[Bob Carroll]

Bob Gill, on Apr 25 2008, 12:46 AM, said:

>I am constantly amazed at how willing some Forum members are at tearing into another person's research, not for the purpose of offering constructive criticism or helpful suggestions, but for the sole purpose of making up a tasteless and stupid joke.

I certainly don't have any constructive criticism to offer because, as I said here years ago, I think won-lost records for quarterbacks are fatally flawed in concept. I don't think it's a question of tweaking something to fix a small problem or make a number more accurate; I think it's not a valid statistic in the first place, like those TV graphics comparing each quarterback's passer rating after the first quarter. That's all.

Clark: here is some constructive criticism:

Gather all your e-mails, notes, and memos about Comeback Wins. Put them all together. When you have constructed a large pile, strike a match and burn the whole thing.

[Clark Heins]

Bob Gill, on Apr 25 2008, 12:46 AM, said:

>Dear Mr. Gill,
>For someone who is suppose to be an experienced, intelligent, and responsible researcher, you have missed several salient points of information: First, and foremost, you mistakingly misidentify me with Joe Montana. I was merely quoting an opinion expressed by Montana. However, I do believe that Montana's opinion is correct.

No, I didn't do that at all. My quote of your post clearly included the fact that Montana had made the original statement. Still, I wonder why you even try to make that claim, since in the next breath you admit that you agree with Montana's statement.


>I am constantly amazed at how willing some Forum members are at tearing into another person's research, not for the purpose of offering constructive criticism or helpful suggestions, but for the sole purpose of making up a tasteless and stupid joke.

I certainly don't have any constructive criticism to offer because, as I said here years ago, I think won-lost records for quarterbacks are fatally flawed in concept. I don't think it's a question of tweaking something to fix a small problem or make a number more accurate; I think it's not a valid statistic in the first place, like those TV graphics comparing each quarterback's passer rating after the first quarter. That's all.

Well, many people feel that way. You're not the only one. I agree with you about those TV graphics as they are absolutely ridiculous. But, like w-l records, they are sexy things which appeal to the public's imagination.
I took my own advice and went back and re-read Mr. McKinley's article and discovered something interesting that I hadn't paid attention to before. He writes, "Obviously, many comeback attempts prove futile. While 603 games since 1996 have featured come-from-behind wins, another 1322 games have ended with close trailer still behind when the final gun sounded."

Let us consider this data: We know, from the information provided to us by Mr. McKinley, that some 56.7% of all games (1474 out of 2598) provide an opportunity for an 8 point or less comeback win in the fourth quarter. (P.S. Once the fourth quarter lead reaches nine or more points, the chances of a comeback win becomes very slim---I would estimate less then 5%. Once the lead reaches ten or more points, only a very small amount of comeback wins occur). However, occuring to Mr. McKinley's figures, by the close of these 2598 games, some 1995 (76.8%) of them "have ended with a close trailer" or have been won outright with a comeback win.
Consider what these figures mean. At the start of the fourth quarter or at sometime during the fourth quarter, the team lbehind is behind by 8 or fewer points in 56.7% of all games played (i.e., 43.3% of the games are pretty well decided). However, by game's end, the trailing teams have either scored a comeback win or have sliced the lead down to eight points or less in 76.8% of all games. What that means is, that somewhere during the course of the fourth quarter, the trailing teams are making considerable headway at getting themselves back into the game. If fact, there has been a 35.5% change in momentum during the course of the fourth quarter. That doesn't mean that too many teams are going to achieve full comeback wins as they simply got themselvres behind by too many points, but it does tell us that the teams who are behind in the fourth quarter, by whatever margin, are putting considerable pressure on those teams who were ahead. It also tells us that it takes a pretty good QB to hold onto the lead in the fourth quarter when faced with this pressure from the opposing team and the opposing QB, both of whom have nothing to lose and, essentially, are playing with house money.

When you begin to add in all the intrinsic factors and take a close look at the available statistics, Montana's opinion doesn't seem so far fetched as it once appeared. Of course, we will never know for sure unless somebody is willing to take a chance and make an effort at making a quantitative study of (for lack of a better term) "QB Dominance". Just as figuring out "QB Comeback Wins Efficiency", it can be done thanks to the pioneering research efforts of Mr. McKinley who really wasn't thinking in this vein, but whose research has opened up many vistas unexpectedly.

[Clark Heins]

Clark Heins, on Apr 25 2008, 04:14 PM, said:

Well, many people feel that way. You're not the only one. I agree with you about those TV graphics as they are absolutely ridiculous. But, like w-l records, they are sexy things which appeal to the public's imagination.
I took my own advice and went back and re-read Mr. McKinley's article and discovered something interesting that I hadn't paid attention to before. He writes, "Obviously, many comeback attempts prove futile. While 603 games since 1996 have featured come-from-behind wins, another 1322 games have ended with close trailer still behind when the final gun sounded."

Let us consider this data: We know, from the information provided to us by Mr. McKinley, that some 56.7% of all games (1474 out of 2598) provide an opportunity for an 8 point or less comeback win in the fourth quarter. (P.S. Once the fourth quarter lead reaches nine or more points, the chances of a comeback win becomes very slim---I would estimate less then 5%. Once the lead reaches ten or more points, only a very small amount of comeback wins occur). However, occuring to Mr. McKinley's figures, by the close of these 2598 games, some 1995 (76.8%) of them "have ended with a close trailer" or have been won outright with a comeback win.
Consider what these figures mean. At the start of the fourth quarter or at sometime during the fourth quarter, the team lbehind is behind by 8 or fewer points in 56.7% of all games played (i.e., 43.3% of the games are pretty well decided). However, by game's end, the trailing teams have either scored a comeback win or have sliced the lead down to eight points or less in 76.8% of all games. What that means is, that somewhere during the course of the fourth quarter, the trailing teams are making considerable headway at getting themselves back into the game. If fact, there has been a 35.5% change in momentum during the course of the fourth quarter. That doesn't mean that too many teams are going to achieve full comeback wins as they simply got themselvres behind by too many points, but it does tell us that the teams who are behind in the fourth quarter, by whatever margin, are putting considerable pressure on those teams who were ahead. It also tells us that it takes a pretty good QB to hold onto the lead in the fourth quarter when faced with this pressure from the opposing team and the opposing QB, both of whom have nothing to lose and, essentially, are playing with house money.

When you begin to add in all the intrinsic factors and take a close look at the available statistics, Montana's opinion doesn't seem so far fetched as it once appeared. Of course, we will never know for sure unless somebody is willing to take a chance and make an effort at making a quantitative study of (for lack of a better term) "QB Dominance". Just as figuring out "QB Comeback Wins Efficiency", it can be done thanks to the pioneering research efforts of Mr. McKinley who really wasn't thinking in this vein, but whose research has opened up many vistas unexpectedly.

I'm sorry. My eyes betrayed me again. The figure 1995 should be 1925 instead.

[Bob Carroll]

Clark Heins, on Apr 25 2008, 04:27 PM, said:

I'm sorry. My eyes betrayed me again. The figure 1995 should be 1925 instead.

As my sainted Irish grandmother used to say, "Mein Gott und Himmell! Vot a dunderskite!" In most games, a team with a two-score lead will attampt to run out the clock. Although they may not start it as soon as the fourth quarter becgins, vert soon the leading team usually increases its rushing attempts and decreases its passing. Their opponents, not being complete idiots, expect this maneuver and will play run-defense. When the trailing side gets the ball, they are often confronted with a defense perfectly willing to trade short gains for time. The result is often a closer final score but the team that was in front tucks a win into its record.

Most of us understood this when we were in grade school. We couldn't have told you the exact numbers, but we could have guessed the percentages. We certainly wouldn't have traced it to the mystical reasons Clark, the QB-Profiler, wants to believe.

[Bryan Lutes]

Clark Heins, on Apr 24 2008, 10:49 PM, said:

I am constantly amazed at how willing some Forum members are at tearing into another person's research, not for the purpose of offering constructive criticism or helpful suggestions, but for the sole purpose of making up a tasteless and stupid joke.

Whenever I have talked specifically about this Comeback Wins subject, I've gotten zero response from Clark. I don't understand why Clark continues to frequent this forum topic, since he has no desire to actually discuss the issue. I find the subject matter to be interesting, but the ideas put forth by Clark and the fascinating Mr. McKinley are just so illogical that something needs to be said.

Clark Heins, on Apr 25 2008, 03:14 PM, said:

Let us consider this data: We know, from the information provided to us by Mr. McKinley, that some 56.7% of all games (1474 out of 2598) provide an opportunity for an 8 point or less comeback win in the fourth quarter.

Just as figuring out "QB Comeback Wins Efficiency", it can be done thanks to the pioneering research efforts of Mr. McKinley who really wasn't thinking in this vein, but whose research has opened up many vistas unexpectedly.

Mr. McKinley's research efforts are far from "pioneering". Unfortunately, since Clark continues to trumpet McKinley's work, I have to continue to "tear into" McKinley's research. As I have said in the past, McKinley's research is flawed from the start because 1) his definition of a "comeback" is incorrect, and 2) he does not take QB performance into account.

I don't understand what is meant by saying McKinley's research has opened up many vistas unexpectedly.

[Clark Heins]

Bryan Lutes, on Apr 29 2008, 11:23 AM, said:

Whenever I have talked specifically about this Comeback Wins subject, I've gotten zero response from Clark. I don't understand why Clark continues to frequent this forum topic, since he has no desire to actually discuss the issue. I find the subject matter to be interesting, but the ideas put forth by Clark and the fascinating Mr. McKinley are just so illogical that something needs to be said.





Mr. McKinley's research efforts are far from "pioneering". Unfortunately, since Clark continues to trumpet McKinley's work, I have to continue to "tear into" McKinley's research. As I have said in the past, McKinley's research is flawed from the start because 1) his definition of a "comeback" is incorrect, and 2) he does not take QB performance into account.

I don't understand what is meant by saying McKinley's research has opened up many vistas unexpectedly.

Bryan---Look around the football research world. Check out the football encyclopedias and all the web sites. Do you notice anything? There is not a single solitary footnote in any reference quide, other than the NFL Record and Fact Book and a few media guides, concerning QB wins and losses and the NFL Record Book only goes back to the 1987 season! Because people like Ken Pullis and Tod Maher and a half dozen others went back and spent years doing the research, we now have QB w-l records going back to 1950s. And, because of the research efforts of others, we even have the w-l records of the pre-1950 so-called "Primary QBs". Somebody had to have the gumption and the dedication and the courage to spent thousands of hours and a lot of money out of theior own pockets digging through newspaper archives to accomplish this, something thought unimportant by the sceptics. But, IT WAS DONE!

And now two more challenges have come along, indirectly presented to us by Mr. McKinley's research: 1. First, how efficient are QBs at obtaining comeback wins? Once again, look around and see if you can find any information, from the NFL or any other source even remotely, telling us how many comeback wins a QB might have accomplished during his career. Well, we have the HOF spewing out a total for John Elway. Of course, that total is completely wrong--- as are the totals for many other QBs---that is, if you can find such figures in the media guides of the various teams. Mr. McKinley was the first and only researcher to ever have touched upon this subject, but even his research is limited by the ten year time frame he used---from 1996-2005. Nor does he separate road and home comeback wins. As far as I know, I'm the only person to have ever attempted finding out these pair of statistics. Why, I don't know---because the research is relatively easy and can be done in a relatively short period of time by any novice willing to make the effort. All one has to do is review the winning games that each QB started and try to determine whether the starting QB was still in the game when the comeback win was achieved. It isn't a 100% fool-proof system, but it gives us a pretty good approximation of how efficient a QB is in the clutch. To make this type of research fool-proof, one would have to have actual game sheets---and the only place I know where to obtain those are in Canton. If for no other reason, such research can and has already corrected many inaccurate records published in media guides and even, in Elway's case, found in the HOF. YES, IT CAN BE DONE and, if I live long enough, it will be done!
2. The second possibility that Mr. McKinley's research presents is a quantitative approximation of how dominant each QB is in manufacturing and then maintaining fourth quarter leads. Once again, look around and see if you can find even the remotest hint of such information presented anywhere within the NFL or outside it. It does not exist because no one has had the courage to attempt such a project before and, even if they wanted to, had no standard by which to base such research upon. Well, now we have a standard---which has been provided to us by Mr. McKinley---and now a world of possibilities open up despite the fact that this type of research is difficult and presents many problems. It is twice as difficult as examining QB comeback wins because the researcher is forced to examine all of a QB's starts---not just wins. I doubt that I'll live long enough to even scratch the surface, but, like comeback wins research, IT NOW CAN BE DONE! But it will take more than merely time and effort and money. It takes the courage to disregard all the nay-sayers, the critics, the second-quessers, and follow your own instincts and what your own mind and heart tells you.

And not by eastern windows only,
When the daylight comes, comes in the light,
In front the sun climbs slow, how slowly,
But westward, look, the land is bright.

From "Say Not The Struggle Nought Availeth"
By Arthur Hugh Clough

[paulksandiego]

Mr. Heins,

What I don't understand is that you push this concept about comeback wins and in another thread degrade the importance of all-pro teams, saying that the concept of all-pro teams is a thing of the past. I won't even get specific as to what I really think about the importance of a quarterback's comeback wins, but let's just say I think some of your concepts are better suited for the Coast-To-Coast AM radio show...I apologize to the administrators, but I'm having a hard time holding my tongue after reading some of this guy's posts over the past week....Remember, this is just my opinion, so no legal threats need to be made. I have no proof that this guy is crazy, it's just my opinion.

Paul

[Bob Gill]

> The second possibility that Mr. McKinley's research presents is a quantitative approximation of how dominant each QB is in manufacturing and then maintaining fourth quarter leads. Once again, look around and see if you can find even the remotest hint of such information presented anywhere within the NFL or outside it. It does not exist because no one has had the courage to attempt such a project before ...


I must admit there's no way I can find even the remotest hint of how good a quarterback is at manufacturing and maintaining a lead, or leading a comeback for a victory. I have no idea about how well Bart Starr, just to take one name, did that. I do know that Starr had outstanding stats as a passer, leading the league several times, and that his team won a very high percentage of its games, including several championships, but that could mean anything. It gives me no clue about Starr's dominance, or how he compares to Neil O'Donnell or Ken Stabler. I can only hope that at some point somebody will have the courage (and yes, it certainly does take courage to look at all that microfilm, as I know from the hundreds of hours I've spent doing so; my friends call me Rambo, for just that reason) to tell me how many fourth-quarter comebacks he led (that is, comebacks from no more than 8 points behind, since comebacks from big deficits don't count), how many road comebacks he presided over, and whether he could do the even more difficult skill (we know it's harder to do since it happens much more often) of maintaining a lead in the fourth quarter. Then maybe I'll know whether he deserves his spot in the Hall of Fame, or not.

[Bryan Lutes]

Clark Heins, on Apr 29 2008, 07:44 PM, said:

Bryan---Look around the football research world. Check out the football encyclopedias and all the web sites. Do you notice anything? There is not a single solitary footnote in any reference quide, other than the NFL Record and Fact Book and a few media guides, concerning QB wins and losses and the NFL Record Book only goes back to the 1987 season! Because people like Ken Pullis and Tod Maher and a half dozen others went back and spent years doing the research, we now have QB w-l records going back to 1950s. And, because of the research efforts of others, we even have the w-l records of the pre-1950 so-called "Primary QBs". Somebody had to have the gumption and the dedication and the courage to spent thousands of hours and a lot of money out of theior own pockets digging through newspaper archives to accomplish this, something thought unimportant by the sceptics. But, IT WAS DONE!

Once again, not only is your response (which, for reasons unknown, is a direct response to me) completely irrelevant, it also implies that I am denigrating the Research Gods by questioning your illogical conclusions.

This reinforces my previous comment that you have no desire to actually discuss the topic of comeback wins.

Clark Heins, on Apr 29 2008, 07:44 PM, said:

And now two more challenges have come along, indirectly presented to us by Mr. McKinley's research: 1. First, how efficient are QBs at obtaining comeback wins? Once again, look around and see if you can find any information, from the NFL or any other source even remotely, telling us how many comeback wins a QB might have accomplished during his career.

Again, you aren’t discussing the specifics of Mr. McKinley’s research…you are merely reiterating your own personal opinion that Mr. McKinley’s research is awesomely fascinating and groundbreaking. There used to be a guy who was very perturbed that a player’s fumble return yards weren’t universally available. To his credit, I don’t remember him ever saying Earnest Byner was better than Leroy Kelly because of fumble return yards, and I don’t remember him ever saying that Jim Brown was overrated because he had a very low fumble return yardage total in away games.

To be blunt, Mr. McKinley’s research does not show how efficient QBs are at obtaining comeback wins, mainly because Mr. McKinley’s criteria for a comeback 1) does not even remotely touch upon a QB’s performance, and 2) categorizes all comeback opportunities as being the same, even if a team was down by 1 point at the start of the 4th quarter.

Clark Heins, on Apr 29 2008, 07:44 PM, said:

Mr. McKinley was the first and only researcher to ever have touched upon this subject, but even his research is limited by the ten year time frame he used---from 1996-2005. Nor does he separate road and home comeback wins. As far as I know, I'm the only person to have ever attempted finding out these pair of statistics. Why, I don't know---because the research is relatively easy and can be done in a relatively short period of time by any novice willing to make the effort. All one has to do is review the winning games that each QB started and try to determine whether the starting QB was still in the game when the comeback win was achieved.

That is not the issue. But you seem to do a great job of promoting your own research efforts, even when we are discussing Mr. McKinley’s fascinating research.

Clark Heins, on Apr 29 2008, 07:44 PM, said:

It isn't a 100% fool-proof system, but it gives us a pretty good approximation of how efficient a QB is in the clutch.

This appears to be the crux of the discussion, yet you have buried it somewhere in the middle of paragraph two. Why does it give us a pretty good approximation of how efficient a QB is in the clutch if it doesn’t look at the performance of the QB, and the definition of ‘clutch’ is trailing by 1 point with 15 minutes left on the clock?

Clark Heins, on Apr 29 2008, 07:44 PM, said:

It does not exist because no one has had the courage to attempt such a project before and, even if they wanted to, had no standard by which to base such research upon. Well, now we have a standard---which has been provided to us by Mr. McKinley---and now a world of possibilities open up despite the fact that this type of research is difficult and presents many problems.

Pretty soon Mr. McKinley will be put forth for research martyrdom? The research I find most interesting is the research has a logical foundation. St. McKinley’s research is somewhat illogical, so the findings of his research are probably met with a shrug of the shoulders and a “so what?” utterance.

[Ken Crippen]

The "Ignore" feature works very well on this board. Just add the username to the list and you never have to read another post from that user again (unless it is quoted by someone not on your ignore list). It is under the "My Controls" link at the top of the board. Follow the "Manage Ignored Users" link about midway down the list along the left side of the page.

Not that I am implying anything. I am just saying...

I would also like to add the following from the Terms of Service:

Quote

The PFRA strives to attract a high quality readership. Intelligent discussion goes hand in hand with minimal moderation. Moderators are on duty to handle occasional conflicts and occasional lapses in judgment. If, over the course of time, a poster's online personality/behavior develops in a way that they become a source of continual conflict then the PFRA reserves the right to restrict the posting privilege of that person. It is up to individuals to conduct themselves in a manner that is socially acceptable to the community at large. The PFRA does not allow mischievous activity from disruptive posters that would result in decreased readership or a devaluation of the product.

Again, I am just saying...

[fgoodwin]

Bryan Lutes, on Apr 29 2008, 10:23 AM, said:

As I have said in the past, McKinley's research is flawed from the start because 1) his definition of a "comeback" is incorrect, and 2) he does not take QB performance into account.

Two questions:

(1) How would you define "comeback"?

(2) How would you take QB performance into account?

[Clark Heins]

Bob Gill, on Apr 29 2008, 09:42 PM, said:

> The second possibility that Mr. McKinley's research presents is a quantitative approximation of how dominant each QB is in manufacturing and then maintaining fourth quarter leads. Once again, look around and see if you can find even the remotest hint of such information presented anywhere within the NFL or outside it. It does not exist because no one has had the courage to attempt such a project before ...


I must admit there's no way I can find even the remotest hint of how good a quarterback is at manufacturing and maintaining a lead, or leading a comeback for a victory. I have no idea about how well Bart Starr, just to take one name, did that. I do know that Starr had outstanding stats as a passer, leading the league several times, and that his team won a very high percentage of its games, including several championships, but that could mean anything. It gives me no clue about Starr's dominance, or how he compares to Neil O'Donnell or Ken Stabler. I can only hope that at some point somebody will have the courage (and yes, it certainly does take courage to look at all that microfilm, as I know from the hundreds of hours I've spent doing so; my friends call me Rambo, for just that reason) to tell me how many fourth-quarter comebacks he led (that is, comebacks from no more than 8 points behind, since comebacks from big deficits don't count), how many road comebacks he presided over, and whether he could do the even more difficult skill (we know it's harder to do since it happens much more often) of maintaining a lead in the fourth quarter. Then maybe I'll know whether he deserves his spot in the Hall of Fame, or not.

Dear Mr. Gill,
Just one minor correction---the comeback totals that Mr. McKinley talked about do include comebacks from greater than 8 points. I too have have some confusion in my mind about figuring out everything Mr. McKinley is saying. Oddly, the proposed equation (Comeback Wins/Expected Comeback Wins) that I came up with actually coincides with Mr. McKinley's t-test figuration (no, I swaer I didn't steal the idea from him), but his is much more sophisticated than mine as it includes things like variances which is fair beyond by math skills.

Football statistics are changing more rapidly then we are aware. In the good old days, a QB was judged by the number of yards he threw for in a season. Then, during the early 1970s, we began rating the QBs by the QB rating system---still in use today. But, most of us realize that judging QBs based upon statistical aveages which occurred between 1960-69, is no longer very logical.

In today's football world, we are witnessing three grand confluences of statistical information which will sweep aside everything we have ever known before. The first of these is the PFRA's Football Project---an attempt to map all the plays from all the games since the NFL started---over ten thousand games in all. Right now, this type of data is only available to a select few. When it becomes available to the general public, it will open up research opportunities to everyone. Most of all, it will open up research opportunities into QBs giving us situational statistics for each QB.
The second project under way isn't quite as encompassing, but it is just as important. Pro-Football Reference.com is planning to publishing the box scores from these ten thousand NFL games. Presently, we only have access to box scores going back to 1994---thanks to John Troan's web site. Box scores give us much more information than simple line scores. As an example, Pro-Football Reference.com's founder Doug Drinen points out that, with box scores, a researcher can determine that QB Richard Todd's TD passes resulted in his team either tying or going ahead in a game approximately 50% of the time while the TD passes of QB Ken Anderson resulted in his team tying or going ahead in only about 30% of the games. Such studies, esoteric as they might seem, can be useful in trying to gage a QB's worth.
The third major thrust is represented by the statisticians like Mr. McKinley and Mr. Drinen. If you follow Mr. Drinen's ramblings, he comes up with the most interesting statistical data about players and teams, but, essentially, he is like most of us---bean counters. Mr. McKinley represents the new breed on the block, the breed found on Football Outsiders.com---the web site Mr. McKinley's article appeared. These guys use statistical models that are incomprehensible to most of us, but, basically, these models involve one thing---making comparisons against a standard. The idea is not new---the NFL stat people did the same thing with QB ratings, but the degree of sophistication is greater.

These three things are eventually going to come together and blow the lid off of our analysis of, not only QBs, but every other skilled position player. But, it might take years for all of this to come about. In the meanwhile, we must use the tools we have available. I can't see into the future, but Mr. McKinley hinted at the possibility. We have the means and the ability to gage QB effectiveness in two compelling areas---comeback ability and maintaining leads. I am not a statistician, but if we can combine these two abilities into one equation, then we can do what every researcher tries to do---make sense out of chaos.

[Aricg]

To me, a game winning drive that happens in the last 5 minutes of the game is far more important than a comeback win calculated during the entire 4th quarter. If the game is tied or your team is trailing in the final 5 minutes, how do quarterbacks perform? This is crunch time and holds a lot more weight to me than a regular comeback.

Think about the great drives that are remembered or great comebacks. These happened in the final minutes of a game. John Elway and The Drive (98 yards) in Cleveland against all odds. Joe Montana in the Super Bowl against the Bengals - 92 yards for a touchdown. There have been many others. So what I guess I am saying is how does a player perform with the game on the line is of great importance in the value of a quarterback, much more than if his team came back early in the 4th quarter.

To make it even more interesting, how many times when a quarterback trailed or was in a tie game with less than two minutes to play did they lead their team to go ahead points.

Just my two cents.

[Aricg]

I quickly looked at Elway's 47 comebacks according to the Hall of Fame and the Broncos. What is more impressive to me than the 47 comebacks is this.

In final 5 minutes of a game - 36 game winning drives
In final 2 minutes of a game - 25 game winning drives

I pulled these numbers very quickly, so I could have made an error by 1 or 2. I am sure if you look at Montana, Marino and some of the other greats - you would see that they would have similar stats. Where Jon Kitna or some of the other average quarterbacks with 4th quarter comebacks probably do not have most of these in the final minutes of the game. This is just a guess on my part, but it makes sense.

Bobby Layne didn't become famous because his team trailed at the beginning of the 4th quarter and then they scored at the beginning of the quarter to give them a lead. He became famous for his last minute heroics.

Once again, I thinking being Great in the Clutch is far more important than just a number. That is why people remember Montana, Layne, Staubach, Elway, Marino, and others and do not remember Testaverde.

[Bryan Lutes]

fgoodwin, on Apr 30 2008, 11:21 PM, said:

Two questions:

(1) How would you define "comeback"?

(2) How would you take QB performance into account?

{Apologies to Ken Crippen for lengthening this thread}

I give you full marks for persistence, if not actual comprehension. You have asked me these questions 3 times now, and I will have responded 3 times.

Before we start, I'd like to point out that the basis of this entire thread was a research piece done by Mr. McKinley of FootballOutsiders. In this piece, the definition of a comeback was if a team trailed by 8 points or less in the 4th quarter and ended up winning the game. I questioned the logic of such definition, because 1) as has been mentioned by several other people, the 1-point deficit with 14:59 remaining scenario exists, and 2) the actual results of the research demonstrated that either we should be preparing busts in Canton for Kerry Collins and Jay Fiedler or that the research concept itself was flawed.

I don't agree with your thought of 'until someone else proposes a better model, Mr. McKinley's must be acceptable'. I'm definitely not an expert on 'comeback wins research', but I believe I have enough basic understanding of football to distinguish logic from folly. I'm not a doctor, either, but I believe I would have enough sense to not go along with my physician's bloodletting treatment. Now, to answer your questions again:

1) As I said before, I would use a sliding scale. This is just a rough sketch without doing any research into the matter other than perusing this thread. We'll have 3 points in time...

Start of 3rd quarter - team must trail by 28 or more points
Start of 4th quarter - team must trail by 15 or more points
Last possession - team must need TD to win, or must need TD to tie then subsequently win in OT

I think this concept would at the very least eliminate the cheapo Jon Kitna 1 point comeback wins.

2) I don't quite understand this question. If it is meant literally, then my answer is that I would look at the QB's statistics during the 'duration of the comeback'. It would require the researcher to look at the play-by-play chart of the 'comeback win' that met the above criteria and track the performance of the QB. This would probably be a big task, much more difficult than looking at a box score, seeing who the QB was, then giving the QB credit for spearheading a late and dramatic rally. But in my mind at least the results would be somewhat logical and meaningful, as opposed to blindly awarding the QB credit.

As for what statistical criteria would a QB have to attain to get credit for a comeback win, that is an interesting question. I went back and looked at the baseball rulebook and found my original comment about starting pitchers needing to "pitch effectively" based on statistical criteria in order to be awarded a win was actually incorrect. There is no statistical criteria...the scorekeeper uses his own judgment to decide if a starting pitcher "pitched effectively" enough to be awarded a win. For QBs, I would say that a QB would have to account for at least 50% of his team's total offensive yardage during the 'duration of the comeback' in order to be credited with a comeback win. That is a totally arbitrary number, but at least it puts forth the idea that the QB actually needs to 'spearhead' a comeback win in order to be credited with 'spearheading a comeback win'.

[Bob Gill]

>As for what statistical criteria would a QB have to attain to get credit for a comeback win, that is an interesting question. I went back and looked at the baseball rulebook and found my original comment about starting pitchers needing to "pitch effectively" based on statistical criteria in order to be awarded a win was actually incorrect. There is no statistical criteria...the scorekeeper uses his own judgment to decide if a starting pitcher "pitched effectively" enough to be awarded a win.


Somewhat off the original subject, I admit, but I wanted to point out that the scorekeeper in baseball has no role in whether a starter gets a win or not. If the starter pitches at least 5 innings (the only requirement) and leaves with a lead that the team never loses, he gets the win. If he leaves after 8 innings leading 7-0, or if he leaves after 5 innings leading 13-9, he gets the win in either case, as long as his team maintains the lead from then on.

"Pitching effectively" is supposed to come into play in one specific and rather unusual case involving relievers, but this isn't a baseball discussion, so I won't get into that. I just wanted to note that there's no judgment involved in crediting wins to starting pitchers.

[Bryan Lutes]

Bob Gill, on May 2 2008, 08:35 AM, said:

>As for what statistical criteria would a QB have to attain to get credit for a comeback win, that is an interesting question. I went back and looked at the baseball rulebook and found my original comment about starting pitchers needing to "pitch effectively" based on statistical criteria in order to be awarded a win was actually incorrect. There is no statistical criteria...the scorekeeper uses his own judgment to decide if a starting pitcher "pitched effectively" enough to be awarded a win.


Somewhat off the original subject, I admit, but I wanted to point out that the scorekeeper in baseball has no role in whether a starter gets a win or not. If the starter pitches at least 5 innings (the only requirement) and leaves with a lead that the team never loses, he gets the win. If he leaves after 8 innings leading 7-0, or if he leaves after 5 innings leading 13-9, he gets the win in either case, as long as his team maintains the lead from then on.

"Pitching effectively" is supposed to come into play in one specific and rather unusual case involving relievers, but this isn't a baseball discussion, so I won't get into that. I just wanted to note that there's no judgment involved in crediting wins to starting pitchers.

That is correct.

[fgoodwin]

Bryan Lutes, on May 2 2008, 06:55 AM, said:

I give you full marks for persistence, if not actual comprehension. You have asked me these questions 3 times now, and I will have responded 3 times.

An answer is a response, but not all responses are answers. If I thought you had answered my questions, I would not have persisted in asking them.

Quote

I don't agree with your thought of 'until someone else proposes a better model, Mr. McKinley's must be acceptable'.

Speaking of comprehension, I never said that -- I said there's a difference between criticism and constructive criticism, and I also noted that no one has offered any better alternatives.

Thanx for the detailed answers.

[Ken Crippen]

I think that we have beaten this topic to death, resurrected it and beaten it again.

[Bob Carroll]

fgoodwin, on May 2 2008, 04:00 PM, said:

An answer is a response, but not all responses are answers. If I thought you had answered my questions, I would not have persisted in asking them.

Speaking of comprehension, I never said that -- I said there's a difference between criticism and constructive criticism, and I also noted that no one has offered any better alternatives.

Thanx for the detailed answers.

No one can offer an alternative to judging quarterbacks by comeback wins because comeback wins do not provide a way to judge quarterbacks. Providing an alternative? Is a lemon merengue pie an alternative to a rocket ship? Is aa anteater an alternative to a tea kettle?

Since comeback wins do not prove anything, the only thing that can be done with them is settle on some sort of definition and then count up the number of games in which a quarterback meets that definition. Unfortunately, since there doesn't seem to be a majority in agreement on a particular definition, we can't even do that.

If we did have an agreed to definition, and if we took the time to add up all the examples that fit, we'd have -- wow -- the number of games in which a team met the definition!. We would NOT, however, have a method by which any sane person would rank quarterbacks.