tag:blogger.com,1999:blog-38600807.post2402514923772669725..comments2023-11-05T04:16:44.937-05:00Comments on Advanced Football Analytics (formerly Advanced NFL Stats): 2013 Seahawks Defense: In the Conversation for Best Ever?Unknownnoreply@blogger.comBlogger24125tag:blogger.com,1999:blog-38600807.post-27193959831201846952014-03-01T04:21:08.584-05:002014-03-01T04:21:08.584-05:00Statistics always have been and always will be not...Statistics always have been and always will be nothing but a tool to make biased claims with no basis in reality. Why do you think politicians quote statistics endlessly? Because they can be used to 'prove' anything whether there is an ounce of truth or not.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-92023017898262080952014-03-01T04:13:06.125-05:002014-03-01T04:13:06.125-05:00These articles are less scientific than the Eugeni...These articles are less scientific than the Eugenics of the 1920s. Truly awful. Blanket assumptions with no basis in reality like Denver had a 'bad day' or 'Seattle holds on 60% of plays'. How about that article about how much better the top offenses are than the top defenses? Great work. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-26526377741011915462014-02-27T22:23:05.440-05:002014-02-27T22:23:05.440-05:00Speaking of a Denver "bad day", here'...Speaking of a Denver "bad day", here's my estimate of their bad day (Brian, avert your eyes, this is probably statistical heresy):<br />Denver averaged 37.9 ppg with a standard deviation of 9.6. Let's a bad day would be Denver scoring 28 (their worst day was against SD, they scored 20).<br />Seattle's defense allowed 14.4 ppg with a standard deviation of 8.8. A good day for Seattle would be allowing 6 points.<br />Average those two and you get 17.<br />This is quick and dirty, but it's my estimate of a Denver "bad" day against Seattle. <br />I'm thinking Denver's performance in the SB was the worst since the Redskins epic meltdown against the Raiders in '84.<br />It was way more than a "bad" day.<br />Anonymoushttps://www.blogger.com/profile/03108711283276399018noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-48294375750087328902014-02-27T21:55:48.119-05:002014-02-27T21:55:48.119-05:00Coming in late to the party here; Dave brings up a...Coming in late to the party here; Dave brings up a great point. EPA is about the best way to evaluate past performances because it catches everything...but that's also why you have to be carefule with it: it catches what we might consider fluky events (fumble recovery touchdowns, kickoff return TD's, James Harrison's 99-yard pick-six, etc.). This whole discussion is based on how you want to work the numbers and whether or not you, personally, feel that a defense "deserves" credit for said "fluky" events. Depending on where you stand, you'll get different answers.<br />All that being said, Seattle just dominated what most consider to be the greatest offense the league has ever seen on the biggest stage and that has to account for something. Maybe not number 1 or 2, but they're in the discussion in my book.<br />Anonymoushttps://www.blogger.com/profile/03108711283276399018noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-45747705014222520142014-02-18T08:41:34.477-05:002014-02-18T08:41:34.477-05:00I think 91 Eagles would be greatest by EPA. They w...I think 91 Eagles would be greatest by EPA. They was flatout stinking and had an even worse offense than the 2000 Ravens. Andersnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-41752787827428139422014-02-10T18:18:34.932-05:002014-02-10T18:18:34.932-05:00Is it possible to do this analysis for just the pa...Is it possible to do this analysis for just the pass defense, rather than defense as a whole? It seems to me that the better a team's offense is, the more important their pass defense becomes relative to the run defense, since the opponents will be playing from behind more often. J.D. Krullnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-47487720511667891852014-02-09T22:34:21.819-05:002014-02-09T22:34:21.819-05:00Brian,
I still think 2000 Baltimores fumble recov...Brian,<br /><br />I still think 2000 Baltimores fumble recovery luck and the number of fumbles vs them makes up almost all EPA difference between them and seattle. Also Baltimore was not forcing more fumbles nor did they have more sacks than SEA so their recovering a lot of fumbles was probably mostly due to randomness. <br /><br />Baltimore recovered 26 fumbles in the regular season. Seattle recovered 11. Thats a difference of 15! If we assume 4 pts of EPA per fumble thats a difference of 44EPA on recovered fumbles or 60/16= 3.75 EP per game. or -3.75 since we are looking at the defensive side.<br /><br />so -8.7 + -3.75= -12.25. <br /><br />Okay... if you say well interceptions are a fairly lucky events as well and Seattle got more of those than Baltimore. In that case we can just look at total turnovers. And Seattle had ten less then Balitmore or 49 to 39. So then we are at 40/16 or 2.5 pts per game advantage<br /><br />That makes Seattles adjusted total -8.75 + -2.5= 11.25. We are still talking about less than half a point per game difference.<br /><br />I think they are definitely in the conversation. <br /><br /><br /><br /><br />Davenoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-16754069983181512662014-02-09T14:14:48.154-05:002014-02-09T14:14:48.154-05:00And Mitch, I won't humiliate you too badly, bu...And Mitch, I won't humiliate you too badly, but you are going to have to get the smackdown.<br /><br />If SEA was so clearly and so drastically ahead of the rest of the league, how was it that they lost to IND, SF, and ARI?<br /><br />I'll assume your answer would (rightly) be: even great teams can have a bad game.<br /><br />Ok, so why wouldn't that apply to other teams? <br /><br />Or are you just some tout?Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-73916681849273471982014-02-09T13:43:34.833-05:002014-02-09T13:43:34.833-05:00Bruce-That's absurd and offensive. The reason ...Bruce-That's absurd and offensive. The reason I mentioned the 'off-night' was to address the silliness of opinions that rest on the outcome of one game, as good as it was.<br /><br />The numbers are what they are. I could have written that the Broncos were crappy all year, and the numbers would still be the same. Continue to shop for whatever numbers suit your own preordained conclusions. Other sites might cater to the hype of recency for clicks, but we don't do that here.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-719881906255994102014-02-09T13:34:31.663-05:002014-02-09T13:34:31.663-05:00The statistical analysis is interesting, and presu...The statistical analysis is interesting, and presumably could be argued for days, as other stats-based sites certainly have concluded that the Seahawks belong in the conversation for the best defense in the Super Bowl era. But when the writer speculates that Denver may simply have had an "off night", and Denver might have lost to an "average team", he reveals bias, that is, he reveals that he started the analysis with a preference for a certain result, and in the stats analysis world, that is a big no-no. The Seahawks simply dismantled the most prolific offense the NFL has ever seen, holding that offense to 30 points under its season average. It was far faster, it was far more physical, and its game plan was just about perfect, all in the most important game of the season, and in some cases of players' careers. Not sure how this model accounts for that kind of domination on that stage, but it certainly should.Bruce M.https://www.blogger.com/profile/00191858781733895281noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-71146820109135903762014-02-07T21:40:44.886-05:002014-02-07T21:40:44.886-05:00Beginning In week 5 I called Seattle the best team...Beginning In week 5 I called Seattle the best team, and said many times they were well above the rest of the league.<br /><br />Since week 5 the model had, Bengals, Saints and Broncos number 1, along with Seattle.<br /><br />At one point when the model did have the Seahawks no. 1 with the saints only .02 behind, I pointed out the the Saints should of been .24 behind not .02.<br /><br />.02 is representive of a close game and a -3.5 spread. Seattle led that game by 16 points most of the game and it took the Saints scoring on 4th down with less than 30 seconds left to close to somewhat of a respectable score.<br /><br />And the Super Bowl speaks for itself.<br /><br />The site tried saying that the model had Seattle no.1 for like 8 or 9 weeks of the season, as though this was an impressive feat.<br /><br />Seattle was well above the rest of the league, having them no.1 for 8 or 9 weeks is a weakness in the model not a strength. And especially having them ONLY .02 above the 2cd place team.<br /><br />I'm just wondering, is your defensive EPA about as good as your model ?Mitchnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-81590234237940261102014-02-06T13:57:03.550-05:002014-02-06T13:57:03.550-05:00Interesting article. Thanks.
But articles like th...Interesting article. Thanks.<br /><br />But articles like this always remind me of a humorous quote by the great Waldo Pepper.<br /><br />"I'm going to be honest with you. There's just no way I'm the best pilot in the world. I'm the second best pilot in the world."<br /><br />Is SEA the best D of all time? Well, being third place by a tiny statistical margin out of some 480 teams (over 15 years), I say close enough. And I say that as a broncos fan.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-72820574146964714472014-02-06T00:30:05.219-05:002014-02-06T00:30:05.219-05:00This and other football stats model all suffer the...This and other football stats model all suffer the same flaws, they could not account for synergistic effects of complementary football among the different units of football or in the case of bad teams "dysergistic" effects with the added complexity of random luck. Add in the effects of different offensive/defensive philosophies, coaching, personnel and match up problems then we have a case here of imperfect and/or coarse models.<br /><br />It might look like a fruitless endeavor but that's why it so fascinating and in my opinion that's why they play the game of football otherwise they would just be modelling it. :)Anonymoushttps://www.blogger.com/profile/13455247722802489844noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-75468690816920354252014-02-05T18:44:35.653-05:002014-02-05T18:44:35.653-05:00Averaged per pass including sacks.
2000 Ravens 5....Averaged per pass including sacks.<br /><br />2000 Ravens 5.3<br />2002 Bucs 4.5<br />2012 Seahawks 4.8<br /><br />Ravens played in a division of very weak offenses<br />Seahawks played in a season that was the easiest in history to pass the ball.<br /><br />Silly conclusion to say this defense isn't one of the best ever.<br /><br />And I especially like how you try to justify your ratings by saying, we must admit Manning and his crew had a off night.<br /><br />That statement seals the deal as far as your credibility and knowledge.<br /><br />The truth is, you don't have any clue whether it was Seattle's incredible defense or Denver's offense having an off night, your guessing because you want it to be.<br /><br />After all, you did call for a close game and you did have Denver no.1 in your model a number of times this season, you were wrong just admit it, it was not likely a off night for Denver, it was more an off night for your model.Mitchnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-89590732194060642782014-02-05T14:57:36.179-05:002014-02-05T14:57:36.179-05:00Playing devil's advocate here. EPA doesn't...Playing devil's advocate here. EPA doesn't "adjust" for things that are more luck than skill like fumble recoveries and interceptions. <br /><br />How much does "turnover" luck skew this data though.<br /><br />The 2000 Ravens forced 19 fumbles but recovered 26.<br />The 2012 Seahwaks forced 24 fumbles but recovered 15<br />The 2002 bucs forced 8 fumbles and only recovered 8 <br /><br />In general 2000 Ravens got 48 turnovers, 2002 Bucs got 38 and the 2013 seahawks got 39. <br />Davenoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-5799159907731484662014-02-05T13:42:35.130-05:002014-02-05T13:42:35.130-05:00Defensive Penalties.
http://www.sportingcharts.co...Defensive Penalties.<br /><br />http://www.sportingcharts.com/nfl/stats/team-defensive-penalties/2013/<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-58088450809331218822014-02-05T13:40:47.445-05:002014-02-05T13:40:47.445-05:00How many penalties do you think the 2000 Ravens wo...How many penalties do you think the 2000 Ravens would have received under the new rules for hitting a "defenseless" receiver and helmet to helmet contact? That Ravens defense could literally get away with murder. When Seattle's defense has struggled this year, it has always been against a stingy officiating crew. According to this source Seattle was third in defensive penalties this year. That Ravens defense could literally get away with murder.<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-75487663776312522122014-02-05T13:08:32.519-05:002014-02-05T13:08:32.519-05:00I've always thought that the best TB defense w...I've always thought that the best TB defense was in 1999, but that's only the eye test so it'd be great to see all the EPA numbers for them. The 2002 unit certainly scored more TD's though, but I always thought that that was a bit of a fluke.ilikeflowersnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-3409037594632223232014-02-05T11:08:03.901-05:002014-02-05T11:08:03.901-05:00...and worst 25 since 1999:
2008 DET (4.48)
2002......and worst 25 since 1999:<br /><br />2008 DET (4.48)<br />2002 KC (2.86)<br />2010 HST (2.74)<br />2000 SF (2.69)<br />2004 KC (2.66)<br />2005 HST (2.63)<br />2002 MIN (2.62)<br />1999 CLV (2.60)<br />2004 MIN (2.46)<br />2008 DEN (2.29)<br />2004 OAK (2.21)<br />2006 WAS (2.14)<br />2009 DET (2.05)<br />2001 DET (1.89)<br />2000 SL (1.84)<br />2001 ATL (1.80)<br />2013 MIN (1.71)<br />2011 CAR (1.65)<br />2000 SEA (1.61)<br />2000 MIN (1.61)<br />2013 SD (1.56)<br />1999 SF (1.50)<br />2000 ARZ (1.48)<br />2002 CIN (1.44)<br />2004 IND (1.35)<br />Nick Bradleyhttps://www.blogger.com/profile/07772794102967712253noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-38679687469506530212014-02-05T11:07:05.643-05:002014-02-05T11:07:05.643-05:00what I did is did a trend line of year vs. DWPA an...what I did is did a trend line of year vs. DWPA and year vs DEPA. <br /><br />I then had a DWPA and DEPA above expected for each.<br /><br />Then, multiplied them together. <br /><br />Then, converted that score into a z-score. Top 25 defenses since 1999:<br /><br />2000 BLT 6.02 <br />2002 TB 5.50 <br />2003 BLT 5.01 <br />2008 BLT 4.74 <br />2008 PIT 4.31 <br />2013 SEA 3.59 <br />1999 TB 3.59 <br />2005 CHI 3.27 <br />2006 BLT 3.08 <br />2003 NE 2.93 <br />2006 CHI 2.66 <br />2010 PIT 2.40 <br />2000 TEN 2.19 <br />2001 BLT 2.16 <br />2001 CHI 1.99 <br />2010 GB 1.98 <br />2011 SF 1.89 <br />1999 JAX 1.75 <br />2011 BLT 1.74 <br />2008 PHI 1.71 <br />2008 TEN 1.62 <br />2012 CHI 1.61 <br />2000 NYG 1.49 <br />2002 PHI 1.48 <br />2013 CIN 1.48 <br />Nick Bradleyhttps://www.blogger.com/profile/07772794102967712253noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-92150395803081207792014-02-05T09:37:58.726-05:002014-02-05T09:37:58.726-05:00Yes, it seems to me it would be interesting to com...Yes, it seems to me it would be interesting to compare EPA per play as well. Obviously both total EPA and EPA/play are potentially important indicators, and with a run-focused ball-control offense, SEA definitely faced less plays than even the average defense of their caliber would expect to, I'm guessing.Timnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-1222482420764623312014-02-05T08:39:46.980-05:002014-02-05T08:39:46.980-05:00Interesting. I'm curious if it's possible...Interesting. I'm curious if it's possible to control for the number of chances (either plays or drives), since EPA is a counting stat. I think Seattle's defense faced a below-average number of opposing plays per game, but I don't know if that's also true of the other great defenses mentioned.Brian Andersonnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-66003431010904696832014-02-05T08:19:18.074-05:002014-02-05T08:19:18.074-05:00Forgot:
If you use the other WPA measure that inc...Forgot:<br /><br />If you use the other WPA measure that incorporates the Vegas line (pro football reference uses this), you get automatic adjustment for strength of opponent. Nick Bradleyhttps://www.blogger.com/profile/07772794102967712253noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-49044821201867777062014-02-05T08:10:37.885-05:002014-02-05T08:10:37.885-05:00Have you looked at DWPA?
It seems like that would...Have you looked at DWPA?<br /><br />It seems like that would hold over time.<br /><br />If you do that, SEA isnt even close to the 2002 Bucs, or even the 1999 Bucs, or the 2000 Ravens.<br /><br />Rather, the 2013 Seahawks are on par with the 2011 49ers, the 2006 bears, and the 2006 Ravens.Nick Bradleyhttps://www.blogger.com/profile/07772794102967712253noreply@blogger.com