Looking at Flacco's numbers from a couple different angles

[callahan09]

I was just thinking about what aspects of a QB's play are generally associated with winning? Number one for me is not turning the ball over, so I knew interceptions had to play a major role whatever I looked at. But then it came down to yards and touchdowns... Yards per attempt is important because you're either moving the ball down the field, or you're not. Touchdowns seem important at first glance, but perhaps they aren't as important as yards per attempt, because as long as you're moving the ball down field, you're not only limiting the opponent's chance to score, but you're setting your team up to score as well, even if that doesn't translate to a passing touchdown to pad your QB stats. A high yards per attempt and low turnover ratio probably means you are putting your team in position to score, whether it's a field goal, rushing TD, or even a passing TD, and it means you're keeping the opponent offense off the field. Any flaw in this logic?

So I decided to take a look at Yards Per Attempt divided by Interception Percentage. That's (Y/A)/(INT/A), which simplifies to (Y/A) [or, simply, yards per interception].

All of these rankings are based on players with 500+ career attempts. I'll just list the top 20 (this is out of 297 players, since 1960):



1. Aaron Rodgers: 1945/2957, 23973 Yards, 189 TD, 51 INT = 470.1 Y/INT
2. Tom Brady: 4351/6845, 50755 Yards, 376 TD, 145 INT = 350 Y/INT
[b]3. Joe Flacco: 1714/2862, 20305 Yards, 121 TD, 64 INT = 317.3 Y/INT[/b]
4. Matt Schaub: 1879/2912, 22549 Yards, 122 TD, 72 INT = 313.2 Y/INT
5. Neil O'Donnell: 2024/3504, 23399 Yards, 129 TD, 76 INT = 307.9 Y/INT
6. Donovan McNabb: 3501/5930, 40859 Yards, 258 TD, 134 INT = 304.9 Y/INT
7. Steve Young*: 2959/4620, 36450 Yards, 252 TD, 120 INT = 303.8 Y/INT
8. Matt Ryan: 1779/2825, 20197 Yards, 136 TD, 67 INT = 301.4 Y/INT
9. Jeff Garcia: 2390/3893, 26894 Yards, 168 TD, 90 INT = 298.8 Y/INT
10. Philip Rivers: 2402/3793, 29711 Yards, 197 TD, 102 INT = 291.3 Y/INT
11. Joe Montana*: 3869/6125, 46323 Yards, 318 TD, 160 INT = 289.5 Y/INT
12. Drew Brees: 4297/6541, 48899 Yards, 346 TD, 169 INT = 289.3 Y/INT
13. David Garrard: 1440/2343, 16489 Yards, 92 TD, 57 INT = 289.3 Y/INT
14. Tony Romo: 2177/3375, 26569 Yards, 181 TD, 93 INT = 285.7 Y/INT
15. Mark Brunell: 2917/4947, 33905 Yards, 195 TD, 119 INT = 284.9 Y/INT
16. Peyton Manning: 5563/8554, 65166 Yards, 468 TD, 230 INT = 283.3 Y/INT
17. Jason Campbell: 1328/2182, 14682 Yards, 76 TD, 52 INT = 282.3 Y/INT
18. Shaun Hill: 591/954, 6381 Yards, 41 TD, 23 INT = 277.4 Y/INT
19. Sam Bradford: 873/1498, 9378 Yards, 45 TD, 34 INT = 275.8 Y/INT
20. Cam Newton: 590/1002, 7920 Yards, 40 TD, 29 INT = 273.1 Y/INT

But, since a lot of people like to talk about touchdowns, I wanted to come up with a way to incorporate that stat as well. So this time I took the TD/INT ratio and multiplied it by Y/A. I don't really have a name for this, but it looks like an effective efficiency rating to me. Call it what you will (suggestions would be wonderful):


1. Aaron Rodgers: 1945/2957, 23973 Yards, 189 TD, 51 INT = 30.04 TD/INT * Y/A
2. Tom Brady: 4351/6845, 50755 Yards, 376 TD, 145 INT = 19.23 TD/INT * Y/A
3. Steve Young*: 2959/4620, 36450 Yards, 252 TD, 120 INT = 16.57 TD/INT * Y/A
4. Peyton Manning: 5563/8554, 65166 Yards, 468 TD, 230 INT = 15.5 TD/INT * Y/A
5. Tony Romo: 2177/3375, 26569 Yards, 181 TD, 93 INT = 15.32 TD/INT * Y/A
6. Drew Brees: 4297/6541, 48899 Yards, 346 TD, 169 INT = 15.31 TD/INT * Y/A
7. Philip Rivers: 2402/3793, 29711 Yards, 197 TD, 102 INT = 15.13 TD/INT * Y/A
8. Joe Montana*: 3869/6125, 46323 Yards, 318 TD, 160 INT = 15.03 TD/INT * Y/A
9. Matt Ryan: 1779/2825, 20197 Yards, 136 TD, 67 INT = 14.51 TD/INT * Y/A
10. Kurt Warner: 2973/4532, 36296 Yards, 239 TD, 142 INT = 13.48 TD/INT * Y/A
[b]11. Joe Flacco: 1714/2862, 20305 Yards, 121 TD, 64 INT = 13.41 TD/INT * Y/A[/b]
12. Ben Roethlisberger: 2622/4171, 32994 Yards, 211 TD, 125 INT = 13.35 TD/INT * Y/A
13. Donovan McNabb: 3501/5930, 40859 Yards, 258 TD, 134 INT = 13.27 TD/INT * Y/A
14. Matt Schaub: 1879/2912, 22549 Yards, 122 TD, 72 INT = 13.12 TD/INT * Y/A
15. Jeff Garcia: 2390/3893, 26894 Yards, 168 TD, 90 INT = 12.9 TD/INT * Y/A
16. Billy Volek: 340/565, 3802 Yards, 27 TD, 15 INT = 12.11 TD/INT * Y/A
17. Dan Marino*: 5352/9045, 65871 Yards, 452 TD, 276 INT = 11.93 TD/INT * Y/A
18. Shaun Hill: 591/954, 6381 Yards, 41 TD, 23 INT = 11.92 TD/INT * Y/A
19. Rich Gannon: 2687/4446, 30434 Yards, 191 TD, 113 INT = 11.57 TD/INT * Y/A
20. David Garrard: 1440/2343, 16489 Yards, 92 TD, 57 INT = 11.36 TD/INT * Y/A


One thing you'll notice is that Aaron Rodgers seems to be in a league of his own in both stats. He is 34% higher than the 2nd ranked player in Y/INT, and 56% higher in the efficiency score. That's pretty crazy. Rodgers and Brady represent #1 and #2 by both measures.

How does this compare to traditional QB Rating, then (which takes completion percentage into account)?


1. Aaron Rodgers: 1945/2957, 23973 Yards, 189 TD, 51 INT = 104.8 QB Rating
2. Steve Young*: 2959/4620, 36450 Yards, 252 TD, 120 INT = 95.7 QB Rating
3. Tom Brady: 4351/6845, 50755 Yards, 376 TD, 145 INT = 95.4 QB Rating
4. Peyton Manning: 5563/8554, 65166 Yards, 468 TD, 230 INT = 95.1 QB Rating
5. Tony Romo: 2177/3375, 26569 Yards, 181 TD, 93 INT = 95 QB Rating
6. Drew Brees: 4297/6541, 48899 Yards, 346 TD, 169 INT = 94.8 QB Rating
7. Kurt Warner: 2973/4532, 36296 Yards, 239 TD, 142 INT = 94.6 QB Rating
8. Philip Rivers: 2402/3793, 29711 Yards, 197 TD, 102 INT = 93.6 QB Rating
9. Joe Montana*: 3869/6125, 46323 Yards, 318 TD, 160 INT = 92.7 QB Rating
10. Ben Roethlisberger: 2622/4171, 32994 Yards, 211 TD, 125 INT = 91.8 QB Rating
11. Matt Schaub: 1879/2912, 22549 Yards, 122 TD, 72 INT = 91.8 QB Rating
12. Matt Ryan: 1779/2825, 20197 Yards, 136 TD, 67 INT = 90.5 QB Rating
13. Chad Pennington: 1764/2687, 19241 Yards, 110 TD, 72 INT = 89.1 QB Rating
14. Daunte Culpepper: 2089/3333, 25133 Yards, 157 TD, 111 INT = 87.6 QB Rating
15. Jeff Garcia: 2390/3893, 26894 Yards, 168 TD, 90 INT = 86.8 QB Rating
[b]16. Joe Flacco: 1714/2862, 20305 Yards, 121 TD, 64 INT = 86.3 QB Rating[/b]
17. Bart Starr*: 1649/2812, 22810 Yards, 148 TD, 109 INT = 86.1 QB Rating
18. Brett Favre: 6781/10960, 77693 Yards, 552 TD, 366 INT = 86.1 QB Rating
19. Carson Palmer: 2587/4147, 29677 Yards, 190 TD, 131 INT = 86 QB Rating
20. Shaun Hill: 591/954, 6381 Yards, 41 TD, 23 INT = 85.9 QB Rating

Well, the top 6 players are all the same, just with Steve Young and Tom Brady flipped, so it's a rather similar measure. 9 of 10 players appear in both top 10's, with the exception being Matt Ryan appearing on my ranking, and Ben Roethlisberger on the QB Rating ranking. So since the primary noticeable differences seem to be that my score ranks Brady over Young and Ryan over Roethlisberger, I'll do a comparison between them, and let you decide which score ranks the players more appropriately.

First, Brady vs Young:

Completion Percentage: 63.6% vs 64.0%
Yards Per Attempt: 7.41 vs 7.89
Passes Per TD: 18 vs 18
Passes Per INT: 47 vs 38

Adjusting Young's attempts up to Brady's (6845), Young would be expected to complete 33 more passes, gain 3249 more yards, throw 3 fewer touchdowns, and 33 more interceptions.

Essentially, to make the transition from Brady to Young, you'd be trading 3 touchdowns for 3000 yards and 33 interceptions. From my perspective, if someone told me I could trade in my quarterback for another who would give me 3000 more yards, but 33 more interceptions and 3 fewer touchdowns over their career, I'd decide to stick with what I've got. That's my eyeball test, and to me, I feel Brady deserves to be ranked above Young based purely on those numbers. Feel free to comment on this with your own perspective.

Next, Ryan vs Roethlisberger:


Completion Percentage: 63.0% vs 62.9%
Yards Per Attempt: 7.15 vs 7.91
Passes Per TD: 21 vs 20
Passes Per INT: 42 vs 33

Adjusting Ryan's attempts up to Roethlisbergers (4171), Ryan would be expected to complete 5 more passes, gain 3174 fewer yards, throw 10 fewer touchdowns, and 26 fewer interceptions.

Basically, making the transition from Ryan to Roethlisberger would mean you'd be gaining 3000 more yards and 10 more touchdowns, but at the cost of 26 more interceptions. This one isn't as clear cut to me as the Brady/Young comparison, because we're not only gaining more yards, but also more touchdowns in exchange for all those interceptions. But I still personally like my ranking better, because if I'm really given this choice, that 10 more TDs for 26 more INTs figure doesn't look appealing at all. INTs lose games a lot more than TDs win them, because once you've thrown that INT, you've taken away your teams chance to score as well as giving the opponent a chance to score. Whereas merely *not* throwing a touchdown doesn't mean your team can't score by means of field goal or rushing attempt, and even without scoring at all, at least you aren't aiding the other team in the field position and opportunity game the way an interception does.

In a way, I feel that interceptions aren't weighted enough in most QB performance metrics these days. Perhaps Joe Flacco ranks among the greatest of all time in terms of [i]winning games[/i], because he has such a great Yards/INT ratio as I demonstrated above. Throwing interceptions is one of the surefire ways to lose a game for your team, and protecting the football is a great way to stay in control and win the game, all other factors independent...

Anyway, that's what I've got for now.

Thoughts? Notes? Criticisms?

[BallHawker]

Completely agree; been saying this all along. You should apply for a job to be an NFL analyst, my friend. Only criticism is that you also need to account for fumbles

Also, I hope you created a script to do all the typing for you.. Even using a spreadsheet, it would take so long!

[BALTIMOREFLACCO]

Espn needs to hire you sir

[callahan09]

Hah, yeah, I do write scripts to crunch the numbers, sort, and output in a readable format, so it doesn't take me too long to put together these rankings.

Since this topic title doesn't just imply discussion of Y/INT and that new efficiency model I created, I think it's a good place to post other aspects and interpretations of Flacco's numbers.

I've got another one:

Since 2010, we can split Flacco's performances up by first half of the season, second half, and post-season. For the second half numbers, we're going to have one fewer game than the first half because I'm not counting the 2012 Week 17 game where he essentially didn't play.

[u]First 8 Games[/u]
18-6 Record
494/848 (58.3%), 5958 Yards (7.03 Y/A), 31 TD, 18 INT (1.72 TD:INT), 83.2 QB Rating

[u]Last 8 Games[/u]
16-7 Record
437/706 (61.9%), 5057 Yards (7.16 Y/A), 36 TD, 14 INT (2.57 TD:INT), 92.2 QB Rating

[u]Post-Season[/u]
6-2 Record
150/253 (59.3%), 2012 Yards (7.95 Y/A), 18 TD, 2 INT (9.00 TD:INT), 105.0 QB Rating

And here's the progression of passer rating from first half, to second half, to post-season over that 3 year span:

[img]http://i50.tinypic.com/2prahqq.gif[/img]

[hawkprey]

Interesting. Here's my analysis...
[media]http://www.youtube.com/watch?v=XvtgUu_z5TI&feature=youtu.be[/media]

[Purple Nurple]

My head hurts.

[callahan09]

[quote name='Purple Nurple' timestamp='1363026234' post='1385344']
My head hurts.
[/quote]

I'm sorry...

I do notice one thing with the efficiency rating that I didn't consider before... It counts yards on a per attempt basis, but not touchdowns & interceptions, which are only counted as a ratio between each other, but not as a percentage of pass attempts. So given the same number of attempts and the same number of yards, two totally different TD:INT numbers will result in the same exact rating as long as their ratio is the same. For example, let's look at Flacco's average season over the past 4 years:

313/515, 3666 Yards, 22 TD, 11 INT, 87.7 QB Rating

That would be 7.12 Y/A and a 2:1 TD:INT ratio, which comes out to 14.24 points based on my efficiency model.

Now, imagine he had exactly the same number of attempts, and yards, but double the touchdowns and interceptions. Now his numbers look like this:

313/515, 3666 Yards, 44 TD, 22 INT, 93.1 QB Rating

He will still have the same individual components in the efficiency model formula: 7.12 Y/A and 2:1 TD:INT ratio, giving him the exact same 14.24 points score. In terms of passer rating, this would bring him up from an 87.7 to a 93.1. So I'm not sure if there's some tweaking I should do to incorporate the TD & INT numbers in a way that it's not purely based on ratio.

I can't decide, and would like some input.

Would you personally take 22 more touchdowns in exchange for 11 more interceptions? I'm torn... 11 more interceptions for a total of 22 is pretty awful, and would really hurt the team... On the other hand, 22 more touchdowns is a lot of points. However with the same number of yards, is it likely that without those 22 touchdowns the team could be scoring in other ways, and with 11 fewer INTs preventing the other team from scoring, thus kind of evening out to actually be of about equal quality in performance? I don't know. 44 touchdowns sure looks a lot nicer than 22, but 22 INTs is also an awful lot, and would have led the league in INTs by 3 this season. Drew Brees this year had a ton of touchdowns and a ton of INTs, and it didn't do his team any good winning more games than Flacco who had about half the TDs and half the INTs, so maybe the bulk isn't necessarily the better, and the ratio is the most important aspect? I don't know. I'm open to differing opinions on the matter.

[Purple Nurple]

No need to apologize, dude. That's just a lot to absorb. You're good!

[hawkprey]

[quote name='callahan09' timestamp='1363030396' post='1385419']
I'm sorry...

I do notice one thing with the efficiency rating that I didn't consider before... It counts yards on a per attempt basis, but not touchdowns & interceptions, which are only counted as a ratio between each other, but not as a percentage of pass attempts. So given the same number of attempts and the same number of yards, two totally different TD:INT numbers will result in the same exact rating as long as their ratio is the same. For example, let's look at Flacco's average season over the past 4 years:

313/515, 3666 Yards, 22 TD, 11 INT, 87.7 QB Rating

That would be 7.12 Y/A and a 2:1 TD:INT ratio, which comes out to 14.24 points based on my efficiency model.

Now, imagine he had exactly the same number of attempts, and yards, but double the touchdowns and interceptions. Now his numbers look like this:

313/515, 3666 Yards, 44 TD, 22 INT, 93.1 QB Rating

He will still have the same individual components in the efficiency model formula: 7.12 Y/A and 2:1 TD:INT ratio, giving him the exact same 14.24 points score. In terms of passer rating, this would bring him up from an 87.7 to a 93.1. So I'm not sure if there's some tweaking I should do to incorporate the TD & INT numbers in a way that it's not purely based on ratio.

I can't decide, and would like some input.

Would you personally take 22 more touchdowns in exchange for 11 more interceptions? I'm torn... 11 more interceptions for a total of 22 is pretty awful, and would really hurt the team... On the other hand, 22 more touchdowns is a lot of points. However with the same number of yards, is it likely that without those 22 touchdowns the team could be scoring in other ways, and with 11 fewer INTs preventing the other team from scoring, thus kind of evening out to actually be of about equal quality in performance? I don't know. 44 touchdowns sure looks a lot nicer than 22, but 22 INTs is also an awful lot, and would have led the league in INTs by 3 this season. Drew Brees this year had a ton of touchdowns and a ton of INTs, and it didn't do his team any good winning more games than Flacco who had about half the TDs and half the INTs, so maybe the bulk isn't necessarily the better, and the ratio is the most important aspect? I don't know. I'm open to differing opinions on the matter.
[/quote]
You have to decide how you're looking at this. Are you doubling someone's TD + INT/ attempt? Or are you adding an extra 2:1 ratio? Because it just so happens that a 2:1 ratio is the median point for what you expect out of a starter. In Joe's case it doesn't really do much because he's already sitting at 2:1. But add 22 TDs and 11 INTs to Ryan Tannehill, for example, and that's a much better stat line for him. Double his 12 TDs and 13 INTs and now he looks much worse. That's why you're having trouble deciding which you want for Joe.

[BloodRaven]

It's still preferable to be near the top of both lists, especially when you're paid more than anyone else.

[pep_meister]

So flacco shouldn't have been paid like Brady and potentially more than rogers?

[callahan09]

Here's another graph on the breakdown of games throughout the season. This is his average passer rating in games 1-8, 9-16, and post-season games, for 2010, 2011, and 2012, shown individually on the graph this time, and then with the curve fitting applied.

[IMG]http://i50.tinypic.com/2i5f6o.png[/IMG]

[callahan09]

[quote name='pep_meister' timestamp='1363056225' post='1387777']
So flacco shouldn't have been paid like Brady and potentially more than rogers?
[/quote]

Why shouldn't he have been paid like Brady? He is right up there with him on this list, and keep in mind that this is Flacco in his first five years, compared to Brady as a veteran in his prime years over these past 5 seasons. Also, we're paying Flacco for what we expect him to do, not what he has already done. Additionally, Brady already received his monster pay-day, and now he is looking for smaller overall money in exchange for big guaranteed money to finish out his career, so to say Flacco doesn't deserve to paid like Brady is a bit ridiculous. The two have led arguably the most similar careers two quarterbacks could have in NFL history. Their passer rating through their first five years is about dead even, their interception rates are very similar, their win rates are almost even, playoff wins are equal (though Flacco did it almost entirely on the road, while Brady has played almost every playoff game of his career at home and usually got all the help he needed from his teammates, where we know Flacco has occasionally been let down)

How about this, I'll post the rankings for all active QBs' first five seasons (with at least 500 attempts total in their first 5 seasons):

Yards Per INT:

1. Aaron Rodgers: 754/1178, 9224 Yards, 63 TD, 22 INT = 419.3 Y/INT
2. Joe Flacco: 1714/2862, 20305 Yards, 121 TD, 64 INT = 317.3 Y/INT
3. Matt Ryan: 1779/2825, 20197 Yards, 136 TD, 67 INT = 301.4 Y/INT
4. Tom Brady: 1433/2322, 15876 Yards, 108 TD, 55 INT = 288.7 Y/INT
5. Jason Campbell: 1196/1966, 13247 Yards, 68 TD, 46 INT = 288 Y/INT
6. Philip Rivers: 997/1617, 12219 Yards, 85 TD, 43 INT = 284.2 Y/INT
7. Matt Hasselbeck: 794/1327, 9389 Yards, 50 TD, 34 INT = 276.1 Y/INT
8. Sam Bradford: 873/1498, 9378 Yards, 45 TD, 34 INT = 275.8 Y/INT
9. Cam Newton: 590/1002, 7920 Yards, 40 TD, 29 INT = 273.1 Y/INT
10. Shaun Hill: 579/938, 6176 Yards, 39 TD, 23 INT = 268.5 Y/INT
11. Kyle Orton: 1134/1952, 12774 Yards, 71 TD, 48 INT = 266.1 Y/INT
12. Matt Schaub: 527/830, 6317 Yards, 30 TD, 25 INT = 252.7 Y/INT
13. Andrew Luck: 367/681, 4662 Yards, 23 TD, 19 INT = 245.4 Y/INT
14. Byron Leftwich: 839/1433, 9500 Yards, 52 TD, 39 INT = 243.6 Y/INT
15. Matt Cassel: 620/1048, 6870 Yards, 39 TD, 29 INT = 236.9 Y/INT
16. Matthew Stafford: 1142/1906, 13187 Yards, 83 TD, 56 INT = 235.5 Y/INT
17. Drew Brees: 1156/1851, 12667 Yards, 82 TD, 54 INT = 234.6 Y/INT
18. Carson Palmer: 1381/2166, 15696 Yards, 107 TD, 67 INT = 234.3 Y/INT
19. Tony Romo: 866/1372, 10952 Yards, 83 TD, 47 INT = 233 Y/INT
20. Michael Vick: 784/1445, 9640 Yards, 54 TD, 42 INT = 229.5 Y/INT
21. Blaine Gabbert: 372/691, 3876 Yards, 21 TD, 17 INT = 228 Y/INT
22. Charlie Batch: 747/1334, 9063 Yards, 49 TD, 40 INT = 226.6 Y/INT
23. Andy Dalton: 670/1116, 7451 Yards, 47 TD, 33 INT = 225.8 Y/INT
24. Colt McCoy: 409/702, 4388 Yards, 21 TD, 20 INT = 219.4 Y/INT
25. Ben Roethlisberger: 1361/2183, 17213 Yards, 116 TD, 81 INT = 212.5 Y/INT
26. Peyton Manning: 1799/2922, 21176 Yards, 139 TD, 102 INT = 207.6 Y/INT
27. David Carr: 1243/2070, 13391 Yards, 59 TD, 65 INT = 206 Y/INT
28. Josh Freeman: 1101/1873, 12963 Yards, 78 TD, 63 INT = 205.8 Y/INT
29. Jay Cutler: 1380/2249, 16318 Yards, 106 TD, 80 INT = 204 Y/INT
30. Matt Moore: 448/758, 5268 Yards, 33 TD, 26 INT = 202.6 Y/INT
31. Matt Leinart: 357/618, 4011 Yards, 15 TD, 20 INT = 200.6 Y/INT
32. Brandon Weeden: 297/517, 3385 Yards, 14 TD, 17 INT = 199.1 Y/INT
33. Eli Manning: 1389/2477, 15920 Yards, 106 TD, 81 INT = 196.5 Y/INT
34. Chad Henne: 812/1373, 9198 Yards, 42 TD, 48 INT = 191.6 Y/INT
35. Christian Ponder: 458/774, 4788 Yards, 31 TD, 25 INT = 191.5 Y/INT
36. Mark Sanchez: 1123/2024, 13247 Yards, 77 TD, 72 INT = 184 Y/INT
37. Kevin Kolb: 340/572, 4037 Yards, 20 TD, 22 INT = 183.5 Y/INT
38. Tarvaris Jackson: 369/638, 4148 Yards, 24 TD, 23 INT = 180.3 Y/INT
39. Brady Quinn: 296/550, 3043 Yards, 12 TD, 17 INT = 179 Y/INT
40. Alex Smith: 864/1514, 9399 Yards, 51 TD, 53 INT = 177.3 Y/INT
41. Tyler Thigpen: 272/504, 3192 Yards, 21 TD, 18 INT = 177.3 Y/INT
42. Bruce Gradkowski: 362/680, 3883 Yards, 20 TD, 23 INT = 168.8 Y/INT
43. Derek Anderson: 756/1436, 9148 Yards, 53 TD, 55 INT = 166.3 Y/INT
44. Ryan Fitzpatrick: 424/734, 4104 Yards, 21 TD, 27 INT = 152 Y/INT
45. John Skelton: 320/602, 3707 Yards, 15 TD, 25 INT = 148.3 Y/INT

As you can see, Flacco is above Brady here.

And in efficiency:


1. Aaron Rodgers: 754/1178, 9224 Yards, 63 TD, 22 INT = 22.42 TD/INT * Y/A
2. Philip Rivers: 997/1617, 12219 Yards, 85 TD, 43 INT = 14.94 TD/INT * Y/A
3. Matt Ryan: 1779/2825, 20197 Yards, 136 TD, 67 INT = 14.51 TD/INT * Y/A
4. Tony Romo: 866/1372, 10952 Yards, 83 TD, 47 INT = 14.1 TD/INT * Y/A
5. Tom Brady: 1433/2322, 15876 Yards, 108 TD, 55 INT = 13.43 TD/INT * Y/A
6. Joe Flacco: 1714/2862, 20305 Yards, 121 TD, 64 INT = 13.41 TD/INT * Y/A
7. Carson Palmer: 1381/2166, 15696 Yards, 107 TD, 67 INT = 11.57 TD/INT * Y/A
8. Ben Roethlisberger: 1361/2183, 17213 Yards, 116 TD, 81 INT = 11.29 TD/INT * Y/A
9. Shaun Hill: 579/938, 6176 Yards, 39 TD, 23 INT = 11.16 TD/INT * Y/A
10. Cam Newton: 590/1002, 7920 Yards, 40 TD, 29 INT = 10.9 TD/INT * Y/A
11. Matt Hasselbeck: 794/1327, 9389 Yards, 50 TD, 34 INT = 10.4 TD/INT * Y/A
12. Drew Brees: 1156/1851, 12667 Yards, 82 TD, 54 INT = 10.39 TD/INT * Y/A
13. Matthew Stafford: 1142/1906, 13187 Yards, 83 TD, 56 INT = 10.25 TD/INT * Y/A
14. Jason Campbell: 1196/1966, 13247 Yards, 68 TD, 46 INT = 9.96 TD/INT * Y/A
15. Peyton Manning: 1799/2922, 21176 Yards, 139 TD, 102 INT = 9.88 TD/INT * Y/A
16. Kyle Orton: 1134/1952, 12774 Yards, 71 TD, 48 INT = 9.68 TD/INT * Y/A
17. Jay Cutler: 1380/2249, 16318 Yards, 106 TD, 80 INT = 9.61 TD/INT * Y/A
18. Andy Dalton: 670/1116, 7451 Yards, 47 TD, 33 INT = 9.51 TD/INT * Y/A
19. Matt Schaub: 527/830, 6317 Yards, 30 TD, 25 INT = 9.13 TD/INT * Y/A
20. Byron Leftwich: 839/1433, 9500 Yards, 52 TD, 39 INT = 8.84 TD/INT * Y/A
21. Matt Moore: 448/758, 5268 Yards, 33 TD, 26 INT = 8.82 TD/INT * Y/A
22. Matt Cassel: 620/1048, 6870 Yards, 39 TD, 29 INT = 8.82 TD/INT * Y/A
23. Michael Vick: 784/1445, 9640 Yards, 54 TD, 42 INT = 8.58 TD/INT * Y/A
24. Josh Freeman: 1101/1873, 12963 Yards, 78 TD, 63 INT = 8.57 TD/INT * Y/A
25. Eli Manning: 1389/2477, 15920 Yards, 106 TD, 81 INT = 8.41 TD/INT * Y/A
26. Charlie Batch: 747/1334, 9063 Yards, 49 TD, 40 INT = 8.32 TD/INT * Y/A
27. Andrew Luck: 367/681, 4662 Yards, 23 TD, 19 INT = 8.29 TD/INT * Y/A
28. Sam Bradford: 873/1498, 9378 Yards, 45 TD, 34 INT = 8.29 TD/INT * Y/A
29. Christian Ponder: 458/774, 4788 Yards, 31 TD, 25 INT = 7.67 TD/INT * Y/A
30. Tyler Thigpen: 272/504, 3192 Yards, 21 TD, 18 INT = 7.39 TD/INT * Y/A
31. Mark Sanchez: 1123/2024, 13247 Yards, 77 TD, 72 INT = 7 TD/INT * Y/A
32. Blaine Gabbert: 372/691, 3876 Yards, 21 TD, 17 INT = 6.93 TD/INT * Y/A
33. Tarvaris Jackson: 369/638, 4148 Yards, 24 TD, 23 INT = 6.78 TD/INT * Y/A
34. Colt McCoy: 409/702, 4388 Yards, 21 TD, 20 INT = 6.56 TD/INT * Y/A
35. Kevin Kolb: 340/572, 4037 Yards, 20 TD, 22 INT = 6.42 TD/INT * Y/A
36. Derek Anderson: 756/1436, 9148 Yards, 53 TD, 55 INT = 6.14 TD/INT * Y/A
37. Alex Smith: 864/1514, 9399 Yards, 51 TD, 53 INT = 5.97 TD/INT * Y/A
38. David Carr: 1243/2070, 13391 Yards, 59 TD, 65 INT = 5.87 TD/INT * Y/A
39. Chad Henne: 812/1373, 9198 Yards, 42 TD, 48 INT = 5.86 TD/INT * Y/A
40. Brandon Weeden: 297/517, 3385 Yards, 14 TD, 17 INT = 5.39 TD/INT * Y/A
41. Bruce Gradkowski: 362/680, 3883 Yards, 20 TD, 23 INT = 4.97 TD/INT * Y/A
42. Matt Leinart: 357/618, 4011 Yards, 15 TD, 20 INT = 4.87 TD/INT * Y/A
43. Ryan Fitzpatrick: 424/734, 4104 Yards, 21 TD, 27 INT = 4.35 TD/INT * Y/A
44. Brady Quinn: 296/550, 3043 Yards, 12 TD, 17 INT = 3.91 TD/INT * Y/A
45. John Skelton: 320/602, 3707 Yards, 15 TD, 25 INT = 3.69 TD/INT * Y/A

So he is almost identical to Brady by that measure.

You put the stat lines together, you put the play-off success together, you come to the conclusion that Joe Flacco has basically been the new Tom Brady, and he deserves to be paid like him.

Now, on Aaron Rodgers... yes, he deserves to be the highest paid QB in the NFL. But he hasn't had his contract day yet. When he does, he will almost certainly be paid better than Flacco, unless he goes all white knight and purposefully signs an under-valued contract to help his team's cap situation.

[callahan09]

Alright, so I was looking at some numbers and came to the realization that when a player has thrown 0 interceptions, both metrics are kind of worthless (they will always yield a score of "Infinity" due to division by 0).

I'm not quite sure how to work around that, but I had some ideas...

For the Yards Per INT stat, I figured I'd just add 1 to interceptions if interceptions is 0, so for someone who has thrown, say, 350 yards with no interceptions, they'll just get 350 Y/INT assigned to them. Seems fair at a glance because that's how many yards they've thrown and haven't thrown an INT, but their very next pass could be an INT. It may not be perfect, but I think it's better than assigning "Infinity Y/INT" to the player. The biggest problem with it as that someone who has also thrown 350 yards, but DID throw 1 INT, will get the exact same value of 350 Y/INT, despite being obviously worse for having thrown an INT. So there must be a better way, I just haven't thought of it yet. Maybe I should just leave the attributed number as "Infinity" but sort them into the ranking list as if the value were 350? Any ideas would be appreciated.

I did a similar thing for the TD/INT * Y/A score, where if INTs was 0 I added 1 to TDs and INTs. So someone who threw 1 TD, 0 INT, gets calculated as if they threw 2 TD, 1 INT. Again, the implication here is that someone who throws 1 TD, 1 INT gets the same score as someone who throws 0 TD, 0 INT. Not sure I can agree with that. Also, someone who throws 2 TD, 1 INT gets the same score as someone who throws 1 TD, 0 INT.

So there is some imperfection here. Players with 0 INTs really throw the system out of whack because they always earn a "perfect" score regardless of the other factors, and I don't really know how to do away with that phenomenon without throwing all the other scores out of whack by making an adjustment.

[callahan09]

OK I came up with a revision. Now we're going to be taking Yards Per Attempt + 100 (this eliminates an issue with negative yards per attempt), and multiplying it by a new formula: TD% + (1 - INT%) + (TD Ratio), where TD Ratio = TD / (TD + INT). This new TD Ratio formula eliminates the problem where 0 interceptions was causing an infinite TD:INT ratio. So now instead of seeing that we have a 9 to 1 TD:INT ratio, we see that we have a 0.9 TD Ratio (as an example), while any amount of TDs to 0 INTs = 1.0 TD Ratio.

The other major factor here is that we are including the rate of touchdowns & interceptions (against attempts) in the equation. This way we can see that 44 TD : 22 INT is not the same thing as 22 TD: 11 INT, despite them having the same TD Ratio. The most important implication of this is that now there is a differentiation between amount of touchdowns when the player has 0 interceptions (without this, 10 TD : 0 INT just counts as the same 1.0 TD Ratio as 1 TD: 0 INT, even though 10 : 0 is obviously better).

Finally we subtract 100 from the final value to give us a baseline of 0 when the player had 0 yards, 0 touchdowns, and 0 attempts. This means that the maximum efficiency value is 497. To get to that lofty figure, one would require 99 yards and a touchdown on *every* attempt. (99 Y/A + 100) * (1 TD% + (1 - 0 INT%) + (1 TD Ratio)) = 199 * 3 = 497.

The minimum value is thus -100, which would require the player to throw an interception on every attempt.

I kind of like this because since traditional QB Rating topped out at 158.3 with 77.5% completion percentage, 12.5 yards per attempt, 11.875% touchdown percentage, and 0% interception percentage, we never could under that model see that someone with, say, an 88% completion percentage and 15% touchdown rate, was actually performing better than someone who had the bottom-level figures for the highest rating. With this, we can never top out the rating unless the player literally tops out the maximum possibility from every attempt (99 yards and a touchdown).

Hopefully that made some sense to you, and this should make things a little clearer: Here are the 2012 Regular Season Qualifying Passers (224+ Attempts):


1. Aaron Rodgers: 371/552, 4295 Yards, 39 TD, 8 INT = 103.3 Efficiency
2. Tom Brady: 401/637, 4827 Yards, 34 TD, 8 INT = 99.1 Efficiency
3. Robert Griffin III: 258/393, 3200 Yards, 20 TD, 5 INT = 98.8 Efficiency
4. Peyton Manning: 400/583, 4659 Yards, 37 TD, 11 INT = 96.1 Efficiency
5. Ben Roethlisberger: 284/449, 3265 Yards, 26 TD, 8 INT = 93.6 Efficiency
6. Russell Wilson: 252/393, 3118 Yards, 26 TD, 10 INT = 90.3 Efficiency
7. Drew Brees: 422/670, 5177 Yards, 43 TD, 19 INT = 86.3 Efficiency
8. Matt Ryan: 422/615, 4719 Yards, 32 TD, 14 INT = 85.7 Efficiency
9. Joe Flacco: 317/531, 3817 Yards, 22 TD, 10 INT = 83.3 Efficiency
10. Matt Schaub: 350/544, 4008 Yards, 22 TD, 12 INT = 78.8 Efficiency
11. Eli Manning: 321/536, 3948 Yards, 26 TD, 15 INT = 77.7 Efficiency
12. Philip Rivers: 338/527, 3606 Yards, 26 TD, 15 INT = 76.8 Efficiency
13. Andy Dalton: 329/528, 3669 Yards, 27 TD, 16 INT = 76.3 Efficiency
14. Cam Newton: 280/485, 3869 Yards, 19 TD, 12 INT = 75.7 Efficiency
15. Josh Freeman: 306/558, 4065 Yards, 27 TD, 17 INT = 75 Efficiency
16. Sam Bradford: 328/551, 3702 Yards, 21 TD, 13 INT = 74.2 Efficiency
17. Carson Palmer: 345/565, 4018 Yards, 22 TD, 14 INT = 74.1 Efficiency
18. Tony Romo: 425/648, 4903 Yards, 28 TD, 19 INT = 73.1 Efficiency
19. Ryan Fitzpatrick: 306/505, 3400 Yards, 24 TD, 16 INT = 72.5 Efficiency
20. Christian Ponder: 300/483, 2935 Yards, 18 TD, 12 INT = 71 Efficiency
21. Blaine Gabbert: 162/278, 1662 Yards, 9 TD, 6 INT = 70.7 Efficiency
22. Jay Cutler: 255/434, 3033 Yards, 19 TD, 14 INT = 69.8 Efficiency
23. Andrew Luck: 339/627, 4374 Yards, 23 TD, 18 INT = 67.8 Efficiency
24. Michael Vick: 204/351, 2362 Yards, 12 TD, 10 INT = 65.6 Efficiency
25. Matthew Stafford: 435/727, 4967 Yards, 20 TD, 17 INT = 65 Efficiency
26. Nick Foles: 161/265, 1699 Yards, 6 TD, 5 INT = 64.9 Efficiency
27. Chad Henne: 166/308, 2084 Yards, 11 TD, 11 INT = 60.1 Efficiency
28. Ryan Tannehill: 282/484, 3294 Yards, 12 TD, 13 INT = 57.9 Efficiency
29. Jake Locker: 177/314, 2176 Yards, 10 TD, 11 INT = 57.5 Efficiency
30. Brandon Weeden: 297/517, 3385 Yards, 14 TD, 17 INT = 54 Efficiency
31. Mark Sanchez: 246/453, 2883 Yards, 13 TD, 18 INT = 49.8 Efficiency
32. Matt Cassel: 161/277, 1796 Yards, 6 TD, 12 INT = 39.7 Efficiency

[callahan09]

I do have an alternate as well, which is that we divide the (Y/A + 100) by 2, and then do not subtract anything from the final score, as opposed to subtracting 100 from the final score. The biggest effects of this are that the baseline (0 Yards, 0 TD, 0 INT) becomes 50, the minimum becomes 0 (instead of a negative number), and the maximum shrinks from 497 to 298.5.

The scores do look a little different, mostly in that it sort of bunches them closer together. Whereas on the previous display method, the range was 39.7 (Cassel) to 103.3 (Rodgers), on this one the range is much closer (69.8 to 101.6), which is actually closer to the range on display from the traditional Passer Rating for these players (66.7 to 108.0). The reason for the bunching is obviously that the possible range has shrunk from -100 to 497, to 0 to 298.5, which is a range of exactly half the size.

Here's the list with this display type:


1. Aaron Rodgers: 371/552, 4295 Yards, 39 TD, 8 INT = 101.6 Efficiency
2. Tom Brady: 401/637, 4827 Yards, 34 TD, 8 INT = 99.5 Efficiency
3. Robert Griffin III: 258/393, 3200 Yards, 20 TD, 5 INT = 99.4 Efficiency
4. Peyton Manning: 400/583, 4659 Yards, 37 TD, 11 INT = 98 Efficiency
5. Ben Roethlisberger: 284/449, 3265 Yards, 26 TD, 8 INT = 96.8 Efficiency
6. Russell Wilson: 252/393, 3118 Yards, 26 TD, 10 INT = 95.1 Efficiency
7. Drew Brees: 422/670, 5177 Yards, 43 TD, 19 INT = 93.1 Efficiency
8. Matt Ryan: 422/615, 4719 Yards, 32 TD, 14 INT = 92.9 Efficiency
9. Joe Flacco: 317/531, 3817 Yards, 22 TD, 10 INT = 91.7 Efficiency
10. Matt Schaub: 350/544, 4008 Yards, 22 TD, 12 INT = 89.4 Efficiency
11. Eli Manning: 321/536, 3948 Yards, 26 TD, 15 INT = 88.8 Efficiency
12. Philip Rivers: 338/527, 3606 Yards, 26 TD, 15 INT = 88.4 Efficiency
13. Andy Dalton: 329/528, 3669 Yards, 27 TD, 16 INT = 88.2 Efficiency
14. Cam Newton: 280/485, 3869 Yards, 19 TD, 12 INT = 87.9 Efficiency
15. Josh Freeman: 306/558, 4065 Yards, 27 TD, 17 INT = 87.5 Efficiency
16. Sam Bradford: 328/551, 3702 Yards, 21 TD, 13 INT = 87.1 Efficiency
17. Carson Palmer: 345/565, 4018 Yards, 22 TD, 14 INT = 87 Efficiency
18. Tony Romo: 425/648, 4903 Yards, 28 TD, 19 INT = 86.6 Efficiency
19. Ryan Fitzpatrick: 306/505, 3400 Yards, 24 TD, 16 INT = 86.2 Efficiency
20. Christian Ponder: 300/483, 2935 Yards, 18 TD, 12 INT = 85.5 Efficiency
21. Blaine Gabbert: 162/278, 1662 Yards, 9 TD, 6 INT = 85.4 Efficiency
22. Jay Cutler: 255/434, 3033 Yards, 19 TD, 14 INT = 84.9 Efficiency
23. Andrew Luck: 339/627, 4374 Yards, 23 TD, 18 INT = 83.9 Efficiency
24. Michael Vick: 204/351, 2362 Yards, 12 TD, 10 INT = 82.8 Efficiency
25. Matthew Stafford: 435/727, 4967 Yards, 20 TD, 17 INT = 82.5 Efficiency
26. Nick Foles: 161/265, 1699 Yards, 6 TD, 5 INT = 82.4 Efficiency
27. Chad Henne: 166/308, 2084 Yards, 11 TD, 11 INT = 80.1 Efficiency
28. Ryan Tannehill: 282/484, 3294 Yards, 12 TD, 13 INT = 78.9 Efficiency
29. Jake Locker: 177/314, 2176 Yards, 10 TD, 11 INT = 78.8 Efficiency
30. Brandon Weeden: 297/517, 3385 Yards, 14 TD, 17 INT = 77 Efficiency
31. Mark Sanchez: 246/453, 2883 Yards, 13 TD, 18 INT = 74.9 Efficiency
32. Matt Cassel: 161/277, 1796 Yards, 6 TD, 12 INT = 69.8 Efficiency

The rankings will be the same. It's really just a display method difference, because you can adapt these scores to the other method by multiplying by 2 and subtracting 100.

[zwerrell]

I do econometrics, so this thread is super interesting to me. I would say you could do something involving TD/Completions and INT/Incompletions and interact the two somehow. Perhaps subtracting INT/Inc from TD/Comp, adding 1-(INT/INC) to TD/COMP.

Another thing you could do is try to create a passing 'INDEX' and use that to better compare quarterbacks across eras. Are you familiar with economic concepts like the GDP deflator?

[callahan09]

Here's some interesting stuff:

Flacco threw 20.1% of his career post-season passes during his rookie season. Since 1960, only 3.0% of all other players' post-season passes came from rookies. Flacco's rookie and second year passes account for 32.2% of his career post-season attempts. All other players, only 10.8% from rookies & sophomores.

That means that among all other passers, 89.2% of post-season pass attempts came from players in their third year or later. For Flacco, he only has 67.8% of his attempts coming from his third season or later.

The total passing stats for players in their third season or later, in the post-season are:

14747/26067, 181851 Yards, 1126 TD, 1048 INT

That's good for a 75.9 QB Rating. Or, using my new model above (for now I'm just going to use the smaller range model, with min of 0 and max of 298.5), it's an 81.4.

Flacco's passer rating as a third year+ QB in the post-season is 105.0, significantly better than average, and ranks 31st of all time (among 238 players). That includes guys with only 1 attempt. The ranking among players with at least 3 attempts is 8th (out of 192). With over 115 attempts, he is 1st (out of 76).

His efficiency (that's my model) is 106.0, significantly better than the average again, and ranks 14th of all time (among 238 players, including those with 1 attempt). Among those with at least 3 attempts, he is 7th. With over 115 attempts, he is 1st again.

And here's another thing using my efficiency score... Since 2010, among the 33 quarterbacks who've been in the league all 3 years (so not counting the second year guys and rookies), who have had at least enough attempts that it would qualify for one season (224+):


1. Aaron Rodgers: 1191/1779, 14749 Yards, 126 TD, 29 INT = 101.1 Efficiency
2. Tom Brady: 1284/1990, 15803 Yards, 123 TD, 31 INT = 99.6 Efficiency
3. Peyton Manning: 896/1331, 9874 Yards, 74 TD, 30 INT = 93.7 Efficiency
[u][i][b]4. Joe Flacco: 1085/1815, 13061 Yards, 85 TD, 34 INT = 93.4 Efficiency[/b][/i][/u]
5. Drew Brees: 1450/2152, 16605 Yards, 131 TD, 57 INT = 93.2 Efficiency
6. Alex Smith: 666/1073, 7746 Yards, 49 TD, 20 INT = 93.1 Efficiency
7. Matt Ryan: 1225/1900, 13642 Yards, 96 TD, 40 INT = 93 Efficiency
8. Ben Roethlisberger: 924/1482, 11453 Yards, 69 TD, 32 INT = 92 Efficiency
9. Tim Tebow: 192/407, 2875 Yards, 19 TD, 9 INT = 91.2 Efficiency
10. Tony Romo: 919/1383, 10692 Yards, 70 TD, 36 INT = 90.8 Efficiency
11. Matt Schaub: 956/1499, 11462 Yards, 63 TD, 32 INT = 90.6 Efficiency
12. Matthew Stafford: 941/1529, 10920 Yards, 70 TD, 36 INT = 90.1 Efficiency
13. Philip Rivers: 1061/1650, 12940 Yards, 83 TD, 48 INT = 89.2 Efficiency
14. Eli Manning: 1125/1827, 14102 Yards, 95 TD, 57 INT = 88.6 Efficiency
15. Michael Vick: 710/1182, 8975 Yards, 52 TD, 31 INT = 88.5 Efficiency
16. Kyle Orton: 452/760, 5500 Yards, 30 TD, 18 INT = 88 Efficiency
17. Jay Cutler: 719/1222, 8980 Yards, 57 TD, 38 INT = 86.7 Efficiency
18. Josh Freeman: 943/1583, 11108 Yards, 68 TD, 45 INT = 86.5 Efficiency
19. Jason Campbell: 326/545, 3822 Yards, 21 TD, 14 INT = 86.3 Efficiency
20. Shaun Hill: 269/432, 2891 Yards, 18 TD, 12 INT = 86.1 Efficiency
21. Matt Cassel: 592/1014, 6695 Yards, 43 TD, 31 INT = 84.9 Efficiency
22. Kevin Kolb: 370/625, 4321 Yards, 24 TD, 18 INT = 84.5 Efficiency
23. Ryan Fitzpatrick: 914/1515, 10232 Yards, 71 TD, 54 INT = 84.3 Efficiency
24. Sam Bradford: 873/1498, 9378 Yards, 45 TD, 34 INT = 83.8 Efficiency
25. Carson Palmer: 906/1479, 10741 Yards, 61 TD, 50 INT = 83.5 Efficiency
26. Mark Sanchez: 886/1592, 10264 Yards, 61 TD, 50 INT = 82.8 Efficiency
27. Matt Hasselbeck: 771/1264, 8469 Yards, 44 TD, 37 INT = 82.6 Efficiency
28. Matt Moore: 300/509, 3485 Yards, 22 TD, 19 INT = 82.4 Efficiency
29. Colt McCoy: 409/702, 4388 Yards, 21 TD, 20 INT = 80.4 Efficiency
30. Tarvaris Jackson: 305/508, 3432 Yards, 17 TD, 17 INT = 80.1 Efficiency
31. Chad Henne: 531/910, 6253 Yards, 30 TD, 34 INT = 78.2 Efficiency
32. Derek Anderson: 173/331, 2123 Yards, 7 TD, 10 INT = 74.6 Efficiency
33. John Skelton: 320/602, 3707 Yards, 15 TD, 25 INT = 72.1 Efficiency

[callahan09]

Something I find very interesting is that my rating gives Tim Tebow a 91.2 and puts him in the top 10. Why might that be, when his passer rating is 77.2 and ranks 28th out of 33?

Personally, this situation speaks to exactly what my problem was with passer rating and why I set out to come up with something better.

Tebow ranked 18th in Y/A, 11th in TD%, and 8th in INT%. How many players ranked higher than him in all 3 categories? Just four: Rodgers, Brady, Flacco, and Ryan. Those 4 players are also the only ones who ranked higher than him in both TD% and INT%. The only place he got hurt was in completion percentage, where he ranked dead last with his 47.2%.

So his raw numbers, what he actually achieved in terms of ball protection, yardage, and scoring, were among the best. It was only his completion percentage that was not, and for that he was dragged down to the bottom few of the list, which seems rather harsh to me.

As an experiment, I was curious just how much his completion percentage was hurting him... If he had thrown the ball the exact same number of times for the exact same yards, touchdowns, and INTs, but completed the average of 61.57% of his passes, it would have propelled his passer rating up to an 89.3, and moved him from 28th on the leader board, to 16th.

Given this same number of completions to get to the average completion percentage, and without adjusting his yards, he could have thrown 7 fewer touchdowns and 6 more interceptions, and still had a better passer rating than he actually had. That means that:

251/407, 2875 Yards, 12 TD, 15 INT

Is, according to QB Rating, better than:

192/407, 2875 Yards, 19 TD, 9 INT

Does that seem ridiculous to anyone else? His touchdown:interception ratio goes from 2.1 down to 0.8 (from a differential of +10 to -3). That's a swing of -13. Same number of attempts, same number of yards, and significantly worse in terms of touchdowns and interceptions, yet a higher rating, because of completion percentage.

I don't like, I never have, and I never will, and so that's why I think my rating does a better job of representing how a quarterback has actually played.

He would have gone down to 31st in TD% and 30th in INT% (now with only John Skelton appearing lower in both categories), in exchange for going up to 12th in Completion% (which isn't even as high as he had been ranked in TD% or INT% before), AND actually be considered an improvement.

Not only would this give him a higher QB Rating, but enough so that he goes up a spot on the rankings, as well, to 27th. Meanwhile, my efficiency ranking would drop him from 9th with a 91.2 to 31st with a 76.9.

When you look at QB Rating, it's tough to see how Tim Tebow has a winning record (9-7) in his career. When you look at my rating, perhaps that makes a little more sense.

[zwerrell]

I do econometrics, so this thread is super interesting to me. I would say you could do something involving TD/Completions and INT/Incompletions and interact the two somehow.

What I've done here is created an index from the formula ((TD/COMP)-(INT/INC))*(COMP%)*(YDS/ATT). This bascially (and using imprecise language) measured how much more likely you are to throw a touchdown on a completion than a interception on an incompletion - meaning a person who has a negative number is hurting the team - with how likely you are to complete a pass, and how far down the field you are getting the ball with each pass. Here is a list of the stats from the quarterbacks listed in the stats on ESPN's page.
PLAYER INDX*COMP% [url="http://espn.go.com/nfl/player/_/id/8439/aaron-rodgers"]Aaron Rodgers, QB[/url] 31.85120484 [url="http://espn.go.com/nfl/player/_/id/2330/tom-brady"]Tom Brady, QB[/url] 24.3018792 [url="http://espn.go.com/nfl/player/_/id/14875/robert-griffin-iii"]Robert Griffin III, QB[/url] 21.61692334 [url="http://espn.go.com/nfl/player/_/id/5536/ben-roethlisberger"]Ben Roethlisberger, QB[/url] 19.81787106 [url="http://espn.go.com/nfl/player/_/id/1428/peyton-manning"]Peyton Manning, QB[/url] 17.75380183 [url="http://espn.go.com/nfl/player/_/id/14881/russell-wilson"]Russell Wilson, QB[/url] 16.39442469 [url="http://espn.go.com/nfl/player/_/id/2580/drew-brees"]Drew Brees, QB[/url] 12.31248605 [url="http://espn.go.com/nfl/player/_/id/11252/joe-flacco"]Joe Flacco, QB[/url] 9.731650904 [url="http://espn.go.com/nfl/player/_/id/12473/josh-freeman"]Josh Freeman, QB[/url] 8.299436975 [url="http://espn.go.com/nfl/player/_/id/5526/eli-manning"]Eli Manning, QB[/url] 4.95738354 [url="http://espn.go.com/nfl/player/_/id/13994/cam-newton"]Cam Newton, QB[/url] 4.291613415 [url="http://espn.go.com/nfl/player/_/id/13197/sam-bradford"]Sam Bradford, QB[/url] 2.290453899 [url="http://espn.go.com/nfl/player/_/id/14874/andrew-luck"]Andrew Luck, QB[/url] 2.018975295 [url="http://espn.go.com/nfl/player/_/id/11237/matt-ryan"]Matt Ryan, QB[/url] 1.731348575 [url="http://espn.go.com/nfl/player/_/id/13987/blaine-gabbert"]Blaine Gabbert, QB[/url] 1.335762452 [url="http://espn.go.com/nfl/player/_/id/14012/andy-dalton"]Andy Dalton, QB[/url] 0.720859083 [url="http://espn.go.com/nfl/player/_/id/5615/matt-schaub"]Matt Schaub, QB[/url] 0.474588925 [url="http://espn.go.com/nfl/player/_/id/4459/carson-palmer"]Carson Palmer, QB[/url] 0.057235968 [url="http://espn.go.com/nfl/player/_/id/8664/ryan-fitzpatrick"]Ryan Fitzpatrick, QB[/url] -0.803700857 [url="http://espn.go.com/nfl/player/_/id/5529/philip-rivers"]Philip Rivers, QB[/url] -1.070681319 [url="http://espn.go.com/nfl/player/_/id/9597/jay-cutler"]Jay Cutler, QB[/url] -1.521766415 [url="http://espn.go.com/nfl/player/_/id/13966/christian-ponder"]Christian Ponder, QB[/url] -2.104477377 [url="http://espn.go.com/nfl/player/_/id/2549/michael-vick"]Michael Vick, QB[/url] -3.598759104 [url="http://espn.go.com/nfl/player/_/id/11291/chad-henne"]Chad Henne, QB[/url] -4.086814526 [url="http://espn.go.com/nfl/player/_/id/14877/nick-foles"]Nick Foles, QB[/url] -4.212898232 [url="http://espn.go.com/nfl/player/_/id/12483/matthew-stafford"]Matthew Stafford, QB[/url] -5.000117068 [url="http://espn.go.com/nfl/player/_/id/14876/ryan-tannehill"]Ryan Tannehill, QB[/url] -8.656389404 [url="http://espn.go.com/nfl/player/_/id/13969/jake-locker"]Jake Locker, QB[/url] -9.300243474 [url="http://espn.go.com/nfl/player/_/id/5209/tony-romo"]Tony Romo, QB[/url] -9.593879736 [url="http://espn.go.com/nfl/player/_/id/14878/brandon-weeden"]Brandon Weeden, QB[/url] -11.32973569 [url="http://espn.go.com/nfl/player/_/id/12482/mark-sanchez"]Mark Sanchez, QB[/url] -11.78016331 [url="http://espn.go.com/nfl/player/_/id/8644/matt-cassel"]Matt Cassel, QB[/url]
-24.91642579

You can basically think of this measure as being how much a quarterback helps or hurts his team on each throw. Bad qb's are punished for throwing the ball further down field, good ones rewarded for the same.

[callahan09]

[quote name='zwerrell' timestamp='1363110665' post='1389064']
I do econometrics, so this thread is super interesting to me. I would say you could do something involving TD/Completions and INT/Incompletions and interact the two somehow.

What I've done here is created an index from the formula ((TD/COMP)-(INT/INC))*(COMP%)*(YDS/ATT). This bascially (and using imprecise language) measured how much more likely you are to throw a touchdown on a completion than a interception on an incompletion - meaning a person who has a negative number is hurting the team - with how likely you are to complete a pass, and how far down the field you are getting the ball with each pass. Here is a list of the stats from the quarterbacks listed in the stats on ESPN's page.
PLAYER INDX*COMP% [url="http://espn.go.com/nfl/player/_/id/8439/aaron-rodgers"]Aaron Rodgers, QB[/url] 31.85120484 [url="http://espn.go.com/nfl/player/_/id/2330/tom-brady"]Tom Brady, QB[/url] 24.3018792 [url="http://espn.go.com/nfl/player/_/id/14875/robert-griffin-iii"]Robert Griffin III, QB[/url] 21.61692334 [url="http://espn.go.com/nfl/player/_/id/5536/ben-roethlisberger"]Ben Roethlisberger, QB[/url] 19.81787106 [url="http://espn.go.com/nfl/player/_/id/1428/peyton-manning"]Peyton Manning, QB[/url] 17.75380183 [url="http://espn.go.com/nfl/player/_/id/14881/russell-wilson"]Russell Wilson, QB[/url] 16.39442469 [url="http://espn.go.com/nfl/player/_/id/2580/drew-brees"]Drew Brees, QB[/url] 12.31248605 [url="http://espn.go.com/nfl/player/_/id/11252/joe-flacco"]Joe Flacco, QB[/url] 9.731650904 [url="http://espn.go.com/nfl/player/_/id/12473/josh-freeman"]Josh Freeman, QB[/url] 8.299436975 [url="http://espn.go.com/nfl/player/_/id/5526/eli-manning"]Eli Manning, QB[/url] 4.95738354 [url="http://espn.go.com/nfl/player/_/id/13994/cam-newton"]Cam Newton, QB[/url] 4.291613415 [url="http://espn.go.com/nfl/player/_/id/13197/sam-bradford"]Sam Bradford, QB[/url] 2.290453899 [url="http://espn.go.com/nfl/player/_/id/14874/andrew-luck"]Andrew Luck, QB[/url] 2.018975295 [url="http://espn.go.com/nfl/player/_/id/11237/matt-ryan"]Matt Ryan, QB[/url] 1.731348575 [url="http://espn.go.com/nfl/player/_/id/13987/blaine-gabbert"]Blaine Gabbert, QB[/url] 1.335762452 [url="http://espn.go.com/nfl/player/_/id/14012/andy-dalton"]Andy Dalton, QB[/url] 0.720859083 [url="http://espn.go.com/nfl/player/_/id/5615/matt-schaub"]Matt Schaub, QB[/url] 0.474588925 [url="http://espn.go.com/nfl/player/_/id/4459/carson-palmer"]Carson Palmer, QB[/url] 0.057235968 [url="http://espn.go.com/nfl/player/_/id/8664/ryan-fitzpatrick"]Ryan Fitzpatrick, QB[/url] -0.803700857 [url="http://espn.go.com/nfl/player/_/id/5529/philip-rivers"]Philip Rivers, QB[/url] -1.070681319 [url="http://espn.go.com/nfl/player/_/id/9597/jay-cutler"]Jay Cutler, QB[/url] -1.521766415 [url="http://espn.go.com/nfl/player/_/id/13966/christian-ponder"]Christian Ponder, QB[/url] -2.104477377 [url="http://espn.go.com/nfl/player/_/id/2549/michael-vick"]Michael Vick, QB[/url] -3.598759104 [url="http://espn.go.com/nfl/player/_/id/11291/chad-henne"]Chad Henne, QB[/url] -4.086814526 [url="http://espn.go.com/nfl/player/_/id/14877/nick-foles"]Nick Foles, QB[/url] -4.212898232 [url="http://espn.go.com/nfl/player/_/id/12483/matthew-stafford"]Matthew Stafford, QB[/url] -5.000117068 [url="http://espn.go.com/nfl/player/_/id/14876/ryan-tannehill"]Ryan Tannehill, QB[/url] -8.656389404 [url="http://espn.go.com/nfl/player/_/id/13969/jake-locker"]Jake Locker, QB[/url] -9.300243474 [url="http://espn.go.com/nfl/player/_/id/5209/tony-romo"]Tony Romo, QB[/url] -9.593879736 [url="http://espn.go.com/nfl/player/_/id/14878/brandon-weeden"]Brandon Weeden, QB[/url] -11.32973569 [url="http://espn.go.com/nfl/player/_/id/12482/mark-sanchez"]Mark Sanchez, QB[/url] -11.78016331 [url="http://espn.go.com/nfl/player/_/id/8644/matt-cassel"]Matt Cassel, QB[/url]
-24.91642579

You can basically think of this measure as being how much a quarterback helps or hurts his team on each throw. Bad qb's are punished for throwing the ball further down field, good ones rewarded for the same.
[/quote]

Thanks for you contribution to the discussion.

I have looked into doing something just like that before and after a while I came up with a lot of problems related to it. I didn't look too deeply into your example, but my problem I discovered with TD/Completion and INT/Incompletion is that if we're going to consider interceptions out of incompletions, then somebody who throws the ball 100 times and completes 99 passes, then has 1 INT, will have an AWFUL INT/Incompletions number of 100%. This despite the fact that he only has a 1% interception rate per attempt, which is extremely good.

An example of how this skews things:

9/10, 100 Yards, 1 TD, 1 INT

This is a pretty good stat line at 90% completion percentage, 10 yards per attempt, and 10% touchdown rate, all very, very high numbers. Only thing not great is the 10% interception rate. My efficiency formula gives me an 82.5 for this stat line.

But with your formula, this comes out to: ((1/9) - (1/1)) * (9/10 * 100) * (100/10) = (0.111 - 1) * 90 * 10 = -0.888 * 900 = -800

So that's certainly not going to work, yielding a negative score because our lone incompletion was an interception. What if I throw a bunch of incomplete passes, now? Say, 10 incomplete passes. This is going to wreck my completion percentage (45%), yards per attempt (5.00), and touchdown percentage (5%), but give me a little benefit to interception rate (5%).

The result is that my efficiency rating will lower this score to 78.8, while yours will boost it to a 4.55. From negative 800 to positive 4.55, because of throwing 10 incompletions.

In fact, the way your formula is worked out, any time there are 0 touchdowns, the score has to be either 0 or negative. Any time completion percentage is exactly 50% and touchdowns and interceptions are equal to each other, the score will be 0. And in your method, the higher the completion percentage, the more damaging interceptions are and the less helpful touchdowns are, all other factors being equal. That's because if my completion percentage is higher, it means I've thrown fewer incompletions, and therefore my interception percentage out of those incompletions will be higher; and on the same token my touchdowns out of completions percentage will be lower if I have completed more passes, out of the same number of attempts.

In fact, you've attributed an inverse value to completion percentage.

If I throw the ball 10 times for 100 yards, 1 touchdown, and 1 interception, then my rating will be the following at the various numbers of completions:

1 Completion = (1/1) - (1/9) * 10% * 10 Y/A = 88.9 Rating
5 Completions = (1/5) - (1/5) * 50% * 10 Y/A = 0 Rating
9 Completions = (1/9) - (1/1) * 90% * 10 Y/A = -800 Rating

I do appreciate your contribution and I only went into so much detail in explaining what was wrong with it because I had previously gone down the same road that you suggested and came away after much wasted time realizing that it just wasn't the right approach so I wanted to share with you what I'd discovered.

[callahan09]

In addition, I'd like to again thank you for your contribution, because it made me think a little about various stat line scenarios. I discovered something when I ran the numbers with my efficiency formula on the same 3 stat lines that I used at the end of my previous post.


1. test: 1/10, 100 Yards, 1 TD, 1 INT = 82.5 Efficiency
2. test2: 5/10, 100 Yards, 1 TD, 1 INT = 82.5 Efficiency
3. test3: 9/10, 100 Yards, 1 TD, 1 INT = 82.5 Efficiency

Simply put, completion percentage has ZERO effect on the rating. Perhaps this is a problem... The issue I have with completion percentage being a part of the formula is that it makes it nearly impossible to always give precedence to the better raw stats (yards, touchdowns, and interceptions per attempt), because if completion percentage is in the mix, it will always be capable of giving a stat line with all 3 production stats better, a lower score due to the worse completion percentage. I don't want to see that, ever, so I'm in a quandary. On the one hand, when it comes to playing the game, 100 yards a touchdown and INT are a 100 yards, a touchdown, and an INT, whether you dinked and dunked your way down the field, or whether you threw the ball once and went the distance.

But I'm sure most people don't think it looks *right* to see the same score for someone with a 10% completion percentage as someone with a 90%.

In the end, though, is it really that wrong? I mean, what was different about the 9 incompletions versus the 1? They both at one point found the end zone, they both at one point turned the ball over... The guy who threw a lot of incompletions probably stopped the clock a lot more, but not necessarily given that the other guy could have thrown the ball to the sidelines with the receiver going out of bounds on every toss.

I dunno... I just thought I'd share that little idiosyncrasy of the formula with you all. I don't think there's anything I can do about it that won't ruin the formula in worse ways, so I think I'll leave it as is.

Anybody familiar with Adjusted Yards Per Attempt (which is Yards + 20*TD - 45*INT divided by Attempts) knows that there are some useful metrics on performance and production that don't allow completion percentage to be a part of the equation, and I'd consider this one of them.

[callahan09]

This one isn't directly related to Flacco's numbers, but it does speak volumes about what our quarterback accomplished, that none of our previous quarterbacks could: In 2012, the Ravens defense ranked 12th in points allowed. There have been 6 other seasons in Ravens history where the team did not rank in the top 10 in points allowed, and in those seasons they had a combined 28-51-1 record (0.356 winning percentage), 0 playoff appearances, 0 playoff wins, best season 7-9, and 0 seasons at 8-8 or better.

So to bring this team to where we got in 2012 without a top 10 defense is unprecedented for the Baltimore Ravens, and we couldn't have done it without a QB like Flacco to lead the offense.

Also, in 9 seasons where the defense did not rank in the top FIVE, we are 61-82-1 (0.427 winning percentage), 1 playoff appearance, 0 playoff wins, best season 10-6, only 3 seasons at 8-8 or better