Here they are, since 2000, broken out by year and position. There are the "playmaker" stats +WPA and +EPA, but there's also Tackle Factor (TF). There are conventional stats too, including interesting stuff like passes defended, tackles for a loss, and QB hits.
For now linebackers are all lumped together, without regard to 3-4 or 4-3 scheme. For that matter, tackles and ends are lumped together as well. In the future I'd like to differentiate positions by scheme.
I also plan to add some more stats. For example, "victories" (or some other lame name) would be the number of plays that a defender made that resulted in a loss of expected points (EP) for the offense. They could be stuffs, passes defended, turnovers, sacks, or even tackles 6-yards deep on 3rd and 7. I've left success rate (SR) in the tables, but for defenders I think the total number of successes makes more sense.
The tables are sortable, with the default sort key set to +WPA. I think +WPA may be the most interesting stat, but +EPA is going to be a better indicator of long-term player ability.
- Home
- Unlabelled
- Individual Defender Stats
Individual Defender Stats
By
Brian Burke
published on 5/10/2010
Subscribe to:
Post Comments (Atom)
I would stay away from a name like "victories" because for readers of sites like this one it might be easily confused with "win shares".
Great stuff as usual, Brian.
Brian,
I seem to remember taht when you first came out with Tackle Factor, K Dansby has a TF of 1.01, now he is shown as 1.24.
Have you reworked this number?
What I am wondering is how we can use all these numbers together, to get a picure of the player.
For now, I use these numbers as a guide, to see who is supposed to be good and not so good aand watch their performance to see if it matches.
Great work, and thanks for your efforts!
Brandon in NY
Could you add in career stats as an option under "Year"?
I find it amazing how big of a lead Charles Woodson has above other CB's in the +EPA. It's nice to see from year to year and have all of this information available. It will be interesting once you have time to get the team pages updated to do analysis on different teams.
Re: K Dansby--There are a couple changes. First, I've removed all special teams tackles from both sides of the ledger: player tackles and team total tackles. Also, all ILBs and OLBs are lumped together, which doesn't make much difference.
Stats by team will be available very soon. Individual career stats are almost ready to go except for the guys who changed numbers. (Defenders love to change numbers for some reason.)
Instead of "victories," maybe I should just use "successes." That's a little clearer, and it's consistent with my Success Rate (SR) stat for offensive players.
Wouldn't the number of successes have to be somehow grounded by the number of opportunities for success? I realize this may be a bridge too far, it just seems like it's hard to judge Darrelle Revis by all the times he didn't have an opportunity to make a pick because he was covering a receiver so well that the guy didn't even get a look.
I'm assuming there is some sort of cutoff for the players listed, otherwise that list would have to be much longer. Is it a certain number of snaps?
The cutoff is in games played. I think I set it to 6 or 7.
I think a better cutoff would be snaps...you're missing guys like Jairus Byrd who had 9 interceptions for example.
Yeah snaps would be a better cutoff, but the only place I've seen those kinds of numbers are at Pro Football Focus, and that only goes back to 2007.
Snaps would be great. I would LOVE snap data for a lot of purposes. However, what's publicly available now is not that reliable.
PFF looks like a great site.
Is there a reason you have EPA+/Game, but not WPA+/Game?
Brian, Can you adjust your tackle factor by run/pass ratio? I think it would make the stat better. For example, DT don't get a lot of tackles on pass plays so they should not be penalized for that.
Oops, sorry. The minimum cutoff is set to 20 tackles. I meant to change that to game appearances.
Nothing against +WPA/G, just running out of room.
Run/pass ratio seems like it could be wise adjustment. I'll have to think about how to implement that.