Tsailand Posted October 11, 2017 Share Posted October 11, 2017 Introduction to DVOA The site Football Outsiders have come up with an advanced metric for evaluating how well NFL teams and players actually perform. The short version is that (1) For each play in every NFL game, the computer looks at the game situation (score, time, field position, down and distance), calculates what the average NFL performance in that situation is, and then compares that to the actual result. And then it adjusts for strength of opponent. And then the teams and players involved are scored accordingly, and all the scores for all their plays are averaged up and called DVOA -- "Defense Adjusted Value Over Average". Over the course of a season, you can get a very good picture of who is good and who is bad, especially at the team level. There's a full explanation here: http://www.footballoutsiders.com/info/methods One caveat is that it's hard to separate an individual player's performance from those of his ten teammates on the field with him, so keep that in mind when comparing players from different teams, especially when one team is known to have more talent. For example, a RB or QB who has an elite offensive line is going to have a better DVOA than a equally talented player who has a bad offensive line. Link to comment Share on other sites More sharing options...
Tsailand Posted October 11, 2017 Author Share Posted October 11, 2017 We've played four games, most other teams have played five. Of course, DVOA is a per-play metric, not per-game, so already each unit of each team has hundreds of action plays to be scored on. I am going to provide the rankings, not the raw scores. "1st" or "1" means best in the NFL, "32nd" or "32" means worst. Yes, that's right, the stats say we have played like the second best team in the league so far. Our offense is 11th best, defense is 4th best, and special teams are well below average. Philly is hanging right in there with us, and the other two NFC East teams are bad. Link to comment Share on other sites More sharing options...
bowhunter Posted October 11, 2017 Share Posted October 11, 2017 Excellent post. It would be cool to see how this ranking adjusts each week. Link to comment Share on other sites More sharing options...
Tsailand Posted October 11, 2017 Author Share Posted October 11, 2017 Here's more details showing how our run and pass games are doing, and how we've done vs the run and pass. The last column in each section is average strength of opponents faced so far. Don't double count this; the DVOA stat already takes this into account. This means our defense's raw performance on the field is lower than 4th overall (in fact it's 10th), but when the computer takes into account that they've faced the second hardest set of offenses so far, it bumps us up to 4th. Interestingly, New England's defense has faced the hardest schedule and the computer still says they are the worst. Which certainly matches what we see on the field. You might ask, "If our pass offense is 11th best and our run offense is 20th, how does that average out to 11th overall?" I'm not certain, but I think the answer is that we are passing more in high leverage game situations, while running in less important situations, which will tend to pull the weighed average more towards passing. For example, running on 1st and 2nd down or in garbage time vs passing on 3rd down, or late in tied games. Link to comment Share on other sites More sharing options...
Tsailand Posted October 11, 2017 Author Share Posted October 11, 2017 Now the individual stats. QB: DVOA has loved Kirk the past two past years, ranking him as high as top five, but this year he's only 15th so far. We know he starts slowly and I expect him to climb throughout the season. http://www.footballoutsiders.com/stats/qb RB running: Only Perine has enough rushes (40+) to be "officially" ranked, and they say he's a trash fire: 32nd out of 32. Thompson, with 20 rushes, has been performing better than any full-time running back in the NFL. Kelley with 29 rushes has been doing well, would be ranked 7th if he was ranked. http://www.footballoutsiders.com/stats/rb RBs as pass-catchers: Thompson is second in the NFL, just behind Lamar Miller. Perine and Kelley don't even have four targets each so they don't appear at all. Overall RB ranking: Unfortunately there is not a combined RB stat. However since Thompson is near the top of both rankings, it's safe to say that he is elite at the position, with the caveat that his body can't take the pounding a fulltime RB does. I belive that if we had three Chris Thompsons and platooned them, we would have the best running offense in football. TE: Vernon Davis 2nd in the NFL. Reed 40th out of 41 TEs ranked. This only takes into account pass catching, not blocking. (Gronk is only 8th?!) http://www.footballoutsiders.com/stats/te WR: Pryor is 36th out of 74 ranked. Middle of the pack. None of the others have enough targets to appear on the main ranking, but if they did, Doctson would be 6th, Grant would be 10th (!?), and Crowder would be 70th. You can argue that this is small sample size, Doctson and Grant aren't actually top 10 receivers... so how about Gruden gives them more touches so we can find out for real? I'm ****ing sick of seeing the Kirk-Pryor connection fail. http://www.footballoutsiders.com/stats/wr O-line: If I'm reading this right, It says we're 11th run blocking and 11th pass blocking. http://www.footballoutsiders.com/stats/ol D-line: 29th against the run... and 4th against the pass. http://www.footballoutsiders.com/stats/dl Link to comment Share on other sites More sharing options...
Warhead36 Posted October 11, 2017 Share Posted October 11, 2017 Interesting stats. Cousins is better then 15th but small sample size plus one fewer game played means he'll be ranked lower. Thompson being ranked elite albeit without as many touches isn't surprising. Its a shame he's too small to be a full fledged whole game back but I love him in his whole. DL being 29th against the run seems off. I mean I know we gave up 100 to Gurley and Hunt, but those guys are both top 5 backs at the moment and I don't think the DL was as much the issue there. Being ranked 2nd overall in the league is...weird. Hopefully we can justify that with an ass whooping of the 49ers. Good teams crush bad teams. Link to comment Share on other sites More sharing options...
Long n Left Posted October 12, 2017 Share Posted October 12, 2017 8 hours ago, Warhead36 said: Thompson being ranked elite albeit without as many touches isn't surprising. Its a shame he's too small to be a full fledged whole game back but I love him in his whole. Wait...wait...what? Link to comment Share on other sites More sharing options...
zoony Posted October 12, 2017 Share Posted October 12, 2017 Still way too small a sample size. Take out the raiders game, for instance, where does the team fall? Glad for obvious progress but some of the gains are being blown out of proportion Link to comment Share on other sites More sharing options...
TSO Posted October 12, 2017 Share Posted October 12, 2017 Special Teams must’ve got killed for the two Crowder muffs (and rightly so) because, otherwise, I feel like they’re doing a good job. Link to comment Share on other sites More sharing options...
Tsailand Posted October 12, 2017 Author Share Posted October 12, 2017 On 10/10/2017 at 10:36 PM, bowhunter said: Excellent post. It would be cool to see how this ranking adjusts each week. Depends on how much interest there is. I'll certainly do a full update after the season midpoint. On 10/10/2017 at 10:50 PM, Tsailand said: You might ask, "If our pass offense is 11th best and our run offense is 20th, how does that average out to 11th overall?" I'm not certain, but I think the answer is that we are passing more in high leverage game situations, while running in less important situations, which will tend to pull the weighed average more towards passing. For example, running on 1st and 2nd down or in garbage time vs passing on 3rd down, or late in tied games. Just ignore this, I didn't think the math through. There's no particular reason why a combined ranking should be the average of the component rankings. For example, our D is 4th overall even though we are "only" 7th against the pass and 8th against the run. 13 hours ago, Warhead36 said: Interesting stats. Cousins is better then 15th but small sample size plus one fewer game played means he'll be ranked lower. Kirk has ~150 plays on the season. That's plenty to get a reliable picture. DVOA is a per-play stat, not per-season. So one less game doesn't mean a lower ranking. 13 hours ago, Warhead36 said: DL being 29th against the run seems off. I dunno man, it seems like it's our linebackers and secondary doing the work against the run. Here's how the site assigns credit to the D-line for run stopping: * Losses: 120% value * 0-4 Yards: 100% value * 5-10 Yards: 50% value * 11+ Yards: 0% value And we're 29th at that. Interesting, in the "open field" category, all run plays after 10 yards, we rank 7th, which is damn good -- and doesn't count towards the D-line rating at all. 3 hours ago, zoony said: Still way too small a sample size. Take out the raiders game, for instance, where does the team fall? It's 200+ plays for each of the offensive and defensive units. I don't know where we rank if you throw out the Raiders game, but do you have a particular reason for wanting to do that? 1 hour ago, thesubmittedone said: Special Teams must’ve got killed for the two Crowder muffs (and rightly so) because, otherwise, I feel like they’re doing a good job. It says we're bad at FGs also. See below. The numbers represent additional points lost or gained over the entire season so far, compared to an average team. Minus 2.4 doesn't seem like very much, but think how many close games there are in the NFL. Link to comment Share on other sites More sharing options...
TSO Posted October 12, 2017 Share Posted October 12, 2017 54 minutes ago, Tsailand said: It says we're bad at FGs also. See below. The numbers represent additional points lost or gained over the entire season so far, compared to an average team. Minus 2.4 doesn't seem like very much, but think how many close games there are in the NFL. That’s slightly surprising but, yeah, makes sense. Hopkins two misses were both 50+ yard tries so it’s understandable to a degree, though you’d like him to come away with them since they were on the right side of 50 (I think 52 and 53 yards). I wonder if they take that into account at all? Link to comment Share on other sites More sharing options...
Tsailand Posted October 12, 2017 Author Share Posted October 12, 2017 I'm sure they take FG distance into account. Most of his kicks this season have come from distances so close that other kickers make them 98% of the time. So they do nothing for him (but missing one would hurt him badly). He's 2/2 from 40-49, and 0/2 from 50+. I don't know exactly how their formula works, but they say that adds up to negative 2.4 points and I find it plausible. The expectations for the average NFL kicker these days are very high. It's not like in the old days where kicks were a crap shoot. Link to comment Share on other sites More sharing options...
zoony Posted October 12, 2017 Share Posted October 12, 2017 7 hours ago, Tsailand said: 200+ plays for each of the offensive and defensive units. I don't know where we rank if you throw out the Raiders game, but do you have a particular reason for wanting to do that? Truncated mean Main point is sample size still too small Link to comment Share on other sites More sharing options...
skinny21 Posted October 12, 2017 Share Posted October 12, 2017 I tend to agree with @zoony about wanting to see our rank without the Oakland game (as in, how big of an anomaly was it), but... if you do that, you'd need to throw out our worst game as well. You'd also have to do the same for all other teams... which might be kind of ridiculous. Also like that this serves as serves as a quick reference about our opponents. 49ers, for example are interesting on offense - bottom 4 offense for both pass and run, but a 12 ranking. I am confused as to why they get the 12 ranking (2nd table/chart), but the first table shows them so poor... what am I missing there? Link to comment Share on other sites More sharing options...
MassSkinsFan Posted October 12, 2017 Share Posted October 12, 2017 1 hour ago, skinny21 said: I tend to agree with @zoony about wanting to see our rank without the Oakland game (as in, how big of an anomaly was it), but... if you do that, you'd need to throw out our worst game as well. You'd also have to do the same for all other teams... which might be kind of ridiculous. Also like that this serves as serves as a quick reference about our opponents. 49ers, for example are interesting on offense - bottom 4 offense for both pass and run, but a 12 ranking. I am confused as to why they get the 12 ranking (2nd table/chart), but the first table shows them so poor... what am I missing there? That 12 is a schedule ranking, aka strength of their opponents. It is taken into account when calculating the other DVOA scores. The 12 score means they faced some pretty good defenses. Their 29 Schedule Rank score on D means they've faced really bad offenses. Link to comment Share on other sites More sharing options...
Always A Commander Never A Captain Posted October 12, 2017 Share Posted October 12, 2017 On 10/10/2017 at 11:21 PM, Tsailand said: Now the individual stats. QB: DVOA has loved Kirk the past two past years, ranking him as high as top five, but this year he's only 15th so far. We know he starts slowly and I expect him to climb throughout the season. http://www.footballoutsiders.com/stats/qb D-line: 29th against the run... and 4th against the pass. http://www.footballoutsiders.com/stats/dl DL, that looks about right. The Raiders OL isn't what it used to be, so that's not as impressive. The Chiefs OL was having it's way with Hood, McGee, and McClain. We need more quality depth on the DL. Allen/Ioannidis are looking really good, but that's just the Nickel package. I also really like the ANY/A stat for QB's, and Cousins is presented well there, I believe he's in the Top 10. Link to comment Share on other sites More sharing options...
CTskin Posted October 12, 2017 Share Posted October 12, 2017 31 minutes ago, Alcoholic Zebra said: DL, that looks about right. The Raiders OL isn't what it used to be, so that's not as impressive. The Chiefs OL was having it's way with Hood, McGee, and McClain. We need more quality depth on the DL. Allen/Ioannidis are looking really good, but that's just the Nickel package. I also really like the ANY/A stat for QB's, and Cousins is presented well there, I believe he's in the Top 10. I stated this in another thread- over the last 3 quarters of the KC game, the ToP differential was 33 mins to 12. You just can't expect the big uglies to stay strong with that kind of field time. Take a look at how well the Chiefs did in the 1st quarter when our D wasn't gassed... Link to comment Share on other sites More sharing options...
Tsailand Posted October 12, 2017 Author Share Posted October 12, 2017 10 hours ago, zoony said: Truncated mean Main point is sample size still too small 200+ plays is too small? 5 hours ago, skinny21 said: I tend to agree with @zoony about wanting to see our rank without the Oakland game (as in, how big of an anomaly was it), but... if you do that, you'd need to throw out our worst game as well. You'd also have to do the same for all other teams... which might be kind of ridiculous. Throwing out one data point sometimes make sense, like with a running back who crashes into the line for 1 yard 29 times, then rips off a 91 yard run. His average is 4.0 YPR, but he's still trash who will cost you games with his inability to get first downs. (I believe DVOA will evaluate him as trash, because it doesn't simply average yards. Those 29 rushes will all count as failed plays unless they came on 2nd/3rd down with 1 yard to go. 29 failed plays outweighs one very successful play) However, you and zoony aren't asking to throw out one data point, but 50+. So unless you think there is something anomalous about the game, the request doesn't make sense. It isn't enough to point out that our D played well that day. However, if for example you had evidence that the Raiders offense all had the flu that day, or were tanking to punish Carr for not kneeling, or whatever, it would be reasonable to throw all 50+ plays from that day. 2 hours ago, Alcoholic Zebra said: I also really like the ANY/A stat for QB's, and Cousins is presented well there, I believe he's in the Top 10. He's 4th in ANY/A. I like that stat too, but I don't want to discuss Cousins much here, because it will take over the thread. Feel free to post about ANY/A and QBR in the Cousins thread :-) Link to comment Share on other sites More sharing options...
skinny21 Posted October 12, 2017 Share Posted October 12, 2017 @TsailandFirst of all, thanks for putting this together. I'm absolutely fine with the overall data, but yes, the reason I'd be curious to see how things change without the Raiders game is because I did feel it was an anomaly. Just from a historical perspective, it was abnormal for this team. So, for example, it's possible that Manusky just caught their OC flat footed and wound up thoroughly outcoaching him. Or that our D happens to match up perfectly with the Raiders (tackling well to limit YAC, and our dline just having the right guys across from them, etc.). Or, it could just be that game was the realized potential of the D. Regardless, I'm not arguing that we should throw out that game, just that I wonder what we look like without it. Link to comment Share on other sites More sharing options...
Tsailand Posted October 17, 2017 Author Share Posted October 17, 2017 On 10/12/2017 at 6:15 PM, skinny21 said: Regardless, I'm not arguing that we should throw out that game, just that I wonder what we look like without it. I don't see that information available on the site. Maybe with a premium membership. As the season continues, the weight of just one game decreases. They've updated individual player rankings after week six. Team rankings will refresh later tonight. QB: Kirk is up to 8th now. RB (minimum 48 rushes to be ranked): Perine is still the worst in the NFL, 34th out of 34. Thompson with 36 rushes is performing around the level of the 9th ranked. Kelley with 29 rushes would be 10th. RBs as pass-catchers: Thompson is best in the NFL, no surprise after the screen passes he took Sunday. Perine has only five passes, and you need twelve to be ranked, but he's looking good, equivalent to 5th in the NFL. TE: Davis 1st in the NFL per-play. His total value is lower than other TEs who have gotten far more targets; 14 to him compared to 40+ for Gronk and Kelce. Reed is 39th out of 42. Probably secretly hurt. WR: Pryor down to 52nd out of 78. Crowder 74th. Grant and Doctson looking good with small sample sizes. O-line: 17th run blocking, 9th pass blocking. D-line: 28th against the run, 11th against the pass. My takeways are the same as before: (1) Kirk is continuing to climb because he's a very good QB and week 1 was an anomoly (2) Perine is trash (3) Davis is a stud, target him more please. (4) Doctson and even Grant need more targets Link to comment Share on other sites More sharing options...
Tsailand Posted October 17, 2017 Author Share Posted October 17, 2017 Team efficiency stats after week 6: Overall, we've dropped to 6th and the Eagles have risen to 2nd. Offense is up to 8th overall. 4th passing and 22nd running. Defense is down to 10th overall. 10th against the pass, 15th against the run. Link to comment Share on other sites More sharing options...
zCommander Posted October 18, 2017 Share Posted October 18, 2017 44 minutes ago, Tsailand said: Team efficiency stats after week 6: Overall, we've dropped to 6th and the Eagles have risen to 2nd. Offense is up to 8th overall. 4th passing and 22nd running. Defense is down to 10th overall. 10th against the pass, 15th against the run. Just now on NFLN they have us as #7 team playing #2 team on Monday. Where is the discrepancy coming from? Link to comment Share on other sites More sharing options...
Tsailand Posted October 18, 2017 Author Share Posted October 18, 2017 We're listed as 6th on the FO site: Link to comment Share on other sites More sharing options...
Warhead36 Posted October 18, 2017 Share Posted October 18, 2017 Thanks for digging up all the info Tsai. Offensively it appears we're similar to last year, but suffering from not having the same good WRs we did last year and Reed being hurt too. Tells you how good Cousins is that he's still basically a top 10 QB after losing two good experienced WRs. Pay him. Still a bit surprised our run D has sagged so much. I know we've given up some but didn't think it'd be that low. Not looking forward to seeing what it'll look like with the injury to Allen. We have a good team though. Its a shame we've played such a brutal schedule. If we were in the AFC North we're probably 5-0. Link to comment Share on other sites More sharing options...
BRAVEONTHEWARPATH93 Posted October 18, 2017 Share Posted October 18, 2017 Further confirmation we have a shot to have a good year Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.