## Guest Post: Bryan Frye and Career ANY/A+ Value Added

Recently, I posted a quick and dirty method to measure quarterback career value above average and above replacement. I used Adjusted Yards per Pass Attempt as the foundational stat because its inputs (yards, touchdowns, interceptions, and attempts) are on record back to 1932.

Today, I wanted to use the same model with Adjusted Net Yards per Dropback (ANY/A) as the base metric. I believe ANY/A is a more accurate reflection of quarterback production, but it does have the downside of only being recorded back to 1969 in Pro Football Reference’s database.

Thus, while the previous post covered every passer in the official stat era, this post will only cover value added since 1969. This means greats like Sammy Baugh and Sid Luckman are completely overlooked, while legends like Johnny Unitas and Joe Namath only have their worst years included (an unfortunate byproduct of this study’s limitations, to be sure).

In case you didn’t want to click back through the previous article to see the details of the formula, I’ll briefly cover the basics here:

I subtract one hundred from each quarterback’s career ANY/A+ score and divide the difference by fifteen. This gives me the number of standard deviations above average the player’s ANY/A was over the course of his career.1 I then multiply the result by the number of career dropbacks (pass attempts plus sacks) each player has in the data set. The product is the total abstract value above average.

To find value above replacement, I do the same thing but use 85 instead of 100 as the baseline from which to subtract ANY/A. It’s far from rocket science.

The Results

The table displays the abstract career value of the 211 quarterbacks with at least 924 dropbacks since 1969.2 Peyton Manning gets too much love around here, so let’s use Dan Marino as an example: Marino played in 242 games and had 8358 pass attempts and 270 sacks, or 8628 combined dropbacks. His career ANY/A+ score was 119, giving him 10929 arbitrary points of value over average and 19557 such points above replacement.

RkQuarterbackGAttSkDBANY/A+ValRepl
1Peyton Manning266938030396831201291122594
2Dan Marino242835827086281191092919557
4Drew Brees21780853318416115841616832
5Brett Favre3021016952510694108570316397
6Joe Montana19253913135704121798613690
7Dan Fouts18156043195923117671312636
8Steve Young16941493584507123691111418
9John Elway23472505167766107362411390
10Warren Moon20868234587281107339810679
11Philip Rivers16453393255664113490910573
12Aaron Rodgers12640473064353121609410447
13Ben Roethlisberger17154234395862111429910161
14Tony Romo1554331248457911648849463
15Ken Anderson1924475398487311342239096
16Kurt Warner1244070260433011646198949
17Boomer Esiason1875205318552310933148837
18Carson Palmer1605443278572110726708391
19Jim Kelly1604779323510210930618163
20Donovan McNabb1675374410578410623148098
21Eli Manning1856227307653410313077841
22Roger Staubach1312958313327112043617632
23Fran Tarkenton1343854274412811233027430
24Jim Everett1584923257518010620727252
25Trent Green1203740259399911231997198
26Steve McNair1614544254479810722397037
27Troy Aikman1654715259497410619906964
28Jeff Garcia1253676181385711230866943
29Rich Gannon1574206302450810824046912
30Mark Brunell1934640390503010516776707
31Drew Bledsoe1946717467718499-4796705
32Matt Ryan1264530218474810618996647
33Vinny Testaverde2336701417711899-4756643
34Dave Krieg213531149458051027746579
35Jim Hart1734352243459510618386433
37Phil Simms1644647477512410310256149
38Kerry Collins1986261337659898-8805718
39Matt Hasselbeck2095330360569010005690
40Craig Morton1753503373387610718095685
41Steve DeBerg206502429653201013555675
43Steve Grogan1493593247384010615365376
44Matt Schaub1413271178344910818395288
45Daunte Culpepper1053199298349710716325129
46Brian Sipe1253439224366310614655128
47Bernie Kosar1263365273363810614555093
48Ron Jaworski188411736344801025975077
49Bob Griese1362743267301011020075017
50Joe Theismann1673602340394210410514993
51Jay Cutler134435428546391013094948
52Ken Stabler184379328140741038154889
53Randall Cunningham1614289484477310004773
54Neil Lomax1083153362351510511724687
55Jeff George131396735843251012884613
56Roman Gabriel1092666168283410917004534
57Mark Rypien104261397271011018074517
58Joe Ferguson1864519312483199-3224509
59Ken O'Brien129360235339551025274482
60Danny White1662950241319110612764467
61Neil O'Donnell125322925934881049304418
62Billy Kilmer1092388158254611118674413
63Bert Jones1022551232278310814844267
65Tommy Kramer129365124939001012604160
66Matthew Stafford93369120538961012604156
67Steve Beuerlein147332833236601024884148
68Chris Chandler1804005380438599-2924093
69Steve Bartkowski129345635638121012544066
70Joe Flacco1224070257432799-2884039
71Doug Williams88250784259110813823973
72Russell Wilson641735164189911620263925
73Marc Bulger96317125434251024573882
74Jim Zorn140314921333621024483810
75Bobby Hebert118312117832991024403739
76Jake Delhomme103293216831001036203720
77Jim Harbaugh1773918361427998-5713708
78Jake Plummer1434350284463497-9273707
79Bill Kenney106243019526251058753500
80Jeff Blake1193241248348910003489
81Lynn Dickey1523125297342210003422
82Jay Schroeder118280820830161024023418
83Aaron Brooks93296323531981012133411
84Jim McMahon119257322627991035603359
85Andy Dalton77249714026371047033340
86Gus Frerotte1473106207331310003313
87Jim Plunkett1573701380408197-8163265
88Elvis Grbac105244512925741046863260
90Brian Griese93279619329891011993188
91Jon Kitna1414442323476595-15883177
92Cam Newton78241918526041035213125
93Joe Namath85208010921891068763065
94Michael Vick1433217316353398-4713062
95Jeff Hostetler152233820725451035093054
96Doug Flutie92215110722581057533011
97Archie Manning1513642396403896-10772961
98Erik Kramer83229912224211034842905
99Alex Smith1263619325394496-10522892
100Daryle Lamonica68140772147911413802859
101Stan Humphries88251614426601011772837
102David Garrard86228117924601023282788
103John Brodie61141647146311312682731
104Chris Miller982892209310198-4132688
105Scott Mitchell99234615525011011672668
106Greg Landry1422252310256210002562
107Gary Danielson101193218321151034232538
108Charley Johnson90160511817231078042527
109Andrew Luck55210611522211022962517
110Kyle Orton872712165287798-3842493
111Sonny Jurgensen68117894127211411872459
112Ryan Fitzpatrick1133473205367895-12262452
114Richard Todd1192967264323195-10772154
115Jason Campbell902518170268897-5382150
116Mike Tomczak1852337114245198-3272124
117Len Dawson85151016716771044472124
119Tony Eason90156417717411011161857
120Bob Berry6197413111051107371842
121Jay Fiedler761717114183110001831
122Bill Munson6411037411771086281805
123Steve Bono88170176177710001777
124Byron Leftwich60160592169710001697
126Bill Nelsen4410215010711075001571
127Rodney Peete1042346244259094-10361554
128Charlie Batch811604161176598-2351530
129Mike Livingston901751152190397-3811522
130James Harris83114910112501032501500
131Colin Kaepernick571361135149610001496
132Vince Ferragamo75161594170998-2281481
133Nick Foles3912307113011021731474
134Don Majkowski901905180208595-6951390
135Kirk Cousins30950429921063971389
136Tony Banks962356227258393-12051378
137John Friesz69136474143899-961342
138Dan Pastorini1403055246330191-19811320
139Robert Griffin37106310111641021551319
140Ryan Tannehill642248184243293-11351297
141Josh Freeman622048112216094-8641296
142Matt Cassel1002574196277092-14771293
143Craig Erickson521092821174101781252
144Kordell Stewart1252358170252892-13481180
146Shaun Hill46119083127398-1701103
147Gary Hogeboom791325118144396-3851058
148Steve Spurrier78110182118398-1581025
149Johnny Unitas53105774113198-151980
150Bubby Brister992212193240591-1443962
151Brian Hoyer4399961106098-141919
152Marc Wilson1262081210229191-1375916
153Steve Pelluer6597481105598-141914
154Eric Hipple1021546160170693-796910
155Jim Miller37104640108697-217869
156Derek Anderson68154376161993-756863
157Jeff Kemp969166898498-131853
158Josh McCown771956171212791-1276851
159Vince Young60130483138794-555832
160Ty Detmer5494667101397-203810
161Steve Ramsey54921110103196-275756
162Damon Huard6494694104095-347693
164Trent Dilfer1303172263343588-2748687
165Mark Sanchez752267158242589-1778647
166David Woodley58130084138492-738646
167Billy Joe Tolliver791707125183290-1221611
168Frank Reich11893268100094-400600
169Kent Graham811339100143991-863576
170Derek Carr32117255122792-654573
171Rex Grossman54156292165490-1103551
172Steve Walsh80131740135791-814543
173Tarvaris Jackson59107390116392-620543
174Patrick Ramsey379137999293-463529
175Rob Johnson4880614094693-441505
176Tim Couch621714166188089-1379501
177Teddy Bridgewater298498393292-497435
178Mike Pagel1321509115162489-1191433
180Quincy Carter3896080104091-624416
181Vince Evans1001390118150889-1106402
182Kelly Holcomb378937196491-578386
183Joey Harrington812538124266287-2307355
184Trent Edwards389296999890-665333
185Steve Dils10697280105289-771281
186Brandon Weeden3496566103189-756275
187Steve Fuller901066152121888-974244
188Jack Trudeau671644103174787-1514233
189Mark Malone73164886173487-1503231
190Hugh Millen40928126105488-843211
191David Carr942267267253486-2365169
192Bob Avellini731110110122087-1057163
193David Whitehurst5498099107987-935144
194J.P. Losman45952108106087-919141
195Dave Wilson53103999113886-106276
196Dennis Shaw50924124104886-97870
197Blake Bortles301081106118785-11870
198Todd Blackledge468817395484-1018-64
199Dave Brown731634181181584-1936-121
200Joe Pisarcik61898112101083-1145-135
201Scott Brunner73104671111783-1266-149
202Christian Ponder38105795115283-1306-154
203Randy Wright46111985120483-1365-161
204Jack Thompson518458593082-1116-186
205Geno Smith318527492681-1173-247
206Mike Phipps1191799172197183-2234-263
207Kyle Boller671519123164282-1970-328
208Bobby Douglass911178180135881-1720-362
209Danny Kanell4395663101979-1427-408
210Blaine Gabbert37106699116578-1709-544
211Rick Mirer802043199224280-2989-747

As before, I’ll forgo additional commentary and leave that to the esteemed Football Perspective readership.

1. This is known as a z-score, which Chase has used quite a few times on this site. []
2. I chose 924 dropbacks because that is equivalent to six seasons of 154 dropbacks. Danny Tuccitto found 154 dropbacks to be the “mandatory minimum” for comparing QB dropback numbers. I chose six seasons because it seemed like a fair enough timeframe to get rid of the flotsam and jetsam. []
• Topher Doll

I can’t say this enough but Steven Young was insanely efficient. He had three thousand. fewer attempts than Elway but had more value. If he had as many attempts as Peyton Manning he’s have had have almost two thousand more points of value.

• Wonder how many dropbacks Rodgers will have when his replacement value exceeds Young’s. Currently 154 and 971 short, respectively.

• Topher Doll

Rodgers is often compared to Young and for good reason, when he passes this I will be happy for him.

• It depends on how the rest of his career plays out. We know Young’s 4507 dropbacks and 123 ANY/A+ is going to stay the same. We know Rodgers’s dropbacks will increase, but we have no idea what his ANY/A+ will look like in the future. If 2015 was an anomaly and he bounces back for a few more stellar years, he’ll pass Young by mid 2017. However, if 2015 was actually the beginning of the end (I doubt it was, but for the sake of argument), and his ANY/A+ falls, he may never catch Young.

For example, if Rodgers has 3000 more career dropbacks, but his rate stats decline in such a way as to make his career ANY/A+ fall to 114, he’d end up with a Val of 6863, or 48 shy of Young.

I doubt it will happen this way, but that’s a scenario in which Rodgers never reaches Young (or reaches him and subsequently falls behind him).

• I hope it doesn’t happen, but we might be witnessing something similar to what I described in my original reply.

Young was hyper efficient in his prime, but his lofty career ANY/A was saved by retiring before his old man decline seasons.

• Corey

Young had his last real season at 37 and retired at 38 due to concussions — that’s hardly retiring early. His rate stats benefit a little from missing some growing pain seasons sitting on the bench behind Montana, but then he also missed 2 or 3 prime years as well.

Also ANY/A underrates Young because it punishes him for having the higher sack rate of a running QB without crediting him for his rushing value, which in Young’s case was considerable.

Young was also to a certain extent ahead of his time — can you imagine 1992-1994 Young playing in a modern offense optimized for a running QB, with zone-reads and run/pass options and all the rest of it?

Good points. I guess a confluence of factors cancel each other out with regard to Young’s stats. His stint in Tampa drags down his career numbers, and he wasted several prime years sitting behind Montana (Young’s rate stats were excellent in his rare late 80’s appearances). I didn’t mean to imply that he retired early per se; rather he retired before his play significantly declined, which happens to most great QB’s. Young was awful in his four starts in 1999 before the career ending concussion, so it’s reasonable to believe that his numbers would have cratered if he had played another few years.

• Most of them really only have one of those bad ending seasons, at most, and quit right? At least since the 80’s. Elway and Montana had none while Marino and Manning had one. Favre ended up having one, though it would have been two had he been done after 06.

Digging deeper, you’re right about modern QB’s usually having only one bad season at the end of their careers. However in earlier eras QB’s tended to hang on for several seasons way past their prime (Stabler and Namath are prime examples). My guess is that modern conditioning and medicine allow QB primes to last into their mid to late 30’s, unlike the old days when guys were physically broken at 33 or 34.

I’m fascinated to see how Brady’s final years unfold. He wants to play until he’s 45, but will he hang on even after his body breaks down?

• The sack rates of Marino and Manning can’t be harped on enough. Just amazing.

• Richie

I always wondered if Marino tried too hard to avoid sacks and threw more interceptions because of it. But it looks like only Peyton Manning (5.8) has a lower Sack%+INT% than Marino (6.1). Guys like Aaron Rodgers and Steve Young have higher numbers just for their Sack %.

• I was looking for another comment I made on either this or another article and stumbled on this one. I recently did a study on all QBs from 1992-2015 and found that the idea of exchanging sacks for interceptions isn’t really supported by the data (at least not in general; individual cases always apply). Quarterbacks who are good at avoiding sacks tend to be pretty good at avoiding interceptions as well. Quarterbacks who eat too many sacks also tend to throw too many picks. They also happen to wash out of the league quickly enough for us to forget about them. It turns out that, in general, good QBs are good because the are good at lots of QB things. Bad QBs tend to be bad at a lot of QB things. Anecdotal arguments like Ben, Alex Smith, and Rodgers have taken primacy in our heads and helped us believe their examples are representative of everyone’s.

I was a little surprised when I found this, especially since I did the study in order to prove the narrative correct. I always enjoy the feeling of being completely wrong but learning something new and exciting.

Philip Rivers and Tony Romo are still criminally underrated.

• sacramento gold miners

I’ll take Dan Fouts over Rivers and Troy Aikman over Romo any day of the week. If Fouts and Aikman were playing today, I think they would put up better numbers.

I’d give a slight edge to Fouts over Rivers, but I think Romo has already surpassed Aikman. Romo can create big gains from broken plays, something Aikman rarely did. Also until recently Romo played behind mediocre offensive lines while Aikman had the benefit of Larry Allen and company.

• sacramento gold miners

We’re going to have to agree to disagree here. Aikman had the better arm, and was better in the big moments. If we put #8 with the recent Cowboy playoff teams, I think they advance further in the playoffs.

• Richie

Aikman is tough for me to evaluate. It seems like he had it so easy for his career, with a great supporting cast (on both sides of the ball). He lost Irvin in 1999, and had an OK season. Then Aikman got hurt in 2000 and retired. Emmitt Smith was still going strong at that point. The other season where he lost Irvin for a bunch of games (96) was one of his least efficient seasons.

Aikman was good, but I just don’t remember him ever having to prove himself by carrying the team like most of the other great QB’s had to do at one time or another. Aikman and Montana (for example) each started about the same number of games for their careers, but Montana had twice as many 4th quarter comebacks.

I don’t know if I’ve ever seen an analysis, but has any QB ever had a better supporting cast for their career than Aikman?

In the modern era, I don’t think anyone has had a better supporting cast than Aikman. As you said, he had great teammates at almost every position for most of his career. Perhaps Montana’s cast comes close, with his elite skill positions and defenses, but his o-lines weren’t anything special. I agree that Aikman never had to carry his team, and frankly I’m not sure if he deserves to be in the HoF.

• Among Hall of Fame QBs, Otto Graham, Sid Luckman, Norm Van Brocklin, Johnny Unitas, Bart Starr, Len Dawson, Terry Bradshaw, Joe Montana, and Steve Young all had arguably comparable (or better) supporting casts. Graham and Starr probably had the greatest supporting casts of any QB in history.

Young may not have had 100% of Aikman’s support, but he certainly was close, in my mind. In 1991, he was on a team that managed to go 10-6 when he was 5-5 as a starter (5-1 behind Bono). They were third in points scored and fourth in points allowed. Then they led the league in points scored four years in a row, while their defense ranked 3, 16, 6, and 2. He broke the huddle with Jerry Rice and Ricky Watters, which is a comparable duo to Irvin/Smith. Brent Jones was comparable to Novacek too, I think. And, while they didn’t have a fancy name like “The Great Wall,” Wallace, McIntyre, Sapolu, and Barton were good to great linemen (not on the level of Erik Williams or Larry Allen, of course). Young hardly got a bad lot in life. He was no early career Elway or late career Marino, that’s for sure.

(I say all of this having grown up with Montana, Young, and Rice posters decorating my bedroom wall…and Lyle Alzado, who was my mom’s favorite player.)

As your list illustrates, all of the QB’s with stacked supporting casts played before or in the infancy of the salary cap. Even after normalizing stats, this makes today’s elite QB’s appear even more impressive, given the constant roster churn they have to deal with.

• I agree. I know Favre takes a lot of crap around here, but the guy had some seasons where his supporting cast was downright pedestrian. If he had the offensive teammates that Trent Green had in the early 2000s, I doubt Green ever cracks the top 50 of these lists, while Favre shoots up a few spots. Same goes for McNabb, who had 21 games’ worth of play with Terrell Owens and put up Peyton Manning type numbers in those games.

For the most part, QBs who have great stats and win a lot of games don’t get there without help. It takes some remarkable play to make the HOF with a supporting cast like Sonny Jurgensen’s.

Three straight MVP’s with Antonio Freeman as his #1 target is damn impressive. I’ve knocked Favre in the past for not having a high enough peak, but that was a bit misguided given that his numbers during that stretch don’t do his play justice.

• Richie

Yeah, this is why I begrudgingly have to keep moving Tom Brady up my list of best QB’s. He’s had some seasons with pretty weak weapons on offense. (He’s also had some pretty good ones.) But they definitely change often.

Agreed. Brady’s 01-06 seasons look more impressive in retrospect, given that his receivers from those teams accomplished zero without him.

• Richie

Especially 04-06. From 01-03 Brady’s stats were basically like Ryan Tannehill. It’s 04 and on where he started putting up All-Pro/HOF numbers.

I actually consider 04-06 the most impressive stretch of Brady’s career. Almost anyone looks good throwing to Moss or Gronk, but it takes a special QB to be great with Reche Caldwell and Deion Branch.

• WR

Your position on Brady v. Tannehill is off the mark. I would argue that Brady’s worst season was significantly better than Tannehill’s best season, so far. Tannehill has an ANYPA+ figure of 93, with a career high of 96 in 2014. Brady through 2003 was at 105, with year-by-year figures of 102, 105, and 107.

I”ll agree that Brady has been much better since 2004, but the idea that he was ever a statistical weakling is wrong. And of course, if you put any weight on team achievements, Brady blows Tannehill out of the water. I’ve studied Brady’s career closely, and I think we can pinpoint the game in which he took a step forward statistically. It was the game in Miami about halfway through 2003, when he threw an OT TD pass to Troy Brown. Ever since then, Brady has been performing at an elite level, and it’s not coincidence that the Miami game was near the beginning of the 21 game win streak.

The myth that Brady had weak numbers at the start of his career, like the claim that Manning usually plays poorly in the playoffs, doesn’t stand up to scrutiny, and is based on impressions people formed years ago.

• Richie

Or, Young was no early career Young (on the Bucs).

• Dr__P

Folks might not say the number of concussions that Aikman had was an indicator of how easy he had it in his career

• Richie

Every QB gets sacked/hit regardless of how good his teammates are.

• Dr__P

Yes every QB gets sacked, but not everyone retires earlier due to repeated head and back injuries. The point was that he did NOT have an easy early part of his career.

• Richie

Yeah, he was on a pretty bad team in 1989. They were better in 1990, and by 1991 they were pretty much THE COWBOYS. Yeah, he got sacked a little more frequently in his early seasons. But he was above average at avoiding sacks for most of his career.

Those extra sacks he took in 1990 and 1991 aren’t specifically the reason he had to retire early. He was sacked 259 times in his career, plus probably took at least another 259 hits. That’s normal. Unfortunately, one or two of those hits may have been a little harder than average, or hit just the right spot. Or maybe Aikman had a predisposition to the effects of getting hit.

But if you are saying that Aikman had to retire early because he was playing on teams with bad offensive lines that were worse at protecting him than other good teams in the NFL, I just don’t see any evidence that is the case.

• It is hard for me to separate Fouts from A. Don Coryell, B. the Mel Blount Rule, and C. Charlie Joiner/John Jefferson/Kellen Winslow/Wes Chandler. Like you, the edge I give to Fouts over Rivers is only a slight one.

I think Aikman played admirably within the structure of his team, and I think he lost quite a few career TD passes to Emmitt Smith runs, whereas Steve Young had great goal line opportunities (of course part of that is Young’s running ability making him the perfect bootleg goal line QB). Aikman didn’t have the improvisational skill of Romo, but he did play remarkably well when his team needed a little more from him in the playoffs. I’m fine saying Romo is the more intrinsically talented QB, while Aikman had the superior career.

I think of Aikman as a poor man’s Brady – similar playing styles and postseason success, but Aikman was not quite as good and didn’t last nearly as long.

I have no rational reason for putting Fouts ahead of Rivers, it’s just my gut feel.

• I think of Aikman and Brady as actual rich men.

• Topher Doll

Both are hurt by people who only judge QB’s by Super Bowl success but both are just far and way better than most give them credit for and if you look purely at their passing numbers they stack up well with the best.

• I mentioned this with AY/A, too, but I’d love to see a refinement of this methodology that smoothes is out across eras. By both methods (Val and Repl), three of the top four, and 7 of the top 14, started last season. This is interesting and potentially useful data, but we shouldn’t take it to imply that (for instance) Philip Rivers is the 10th- or 11th-best quarterback of all time.

• I quit checking the comments on my guest posts after a few days, so I completely missed the one you made in the AY/A post. Sorry about that; I’m glad you were patient enough to bother writing another comment in this post.

I agree with your assessment here (and in the previous comment). I tried to shy away from calling either post a “greatest of all time” or “greatest since the merger” measurement, since that’s not what they are at all. Your concerns are just one of the reasons that is the case (no accounting for supporting cast, coaching, weather, and SOS, to name a few).

That said, I don’t have any immediate plans to account for any of that because I don’t think the results would be dissimilar to Chase’s Adjusted Dropbacks version of the GQBOAT series. The only real difference would be that he is using actual ANY/A relative to average, and I am using standard deviations of ANY/A relative to average.

On my own site, I do plan to use TAY/P (and its derivatives) to go back to 1932 for QB measurements, even though QBs weren’t really QBs then. I plan on prorating schedules, and I have a few methods in mind for adjusting for the increased volume of QB plays in the modern era.

I also have a few models that attempt to account for the strength of the league as a whole, but those ding pre-merger players pretty severely.

• Would love to see a more in depth analysis with rushing yards, rushing touchdowns, and fumbles incorporated. I have a feeling Steve Young would look even better.