hopefully he'll stick around though, he's still a stud in my heart.
Thanks for taking the time to read my (SUPER LONG, sorry) fanpost! If you'd like me to explore another topic, have a question about this article, or think I'm totally bonkers then tell me down in the comments section!
P.S. Notice that I said "suggest" in my analysis, rather than a more definitive verb. Statistics, while really cool, do not give us certainty about anything. I could be proven completely wrong, and Ware could have several monster seasons in line for us. Statistics only gives us likely scenarios. If any of you want to use statistics in the future, it would be helpful to remember this fact.