Comeback Wins is a Good Statistic, but Incomplete
Mike Sando, a Senior Writer at ESPN, published a list on Wednesday answering the question, “who is the best QB since 1978?” He called it the GOAT Index for Modern QBs. How Sando compiled the list was by outsourcing the question to 10 former coaches: Pete Carroll, Tony Dungy, Mike Holmgren, Howard Mudd, Wade Phillips, Mike Reinfeldt, Ray Rhodes, Mike Shanahan, Al Saunders and Norv Turner. By taking everybody’s opinion into account they actually came up with a fairly decent list (1. Brady | 2. P. Manning | 3. Montana | 4. Elway | 5. Rodgers | T-6. Marino | T-6. Favre | 8. Young | 9. Brees | 10. Fouts). Whether or not you agree with the list though is besides the point. One of the metrics they used, and one that multiple coaches pointed out when ranking a QB higher than another was “Comeback Wins.” Here is the example from Al Saunders along with his explanation of why he chose Tom Brady, Peyton Manning and Joe Montana as his top guys:
Notice how at the end he mentions comeback wins. To him, comeback wins is an extremely telling statistic because it reveals how well a guy plays in the fourth quarter and what type of quarterback he is in the clutch. To me though, comeback wins always falls short for two reasons that are strongly tied: it completely ignores how the quarterback performed in the previous three quarters, and it does not take into account how effective he was in producing comeback wins.
Fixing the Comeback Wins Statistic
Comeback wins can be a great statistic, but because it is a number and not a percentage, it does not tell the whole story. If you turn comeback wins into a percentage then everything changes and it becomes a truly fascinating statistic. To better understand what the simple tweak accomplishes, let us first take a look at where the top ten falls with comeback wins in its current iteration:
Notice, that list has some damn fan quarterbacks and you get a tremendous amount of overlap with the Top-10 GOAT Index from above. Also, notice the top two quarterbacks (Peyton Manning and Tom Brady) are flipped from the GOAT Index. If you took a look at this list, you might think Peyton Manning is a better performer in the clutch than Tom Brady. If you thought that without this list, you could be mistaken for having ingested too much peyote in your lifetime. Anybody that has paid even a modicum to football over the past two decades knows there is no better QB than Tom Brady in the clutch.
Already something is misleading about the statistics, but how to fix it? This is like looking at Hits in baseball, and not taking into account At Bats. Sure total hits is great, but 150 hits in 400 at bats is drastically different than 100 in 400. To complete our statistic and turn into into a percentage, we need a second sample set. But which one to use? We need something that shows us how effective a QB was at producing comeback wins, not just that he produced them. Maybe for that we can use overall record? Let us take a look at that list:
Here we have our top two guys from the GOAT Index flipped again. We see Tom Brady has a 208-61 record as a quarterback. Peyton Maning was 200-92. Now that we know how many wins each guy had, we can take into account what percentage of their wins came in comeback fashion. Brady has 39 comeback wins to his name out of 208 total victories. 39 divided by 208 is 18.75%. This tells us that fewer than one in five of Brady’s wins comes in comeback fashion. In other words, Brady does the required lifting prior to the fourth quarter in an astounding 81.25% of his victories. For Peyton Manning (45/200) he seals up victory before the fourth a whopping 77.5% of the time.
Take a look at the cutoff on the above chart; Aaron Rodgers. I did this because this form of the comeback statistic is extremely kind to Rodgers and shows there is something amiss with this form of the statistic. Rodgers has produced the same amount of comeback wins as Mark Sanchez (10). At first blush this might seem like a blemish on Rodgers’ record, but taking into account his 99 wins, that means Rodgers wraps up the win prior to the fourth quarter 89.9% of the time. You would think using our new and improved statistic Rodgers might be the best QB of all time because the guy just wins. But the way we have been using the statistic is flawed. It does do a good job of showing how well they played in the prior three quarters, but it does not reveal how well they played with the game on the line and their team down.
Introducing Comeback Wins Percentage
So is there a solution to fixing the comeback wins statistics? Yes, but instead of looking at the wins, we must look at the losses. In the flawed form of the statistic, Rodgers was a football deity. To better understand his fourth quarter performances, we have to take a look at the losses. Rodgers is 99-52 in his career. That means there has been a total of 62 chances for him to produce comeback wins (52 losses plus 10 comeback wins). To get a better gauge (I feel dirty using that word considering I am a Pats fan) on Rodgers’ fourth quarter prowess we have to take the 10 comeback wins and divide them by the comeback chances. Let us call the new statistic Comeback Wins Percentage. When we do that we have a number that better reveals how well he plays in the clutch: 16.13%. Aaron Rodgers, when down in the fourth quarter, produces comeback wins 16.13% of the time.
But what about our top two boys in each statistic, Peyton Manning and Tom Brady? Now that we have a better metric for evaluating four quarter effectiveness, we should get a better feel for each player’s clutchness, and the stats play out exactly like you would expect. Peyton Manning, with his 92 losses and 45 comeback wins has a Comeback Wins Percentage of 32.85%. That is a great number. But the man that is six comeback wins behind him from producing the most comeback wins of all time? Tom Brady? Tom Brady is a monster in Comeback Wins Percentage. Because Brady has only lost 61 games in his career, his Comeback Wins Percentage is an unreal 39%. In other words, when an opposing team is beating Tom Brady in the fourth quarter, there is almost a two in five chance they will go on to lose.
This makes sense in the overall discussion of who is the greatest quarterback of all time; the best quarterback should be a guy that no matter his surrounding personnel, he can snatch victory from the jaws of defeat. The greatest quarterbacks of all time should get better with the game on the line, and nobody does it better that Tom Brady. By slightly tweaking the comeback wins statistic and turning it into a percentage we now have a way to quantify just how good Tom Brady is with the game on the line. Not only is Tom Brady the GOAT overall, but Comeback Wins Percentage says Tom Brady is the greatest clutch quarterback of all time, pulling out victories 39% of the time when down in the fourth quarter.
Nobody that has played quarterback in the NFL has done it better than Tom Brady. He has had fewer opportunities to produce comeback wins because he is almost always ahead in the fourth quarter due to his work in the previous three quarters. And yet, despite the fewer opportunities, when one does arise, Brady is better than any other quarterback who ever lived. That is the sign of the greatest quarterback of all time; no matter what has happened throughout the game, no lead is safe… not even something like a 25 point lead at the end of the third quarter in the super bowl… as if anybody could come back from that type of deficit…