[Clark Heins]

One of the most fascinating articles I have ever come across I found on the FootballOutsiders.com web site written by Jason McKinley in 2006. He presents an analysis of comeback win opportunities by the various QBs who played between 1996-2005. According to his research, there were 2598 regular season and playoff games played during those 10 years. Of those, 1474 of those games had one team trailing another by 8 or fewer points in the 4th quarter---thus presenting an opportunity for a comeback win---56.7%. Of those 1474 games, 603 comeback wins did occur---40.9%. Overall, 23.2% of all games played ended in comebacks.
McKinley ranks the QBs by the number of comeback wins they had compared to the number of comeback win opportunities they had. His top ten (in no particular order), from 1996-2005 were: Jake Plummer 19/47; Peyton Manning 19/48; Vinny Testaverde 19/48; Tom Brady 13/21; John Kitma 15/38; Kerry Collins 17/47; Donovan McNabb 12/27; Marc Bulger 10/15; Jake Delhomme 10/22; Jay Fiedler 10/22. Other notables: Kurt Warner 5/24; Mark Brunnell 14/53; Brian Griese 4/24; Doug Flutie 6/23; Troy Aikman 6/22; Warren Moon 4/16; Brett Favre 16/50; Ben Roethisberger 7/9; Drew Bledsoe 19/61; Steve Young 7/11; John Elway 7/15.

McKinley uses some sort of mathematical formula to rate 162 QBs who had one or more comeback opportunities over those 10 years. According to his rankings, Jake Plummer was No. 1 and Mark Brunnell was 162nd. He points out that the two biggest surprises were Vinny Testaverde being near the top and Kurt Warner being near the bottom.

[Bryan Lutes]

Well, if Mr. Heins describes something produced by the hacks at FootballOutsiders as “fascinating”, that would explain a lot about his unique perspective on football. Again, its odd that someone would disregard the opinions of the beat writers (i.e. the people who get paid to cover football) but at the same time put merit on the thoughts of some pompous Bill James-wannabes who have yet to figure out that football statistics cannot be analyzed in isolation. I just read a FootballOutsiders article that concluded the 1976 Raiders had the toughest ‘postseason road to the Super Bowl title’ in NFL history. They kept quoting this guy named Pythagoras. I guess Pythagoras was a big Raider fan, because he ignored the fact that the Patriots had made the playoffs for the first time in about 15 years, the Steelers had so many guys on injured reserve that Ray Mansfield handled the kicking duties, and the Vikings were 0-3 in Super Bowls up to that point.


As for the article at hand…
http://www.footballo...-analysis/3978/

I kept waiting for the punch-line, as if the article was a satire of a statistical analysis. Lots of talk about t-tests, p-values, Shapiro-Wilk tests…yet the faultiness of the analysis can be found in the opening paragraph:

“Since 1996, 1,474 out of 2,598 regular season and postseason games have featured a team trailing by eight points or less in possession of the ball in the fourth quarter. In 603 of those 1,474 games, the trailing team won.Therefore, nearly a quarter of all victories in the last decade have been the result of late and dramatic rallies.”

I understand that the 4th quarter is the last quarter of a game, but its still 15 minutes long. In the extreme, the Redskins could be trailing by 6 points going into the 4th quarter, the Redskins offense does nothing and has to punt, the Redskins’ defense returns a fumble for a TD to go up by 1 point, and then the Redskins’ defense holds the opposing offense scoreless the rest of the game. Viola! A late and dramatic comeback victory can be chalked up for Jason Campbell.

I don’t get how the writer can make such an erroneous, sweeping conclusion that any victory by a team that trailed by 8 or less points in the 4th quarter is the product of a “late and dramatic rally”…AND that the “late and dramatic rally” is the direct result of the QB’s performance. This basic gap in logic coupled with the wacky results in which 5 of the top 10 QBs are Fiedler, Kitna, Collins, Plummer, and Testaverde, should have been more than enough to prove the worthlessness of such an analysis.

In this past Super Bowl, Tom Brady’s Pats trailed the Giants 10-7 in the 4th quarter. Brady actually led his team to the go-ahead score (14-10), but then the Pats defense gave up a late Giant TD. Brady got the ball back with about 30 seconds left and had to go virtually the length of the field to tie the game. The game ends, the Giants win, yet Brady is given a “failed comeback opportunity”. Huh?

[JohnH19]

One quick look at the list shows how irrelevant it is. Great, good and mediocre QB's are interspersed throughout.

It's similar to those goofy protecting a lead in the late innings or third period stats. Even bad teams have pretty good records in those situations. They just have less opportunity to protect leads because they're a bad team and they're usually behind in games.

John

[Clark Heins]

Bryan Lutes, on Mar 14 2008, 08:08 AM, said:

Well, if Mr. Heins describes something produced by the hacks at FootballOutsiders as “fascinating”, that would explain a lot about his unique perspective on football. Again, its odd that someone would disregard the opinions of the beat writers (i.e. the people who get paid to cover football) but at the same time put merit on the thoughts of some pompous Bill James-wannabes who have yet to figure out that football statistics cannot be analyzed in isolation. I just read a FootballOutsiders article that concluded the 1976 Raiders had the toughest ‘postseason road to the Super Bowl title’ in NFL history. They kept quoting this guy named Pythagoras. I guess Pythagoras was a big Raider fan, because he ignored the fact that the Patriots had made the playoffs for the first time in about 15 years, the Steelers had so many guys on injured reserve that Ray Mansfield handled the kicking duties, and the Vikings were 0-3 in Super Bowls up to that point.


As for the article at hand…
http://www.footballo...-analysis/3978/

I kept waiting for the punch-line, as if the article was a satire of a statistical analysis. Lots of talk about t-tests, p-values, Shapiro-Wilk tests…yet the faultiness of the analysis can be found in the opening paragraph:

“Since 1996, 1,474 out of 2,598 regular season and postseason games have featured a team trailing by eight points or less in possession of the ball in the fourth quarter. In 603 of those 1,474 games, the trailing team won.Therefore, nearly a quarter of all victories in the last decade have been the result of late and dramatic rallies.”

I understand that the 4th quarter is the last quarter of a game, but its still 15 minutes long. In the extreme, the Redskins could be trailing by 6 points going into the 4th quarter, the Redskins offense does nothing and has to punt, the Redskins’ defense returns a fumble for a TD to go up by 1 point, and then the Redskins’ defense holds the opposing offense scoreless the rest of the game. Viola! A late and dramatic comeback victory can be chalked up for Jason Campbell.

I don’t get how the writer can make such an erroneous, sweeping conclusion that any victory by a team that trailed by 8 or less points in the 4th quarter is the product of a “late and dramatic rally”…AND that the “late and dramatic rally” is the direct result of the QB’s performance. This basic gap in logic coupled with the wacky results in which 5 of the top 10 QBs are Fiedler, Kitna, Collins, Plummer, and Testaverde, should have been more than enough to prove the worthlessness of such an analysis.

In this past Super Bowl, Tom Brady’s Pats trailed the Giants 10-7 in the 4th quarter. Brady actually led his team to the go-ahead score (14-10), but then the Pats defense gave up a late Giant TD. Brady got the ball back with about 30 seconds left and had to go virtually the length of the field to tie the game. The game ends, the Giants win, yet Brady is given a “failed comeback opportunity”. Huh?

It seems strange to me that, in a Forum supposedly dedicated to the pusuit of research, that so much time and effort goes into disparaging the research efforts of others. I am not above this character flaw myself. Several months ago I questioned the usefulness of documenting QB sacks by Mr. Turney and his friend. Then, it suddenly dawned on me that I was doing the same thing to John as had been done to me by others about my interest in QB w-l records. More the worst--- as John had encouraged me in my own research. As soon as I could, I apologized. for my utter stupidity.

I found Mr. McKinley's article fascinating for two reasons. First, this type of information on comebacks is very difficult, if not impossible, to find. Oh sure, you can find it on some of the better known QBs like Brett Favre and Peyton Manning, but try to find it on a Doug Flutie or a Kerry Collins.
Secondly, I marvel at "the achieve of, the mastery of the thing!". Haven't you ever asked yourself how many games result in comebacks? Well, thanks to Mr. McKinley's research, we now have a pretty good estimate. Knowing the approximate number of games that result in comebacks, we can go forward and formulate an additional hypothesis about the value of the number of comebacks that each QB attains and that, in turn, leads to a better understanding of each QBs worth. And that, in turn, is what interests me.
It does not matter whether you or I think that Mr. Mckinley's research has any value. It obviously had value to him. He must have spent thousands of hours reviewing thousands of games to obtain this information. I know how hard it is to obtain this type of data on just one QB, let alone to obtain it on 162 QBs.

Research is often plodding and boring and seems to serve little purpose in the grand scheme of things, but, like a farmer plowing his field, his effort often results in untold dividends:

No wonder of it: sheer plod makes plough down
sillion
Shine, and blue-bleak embers, ah my dear,
Fall, gall themselves, and gash gold-vermillion.

Gerard Manley Hopkins, "The Windhover"

[Bob Carroll]

That this organization is dedicated to the pursuit of research is not in question. The Forum is in pursuit of discussion. One thing discussable is whether a bit of research is valuable to others.

If someone enjoys researching quarterbacks’ eye color, that’s obviously of interest to him and may even make a couple of good trivia questions. But when the researcher begins ranking QBs by the blueness of their eyes, we draw the line. He’s welcome to believe anything he likes, but we should recognize his ideas as unlikely and say so before he influences others. That’s our duty.

When I first glanced at this piece of research, I thought it dealt with comebacks when one team led by MORE than eight points going into the fourth quarter. That made sense. Overcoming an opponent’s eight (or more) point lead was a definite accomplishment. Then I realized I’d misread the item.

Trailing by a couple of points in a game, especially in a game where the scoring lead may have gone through many changes, doesn’t strike me as much to come back from. Technically, scoring a safety to win a game by a point may indeed constitute a comeback. It just comes off as so much “So What!” to me. And, because of that, I can’t get excited or even recommend this research.

I wonder how it would play out with the breaking point being over eight instead of under. .

[Clark Heins]

Although I've downplayed the importance of comebacks (it was Joe Montana who pointed out, when he was a color commentator in 1995, that, while it is important to be able to make comebacks, it is more important and much more difficult to maintain 4th quarter leads than overcome them), I've always been fascinated in trying to document them as no one prior to Mr. McKinley has ever taken the time and made the effort to do so. I said in a previous post under this topic that well-known QBs like Brett Favre should be easy to document. Boy, was I mistaken! I've spent the last 3 hours trying to document his comebacks. The only thing I've been able to confirm are the results of Mr. McKinley's research---Favre had 16 comebacks between 1996-2005 (H=home; A=away).: 1996---1H; 1997---zero; 1998 ---1H; 1999---3H; 2000---zero; 2001---2H; 2002---3H; 2003---1H; 1A; 2004---1H; 2A; 2005---1H. In 2006, Favre had 3 comebacks at home. In 2007, Favre had two comebacks, one at home and the other on the road. So, from 1996-2007, Favre had 21 comebacks, 17 of which came at home.

The problem is that various web sites, including the Packers web site, have all sorts of numbers---36, 38, 40--- concerning Favre's comeback totals. Part of the problem is obvious---whoever wrote up Favre's year-by-year biographies for tthe Packers counted overtime games which were either tied at the end of 3 quarters or were tied by the opposing team in the 4th quarter---in other words, games that the Packers were never really behind in the 4th quarter. The Cowboys web site did the exact same thing for Roger Staubach giving him credit for 7 comebacks in games which the Cowboys were never behind. I count at least 3 games, one from 2004 and 2 from 2003 which Favre was given credit for comebacks in games the Packers were never behind in (or at the beginning of) the 4th quarter.

The best article I've been able to find on this subject mentioned that Favre had 27 comebacks entering the Nov. 23, 2003 game against SF. Favre added 2 more comebacks that season for a seeming total of 29. If everything else is correct, he would have ended up with 38 comebacks---which is one of the figures that the Packers web site comes up with. What that means is that Favre would have to had 17 comebacks between 1992-95---an aveage of better than 4 per season.

Several questions remain. How accurate is that figure of 17 comebacks between 1992-95? Also, how many of those comebacks occurred at home? The reason why this question is fascinating is that Favre was a great "homer" and he scored an inordinate number of his comebacks at home. My guess would be that Favre had only 8 road comebacks out of his 38 total. In comparison, Unitas had 15 of his 31 comebacks on the road; Marino had 16 of 37; Elway had 18 of 46; Staubach 10 of 16; Kelly 13 of 23; Montana 23 of 30.

So, what on the surface seems a sure bet, isn't a sure bet at all and, in fact, is difficult to ascertain. That is why Mr. McKinley's work is so fascinating. There is a wonderful research project made possible by McKinley staring us in the face---if some youngster is willing to tackle it. It would be a comparison of how many comebacks a QB attains during his career vs. the number of comebacks he would be expected to attain on aveage. For example, with Favre, take the total number of games that he started in his career (regular season and post-season). Multiply that figure by 23.2% (the percentage of games decided by comebacks). Divide that figure by 2 to give us an estimate of how many comebacks Favre should have on aveage. If my figuring is correct, it should come out to a total of 32 comebacks. But Favre had what appears to be 38 comebacks. Divide 38 by 32 and you have a figure to work with. Favre's figure would be 1.2 which is positive. In other words, Favre had more comebacks than the aveage QB would be expected to have given the same number of starts. Do the same thing with other QBs and see if you have any trends.

[Clark Heins]

One of the biggest pitfalls of trying to work with comeback figures is that the researcher has to be very careful about some of the publicity that various football teams put out about their QBs. I've mentioned the Dallas Cowboys padding Roger Staubach's figures, but the New England Patriots take the cake. In their eagerness to promote Tom Brady, they've given him credit for upwards of 28 comebacks. They add a notation that this figure includes games that were tied in the 4th quarter, but the problem is that few people, including sports writers, make note of the notation. They just give Brady credit for 28 comebacks. The actual total for Brady is 19 comebacks---7 at home, 11 on the road and one at a neutral site. That is still an impressive total If we compare Brady against Brett Favre---remember that Favre had a comeback/expected comeback ratio of just under 1.2. Brady's ratio is just under 1.3. On the surface, Brady would seem to surpass Favre in his ability to lead comebacks. But is that really a good thing? Well, yes and no.

It is important to have a QB who can lead your team from behind in the 4th quarter. In 23.2% of all games played, one QB or the other will lead his team to a comeback win. All other things being equal, an aveage NFL QB should rack up 11.6 comebacks per 100 starts. An aveage over that figure presumedly is good.

But, as always, there is another way of looking at things. A large number of comebacks by a QB are a two-edged sword. They cut both ways. Yes, they tell us that the QB achieving such high figures is good in the clutch. But they also tell us that that QB had his team trailing for three quarters of the game! The more dominant a QB is, the fewer comeback opportunities should exist.

Previously, I had mentioned that three of the most dominant QBs in NFL history were Otto Graham, Roger Staubach and Steve Young. According to my expressed theory, these three QBs should have much lower comeback/expected comeback ratios than eithe Brady or Favre. Let's see if I'm right.


Otto Graham---77 starts (ties not counted); approximately 9 comebacks; 8,9 expected comebacks; comeback/expected comeback ratio 1.01.

Roger Staubach---130 starts; 16 comebacks; 15.1 expected comebacks; comebacks/expected comebacks ratio 1.06.

Steve Young ---157 starts; 9 comebacks; 18.2 expected comebacks; comebacks/expected comeback ratio .495.


Well, according to our figures, both Graham and Staubach had just about what you would expect an aveage NFL QB to have in number of comebacks. In other words, nothing special---especially for "Captain Comeback".
But look at Young. He has much fewer the number of comebacks than would be expected by an aveage NFL QB. In other words, Young has to be among the most dominant, if not the most dominant QB of all-time, because he rarely left his team in a position where they had to make a comeback.

What does all of this prove. Well, essentially you are left with a Bell-shaped curve---both ends of which are exceptional while the rest (the majority) is just aveage. The two ends are exceptional for different reasons however: On one end is Young because of his dominance. On the other end, we might find a QB like Elway or Brady who are also exceptional in that they have many more comebacks than an aveage NFL QB is expected to have. However, which degree of exceptionality is most important. To my way of thinking, just as Joe Montana pointed out, it is more important to be dominant than to be lucky. It is better to be so far ahead in a game so that comebacks are unnecessary than to find yourself continually fighting from behind.

[Bob Carroll]

Alright, let's get a few things straight. We have nothing against anyone researching anything about pro football and then publishing his findings. The findings may be worthless, but that's not bothersome. If someone wants to carefully total all the warts found on the butts of all the NFL quarterbacks, we say that's his right. The quarterbacks might object, but we won't.

However, when someone begins ranking quarterback efficiency by the number of warts, we draw the line. Such a pseudo-stat system is the use of so-called "comeback wins" to show the superiority of one quarterback or another.

I must in honesty admit that I'm not quite sure how the purveyor of this nonsense defines "comeback win." He writes such long, rambling posts that I lose track. I think he means when a team is behind by any number of points at the end of the third quarter and then goes on to win the game. If a team is leading going into the final quarter, then falls behind until finally ending on top at the end, I don't know if that counts or not. Since the lead often goes back and forth in a typical game, it would seem that any time spent behind would qualify a team for making a comeback. If the Packers trail the Bears by three with ten seconds left in the third quarter and then the Pack scores a TD with fifteen seconds gone in the fourth quarter, which team must win to get credit for a comeback?

The way the lead can change hands, it seems that trailing by a point or two into the final quarter doesn't mean a thing. Personally, I would define a comeback as winning a game in which a team trailed by eight or more points at any time in the final quarter. I still don't think you can use that stat to rank quarterbacks; there are far too many factors at play here to give full credit for the win exclusively to the QB. But at least overcoming an 8-point lead in the final quarter would seem to be some sort of achievement.

[Bryan Lutes]

Clark Heins, on Apr 12 2008, 10:06 PM, said:

But look at Young. He has much fewer the number of comebacks than would be expected by an aveage NFL QB. In other words, Young has to be among the most dominant, if not the most dominant QB of all-time, because he rarely left his team in a position where they had to make a comeback.

This would be like rating a relief pitcher's effectiveness solely on their number of blown saves. You look at the data, and you conclude "Frank DiPino was superior to Lee Smith, because Smith has nearly ten times as many blown saves as DiPino".

I wonder how many comeback wins Ed Rutkowski has? I would guess the total is even smaller than Young's. Probably because Rutkowski rarely left his team in a position where they could make a comeback. A sure sign of dominance.

I don't quite understand how Mr. McKinley made a wonderful research project possible. His own research project was flawed from the start due to his highly questionable definition of a "comeback" (although I do give McKinley credit for, at the very least, clearly stating what his definition of a comeback is. Other people are much more ambiguous on this matter). After reading McKinley's comeback criteria, I guessed that the results would show guys like Jay Fiedler and Kerry Collins to be among the best QBs in the NFL because coming back from only one score down with a timeframe of 15 minutes isn't that big of a deal. Fiedler and Collins played for some decent teams in their careers, so one would think that if they won a good amount of games in general, then a few of those wins would qualify as "comebacks" even if the win could directly be attributed to Jason Taylor returning a fumble for a TD.

So, in my opinion, the only thing fascinating about McKinley's research is that I was able to accurately predict the meaninglessness of the results.

[Clark Heins]

Bryan Lutes, on Apr 14 2008, 08:10 AM, said:

This would be like rating a relief pitcher's effectiveness solely on their number of blown saves. You look at the data, and you conclude "Frank DiPino was superior to Lee Smith, because Smith has nearly ten times as many blown saves as DiPino".

I wonder how many comeback wins Ed Rutkowski has? I would guess the total is even smaller than Young's. Probably because Rutkowski rarely left his team in a position where they could make a comeback. A sure sign of dominance.

I don't quite understand how Mr. McKinley made a wonderful research project possible. His own research project was flawed from the start due to his highly questionable definition of a "comeback" (although I do give McKinley credit for, at the very least, clearly stating what his definition of a comeback is. Other people are much more ambiguous on this matter). After reading McKinley's comeback criteria, I guessed that the results would show guys like Jay Fiedler and Kerry Collins to be among the best QBs in the NFL because coming back from only one score down with a timeframe of 15 minutes isn't that big of a deal. Fiedler and Collins played for some decent teams in their careers, so one would think that if they won a good amount of games in general, then a few of those wins would qualify as "comebacks" even if the win could directly be attributed to Jason Taylor returning a fumble for a TD.

So, in my opinion, the only thing fascinating about McKinley's research is that I was able to accurately predict the meaninglessness of the results.

You do make a good point about how figures can be misleading on Young's end of the Bell-curve. I should have made something akin to the following notation: Caveat. Some QBs were simply not good in the clutch like Jeff George who, like Young, would have a very low comeback/expected comeback ratio. In George's case, such a figure would not reflect dominance, but just the opposite. Therefore, on Young's end of the Bell-curve, it would be necessary to separate the wheat from the chaff. (Note: I miscounted Young's comebacks---he had ten comebacks.)

Obviously, there are many problems with this type of research. Just obtaining accurate comeback figures for the various QBs is often a nightmare. At least, Mr. McKinley's research is very helpful in this aspect. What I don't like about his research is when he starts making up complicated mathematical equations which few of us are capable of understanding. I prefer simple straightforward math which is easy to understand. Once you start making up complicated formulas, personal bias enters into the equation.
Another problem area is the largeness or smallness of the sample data. For example, Otto Graham had only 77 starts in the NFL. If my count is off by one or two comebacks, his comeback/expected comeback ratio could be way off. So the formula that I've come up with works much better with QBs who have started many games. It simply wouldn't work with an Ed Rutkowski who had only 3 starts.
Another problem area that you run into are those QBs who may have recorded comebacks in relief of another QB. How do you record comebacks in these cases if you're unsure about which QB was in the game when the comeback occurred? Earl Morrell, who often pulled out games for Unitas and Griese, specifically comes to mind.
So, there are numerous problems with this type of research and who can say whether it is worthwhile or not. At best, it gives us approximations of how ell, or not well, a QB performs in the clutch. It is one tool among many, but I believe it is a worthy tool. Right now we have a very small data base, consisting of comeback figures on a few major QBs, to explore. That is why I'm hoping that some youngster will take up the chase. It is not an easy task, but thanks to the research results of Ken Pullis, it can be done. Pullis' research now gives us the actual games that each QB started. Knowing the actual games, we can then go back and review line and box scores from these games and approximate the number of comebacks for each QB. It just takes time and time is the one commodity that I'm running out of. The research opportunity is sitting there---ripe for the picking. Somebody take up the quest.

[Bob Carroll]

[quote name='Clark Heins' date='Apr 14 2008, 11:06 AM' post='2329']

Obviously, there are many problems with this type of research. Just obtaining accurate comeback figures for the various QBs is often a nightmare. At least, Mr. McKinley's research is very helpful in this aspect.



Actually, since PFRA has made virtually all the linescores gouig back to 1920, it should be easy to find the "comebacks" if someone could settle on a definition.

I believe it was Einstein who raged against complicated math formulas. He preferred the simplicity of add and subtract. However, when he entered eleventh grade, Lester Einsten actually enjoyed algebra.

I have to agree that getting kids hooked on researching comebaxck wins. It'll keep them off the streets, and gives us time to think up other busywork for them.

[Clark Heins]

Clark Heins, on Apr 12 2008, 01:06 AM, said:

Although I've downplayed the importance of comebacks (it was Joe Montana who pointed out, when he was a color commentator in 1995, that, while it is important to be able to make comebacks, it is more important and much more difficult to maintain 4th quarter leads than overcome them), I've always been fascinated in trying to document them as no one prior to Mr. McKinley has ever taken the time and made the effort to do so. I said in a previous post under this topic that well-known QBs like Brett Favre should be easy to document. Boy, was I mistaken! I've spent the last 3 hours trying to document his comebacks. The only thing I've been able to confirm are the results of Mr. McKinley's research---Favre had 16 comebacks between 1996-2005 (H=home; A=away).: 1996---1H; 1997---zero; 1998 ---1H; 1999---3H; 2000---zero; 2001---2H; 2002---3H; 2003---1H; 1A; 2004---1H; 2A; 2005---1H. In 2006, Favre had 3 comebacks at home. In 2007, Favre had two comebacks, one at home and the other on the road. So, from 1996-2007, Favre had 21 comebacks, 17 of which came at home.

The problem is that various web sites, including the Packers web site, have all sorts of numbers---36, 38, 40--- concerning Favre's comeback totals. Part of the problem is obvious---whoever wrote up Favre's year-by-year biographies for tthe Packers counted overtime games which were either tied at the end of 3 quarters or were tied by the opposing team in the 4th quarter---in other words, games that the Packers were never really behind in the 4th quarter. The Cowboys web site did the exact same thing for Roger Staubach giving him credit for 7 comebacks in games which the Cowboys were never behind. I count at least 3 games, one from 2004 and 2 from 2003 which Favre was given credit for comebacks in games the Packers were never behind in (or at the beginning of) the 4th quarter.

The best article I've been able to find on this subject mentioned that Favre had 27 comebacks entering the Nov. 23, 2003 game against SF. Favre added 2 more comebacks that season for a seeming total of 29. If everything else is correct, he would have ended up with 38 comebacks---which is one of the figures that the Packers web site comes up with. What that means is that Favre would have to had 17 comebacks between 1992-95---an aveage of better than 4 per season.

Several questions remain. How accurate is that figure of 17 comebacks between 1992-95? Also, how many of those comebacks occurred at home? The reason why this question is fascinating is that Favre was a great "homer" and he scored an inordinate number of his comebacks at home. My guess would be that Favre had only 8 road comebacks out of his 38 total. In comparison, Unitas had 15 of his 31 comebacks on the road; Marino had 16 of 37; Elway had 18 of 46; Staubach 10 of 16; Kelly 13 of 23; Montana 23 of 30.

So, what on the surface seems a sure bet, isn't a sure bet at all and, in fact, is difficult to ascertain. That is why Mr. McKinley's work is so fascinating. There is a wonderful research project made possible by McKinley staring us in the face---if some youngster is willing to tackle it. It would be a comparison of how many comebacks a QB attains during his career vs. the number of comebacks he would be expected to attain on aveage. For example, with Favre, take the total number of games that he started in his career (regular season and post-season). Multiply that figure by 23.2% (the percentage of games decided by comebacks). Divide that figure by 2 to give us an estimate of how many comebacks Favre should have on aveage. If my figuring is correct, it should come out to a total of 32 comebacks. But Favre had what appears to be 38 comebacks. Divide 38 by 32 and you have a figure to work with. Favre's figure would be 1.2 which is positive. In other words, Favre had more comebacks than the aveage QB would be expected to have given the same number of starts. Do the same thing with other QBs and see if you have any trends.

Well, I've gone back and, using the line score data from the PFRA site, counted up the number of comebacks that Brett Favre had between 1992-95. The total is surprising---only 8 comebacks: 1992---3H; 1993---2H, 2A; 1994---1H; 1995---zero. We had already established that Favre had 21 comebacks, 17 at home and 4 on the road, between 1996-2007. That means that Favre's career total for comebacks is actually 29---23 at home and 6 on the road. The Packers web site had credited Favre with between 35-40 comebacks (like the Cowboys and Patriots, they included tie games in the 4th quarter). Other web sites had credited Favre with anywhere between 31-43 comebacks.



What does this mean? Well, with 275 career starts, Favre should have 31.9 expected comebacks---if he were an aveage NFL QB. His comeback/expected comeback ratio suddenly becomes .909 instead of 1.2 What does this mean? Well, instead of having an extraordinary number of comebacks, Favre actually has fewer comebacks than an aveage QB in the NFL would be expected to have with the same number of starts! This will disappoint all the Brett Favre fans and writers who foam at the mouth about all of his spectacular comebacks. But wait---a golden rainbow emerges from the dark cloud. Fewer than expected comebacks is an indication of dominance---like Steve Young. So, while Favre loses some glory at one end of the Bell-curve, he gains prominance at the other.
Ain't research wonderful!

While I have your attention, I'd like to correct one other thing: Johnny Unitas had 16 of his 31 comebacks on the road, not 15. Other QBs that I've been able to accumulate data on: Ken Stabler had 19 comebacks, 11 on the road; Joe Theisman also had 19 comebacks, 8 on the road and one at a neutral site; Bart Starr had 17 comebacks, 10 on the road. Of necessity, these figures are approximations as it is not always clear whether starters are still in the game when the comebacks occur.

[Bob Carroll]

Clark Heins, on Apr 15 2008, 04:52 PM, said:

Well, I've gone back and, using the line score data from the PFRA site, counted up the number of comebacks that Brett Favre had between 1992-95. The total is surprising---only 8 comebacks: 1992---3H; 1993---2H, 2A; 1994---1H; 1995---zero. We had already established that Favre had 21 comebacks, 17 at home and 4 on the road, between 1996-2007. That means that Favre's career total for comebacks is actually 29---23 at home and 6 on the road. The Packers web site had credited Favre with between 35-40 comebacks (like the Cowboys and Patriots, they included tie games in the 4th quarter). Other web sites had credited Favre with anywhere between 31-43 comebacks.



What does this mean? Well, with 275 career starts, Favre should have 31.9 expected comebacks---if he were an aveage NFL QB. His comeback/expected comeback ratio suddenly becomes .909 instead of 1.2 What does this mean? Well, instead of having an extraordinary number of comebacks, Favre actually has fewer comebacks than an aveage QB in the NFL would be expected to have with the same number of starts! This will disappoint all the Brett Favre fans and writers who foam at the mouth about all of his spectacular comebacks. But wait---a golden rainbow emerges from the dark cloud. Fewer than expected comebacks is an indication of dominance---like Steve Young. So, while Favre loses some glory at one end of the Bell-curve, he gains prominance at the other.
Ain't research wonderful!

While I have your attention, I'd like to correct one other thing: Johnny Unitas had 16 of his 31 comebacks on the road, not 15. Other QBs that I've been able to accumulate data on: Ken Stabler had 19 comebacks, 11 on the road; Joe Theisman also had 19 comebacks, 8 on the road and one at a neutral site; Bart Starr had 17 comebacks, 10 on the road. Of necessity, these figures are approximations as it is not always clear whether starters are still in the game when the comebacks occur.

Did I miss something? Did he just say that when he gets a different answer from what he expects, it shows that his own system is wrong? That what he came up with means something else? That's a certifiable definition of nonsense.

But maybe I'm missing something. I still can't find a definition of a comeback within all the verbiage. I get the impression that he doesn't count ties but everybody else does. I was thinking. Suppose a team enters the final quarter trailing by 21 points; the quarterback gets hot and leads his team to three unanswered touchdowns but his team misses the try for a two-point coversion and loses by a point. The poor quarterback gets no credit for a comeback. The next week in a mud bath ,neither team can muster an offense. With five minutes left, our quarterback is watching from the sideline while his opponent tries to protect a 3-0 lead. Suddenly, there's a fumble! The ball squishes backward over the goal line and is recovered by the defense for a touchdown. Our QB gets a 7-3 comeback.

Who says fact is stranger than fiction?

[Bryan Lutes]

Clark Heins, on Apr 15 2008, 03:52 PM, said:

Joe Theisman also had 19 comebacks, 8 on the road and one at a neutral site.

I do remember that big 'neutral site' comeback of Thiesmann's (do neutral site comebacks count for more than a home comeback, but less than a road comeback?).

At the end of the 3rd quarter, the Redskins were trailing by a whole 4 points (17-13) but were driving down the field with their running game. The Redskins had gained 38 yards rushing on the drive and looked primed to take the lead, but Thiesmann threw an INT at the goalline.

The Redskins D forced a 3-and-out. 11 runs, 0 passes, and 52 yards later, the Redskins were winning 20-17.

The Redskins D forced another 3-and-out. 10 runs and 1 pass later, Thiesmann then tossed a 6-yard TD pass to Charlie Brown for a 27-17 lead.

The Redskins D forced a turnover on downs after 4 plays. 4 runs and 0 passes later, the miraculous comeback was complete.

The Redskins offense had run the ball 27 times in the 4th period, thrown the ball only 4 times (one of which was intercepted at the goalline), and the defense did not allow a first down in the entire period.

Credit Joe Thiesmann.

[Clark Heins]

Bryan Lutes, on Apr 16 2008, 07:23 AM, said:

I do remember that big 'neutral site' comeback of Thiesmann's (do neutral site comebacks count for more than a home comeback, but less than a road comeback?).

At the end of the 3rd quarter, the Redskins were trailing by a whole 4 points (17-13) but were driving down the field with their running game. The Redskins had gained 38 yards rushing on the drive and looked primed to take the lead, but Thiesmann threw an INT at the goalline.

The Redskins D forced a 3-and-out. 11 runs, 0 passes, and 52 yards later, the Redskins were winning 20-17.

The Redskins D forced another 3-and-out. 10 runs and 1 pass later, Thiesmann then tossed a 6-yard TD pass to Charlie Brown for a 27-17 lead.

The Redskins D forced a turnover on downs after 4 plays. 4 runs and 0 passes later, the miraculous comeback was complete.

The Redskins offense had run the ball 27 times in the 4th period, thrown the ball only 4 times (one of which was intercepted at the goalline), and the defense did not allow a first down in the entire period.

Credit Joe Thiesmann.

Joe Theisman---1982 NFL Season: Bert Bell Award as the best American Pro Football Player. Walter Payton Award for his volunteer and charity work. Top rated passer in NFC.

"At QB, Joe Theisman was the perfect complement to the running game. He led the league in accuracy, completing 64% of his passes. Although he threw only 13 TD passes, his high percentage passes kept the Redskins moving relentlessly down the field."---quotation from Thinkquest.com

Joe Theisman---1983 NFL Season: MVP of NFL. Second leading passer in NFL. Led Redskins to record setting scoring performances.


Other notable facts: Finished second in Heisman voting 1970. Academic All-American at Notre Dame. Only NFL QB to lead his team to both the Grey Cup and the Super Bowl. Three time All-Pro. QB rating of 91.7 in post-season games (higher than Tom Brady's). Comeback/Expected comeback ratio almost identical to Brady's. Between 1982-84, compiled a 36-8 w-l record. One of only three NFL QBs (Plunkett and Elway) to have posted 4 post-season wins in a single season. Among the top five winning percentages in NFL history during December---along with Brady, Favre, Montana and Staubach. Still holds most of the Redskins career passing records. I believe he still holds the NFL record for highest yards per pass attempt in post-season games---8.4. One of his Redskin teammates described Theisman, wearer of the single-bar facemask, as, "The toughest player I ever saw."

Not bad for a kid who showed up at Notre Dame weighing 147 pounds; the stunned Notre Dame coach said, "You gotta be kidding."

[Gabe]

Clark Heins, on Apr 16 2008, 02:58 PM, said:

Joe Theisman---Only NFL QB to lead his team to both the Grey Cup and the Super Bowl.

Actually, Joe Kapp led the British Columbia Lions to the Grey Cup in 1963 and 1964, and the Minnesota Vikings to the Super Bowl in 1969. I can only assume that there is something special about the first name, and had several QBs named Namath and Montana played a few years in the CFL, they too would have led their teams to the Grey Cup.

[Clark Heins]

Gabe, on Apr 16 2008, 08:40 PM, said:

Actually, Joe Kapp led the British Columbia Lions to the Grey Cup in 1963 and 1964, and the Minnesota Vikings to the Super Bowl in 1969. I can only assume that there is something special about the first name, and had several QBs named Namath and Montana played a few years in the CFL, they too would have led their teams to the Grey Cup.

You're right. I forgot all about Joe Kapp. I also somehow forgot about Eli Manning winning 4 in a row as a wild card.

Usually, to be successful in the CFL, a QB has to be a pretty good runner. Joe Namath didn't have the legs and Joe Montana's body was just too fragile. Theisman was a pretty good runner in his younger days, running back punts for the Redskins one year. Joe Kapp was sort of the Walter Payton of QBs as he just threw his body full-force into a tackler. No one had ever seen the likes of Kapp before. Billy Kilmer reminded me a bit of Kapp, but Kilmer did have the background as a running back.

[Rupert Patrick]

Clark Heins, on Apr 16 2008, 03:58 PM, said:

Joe Theisman---1982 NFL Season: Bert Bell Award as the best American Pro Football Player. Walter Payton Award for his volunteer and charity work. Top rated passer in NFC.

"At QB, Joe Theisman was the perfect complement to the running game. He led the league in accuracy, completing 64% of his passes. Although he threw only 13 TD passes, his high percentage passes kept the Redskins moving relentlessly down the field."---quotation from Thinkquest.com

Joe Theisman---1983 NFL Season: MVP of NFL. Second leading passer in NFL. Led Redskins to record setting scoring performances.


Other notable facts: Finished second in Heisman voting 1970. Academic All-American at Notre Dame. Only NFL QB to lead his team to both the Grey Cup and the Super Bowl. Three time All-Pro. QB rating of 91.7 in post-season games (higher than Tom Brady's). Comeback/Expected comeback ratio almost identical to Brady's. Between 1982-84, compiled a 36-8 w-l record. One of only three NFL QBs (Plunkett and Elway) to have posted 4 post-season wins in a single season. Among the top five winning percentages in NFL history during December---along with Brady, Favre, Montana and Staubach. Still holds most of the Redskins career passing records. I believe he still holds the NFL record for highest yards per pass attempt in post-season games---8.4. One of his Redskin teammates described Theisman, wearer of the single-bar facemask, as, "The toughest player I ever saw."

Not bad for a kid who showed up at Notre Dame weighing 147 pounds; the stunned Notre Dame coach said, "You gotta be kidding."

Ben Roethlisberger also led the 2005 Steelers to four post season wins in a season.

I've often wondered if the Redskins had beaten the Raiders in Super Bowl 18 if that would have been enough to put Theismann into the HOF. With the leg injury that forced him out of the game in 1985, he would have been viewed as having a short but brilliant career, although it took him a few years to get his chance to start in the NFL, and a sympathy argument could have been made for him. All the other QB's who won back to back Super Bowls have gone on to Canton (as Tom Brady will one day), and it would be tough to keep Theismann out I think when coupled with the most famous broken leg in NFL history. Had the Redskins beaten the Raiders, I think the 83 Redskins would be remembered today as one of the greatest team-seasons in NFL history and arguably the second strongest team-season of the 80's next to the 85 Bears. But Theismann had a terrible day and the Redskins choked and less than two years later his NFL career was over.

[Gabe]

Clark Heins, on Apr 16 2008, 08:14 PM, said:

You're right. I forgot all about Joe Kapp. I also somehow forgot about Eli Manning winning 4 in a row as a wild card.

Usually, to be successful in the CFL, a QB has to be a pretty good runner. Joe Namath didn't have the legs and Joe Montana's body was just too fragile. Theisman was a pretty good runner in his younger days, running back punts for the Redskins one year. Joe Kapp was sort of the Walter Payton of QBs as he just threw his body full-force into a tackler. No one had ever seen the likes of Kapp before. Billy Kilmer reminded me a bit of Kapp, but Kilmer did have the background as a running back.

Well, so much for my theory that having the first name, Joe is the key to success in both the CFL and NFL.

[Clark Heins]

Rupert Patrick, on Apr 16 2008, 10:57 PM, said:

Ben Roethlisberger also led the 2005 Steelers to four post season wins in a season.

I've often wondered if the Redskins had beaten the Raiders in Super Bowl 18 if that would have been enough to put Theismann into the HOF. With the leg injury that forced him out of the game in 1985, he would have been viewed as having a short but brilliant career, although it took him a few years to get his chance to start in the NFL, and a sympathy argument could have been made for him. All the other QB's who won back to back Super Bowls have gone on to Canton (as Tom Brady will one day), and it would be tough to keep Theismann out I think when coupled with the most famous broken leg in NFL history. Had the Redskins beaten the Raiders, I think the 83 Redskins would be remembered today as one of the greatest team-seasons in NFL history and arguably the second strongest team-season of the 80's next to the 85 Bears. But Theismann had a terrible day and the Redskins choked and less than two years later his NFL career was over.

The 1983 Redskins were one of the greatest offensive machines I ever saw---and they were a great joy to watch---the "Fun Bunch". This past year, as the Patriots kept racking up all those points, I kept remembering how the Redskins got beat in the Super Bowl and wondered if history might repeat itself.
As much of a competitor as Theisman was, his career stats just don't add up for HOF status. Like a lot of other QBs, he had problems with INTs. I don't think the leg injury would have mattered because it seemed to me that he slowed down a lot by 1985 as he appeared to have put on weight and wasn't mobile anymore. His mobility was his greatest asset and Joe Gibbs used the moving pocket to get Theisman away from the huge linemen who would block his vision. He could also use his running ability outside the traditional pocket.