This isn't just for you, but for all the Kelly fans (of which I am one; I liked the guy. Class, style, humility...), you're starting a team. You have the choice between Jim Kelly, Dan Marino on one side, or John Elway, Roger Staubach, Peyton Manning, Bob Griese, Ben Roethlisberger, Bart Starr, Jim Plunkett and Eli Manning, I'm taking one of the latter group every single time. I might even take Warner, Favre and Theisman, too.
The object of a football game (or a football season) is not to make the Hall of Fame. It's to win the Super Bowl.
All of these guys over Kelly lol?
Elway: Probably doesn't even get the chance to make a fourth and fifth SB without Terrell Davis and TD was way more the reason they won XXXII than him and despite how iconic his helicopter dive was, he went 12/22 123 yds. 0 TD/1 INT in that game. He turns in that performance with an average RB, he probably loses the game and who knows if he ever makes it to another one with an 0-4 resume being in his head. Also, his average margin of defeat was 32.0 ppg in his losses. Even if you take Kelly's three worst SB losses it's only 21.7 ppg.
Staubach: Maybe. But his defense made SB VI a cakewalk for him by only allowing 3 points and getting 3 takeaways while the ground game averaged 5.25 yards a carry gaining 252 yards compared to his 119.
Here's a breakdown of the Cowboys' scoring drives in that game with the first group of letters being how the Cowboys gained possession of the football, the next part being their starting field position for the drive, and the last part being what type of score followed by how much yardage he accounted for.
FUM-DAL 48-FG[9 yd] (two completions for 29 yards)
Punt-DAL 24-TD (three completions for 34 yards including TD)
KOff-DAL 29-TD (one non-scoring completion for 12 yards)
INT-MIA 9-TD (one completion, a 7 yard TD)
SB XII could've been as bad a bloodbath as SB XXIV considering that the Cowboys D had 8 takeaways but Dallas had to settle for five FG attempts and missed three of them. The summary below shows that three of their five scoring drives were for 35 yards or less due to a combo of their defense being the real MVP and Denver being flat out horrible.
INT-DEN 25-TD (one non-scoring completion for 13 yards)
INT-DEN 35-FG (one incomplete pass)
Punt-DAL 43-FG [43 yd] (two completions for 30 yards)
Punt-DAL 42-TD (two completions for 58 yards and a TD)
FMB-DEN 29-TD (a one-play drive ended by a HB pass for TD)
In SB X, he lost 21-17 while throwing 3 picks and fumbling 3 times which were all recovered by Dallas and being sacked 7 times.
Oddly enough, SB XIII seems to be either his best SB stat line or close to it with him going 17/30 228 3 TD 1 INT but he was still sacked 5 times (no way of knowing which were his fault or the O-line's without going back and watching it though.)
My point through all of this Staubach research is that he was a good QB but far from noteworthy in his SB victories (considering the ridiculous amount of help he got from his running game and D) and due to the way Landry used him he was basically a game manager if you check the box scores from Dallas' scoring drives in their two SB wins with him.
Peyton Manning: I agree with you on that one.
Bob Griese: He missed 9 games during the perfect '72 season and literally only threw five passes in the AFC CG that year. How much credit are you really trying to give him for having one of the deadliest backfields in NFL history combined with a beyond-stout D?
Ben Roethlisberger: Literally a game manager during the SB XL and XLIII runs and then lost the only SB where he could even be argued as being anything more than a game manager. He was below 3500 yards five times in his first seven seasons and below 20 TDs five times as well. With passing stats as highly inflated as they are since 2000, that's a pretty clear indication of a game manager especially when you consider how many high quality RBs he's played with. He's better than that these days but can't seem to make a SB. The fact that he injured himself on a motorcycle clearly violating the terms of his contract and might've raped a girl are just more reasons why it'd be foolish to go with him over Kelly.
Bart Starr: Lombardi's power sweep is what made that offense run. This dude is good but tends to get overrated because of all the titles they won as a team, not a platoon led by a standout QB. He averaged about 16 attempts per game for his entire career. Not a very daunting task when your job is just to hand off for 70% of the game and make about 9 good passes a game.
Plunkett and Manning: You're nearly trolling at this point.
Warner and Favre: I have no problem with them
Theisman: I doubt it but I'll research it later if you want.
Again, blaming one guy for the efforts of 22.
Like I'm the first guy in the history of sports commentary to do that... having said that, how many times have we seen guys - Tom Brady; Kurt Warner was GREAT for this - make the guys around him better. Some guys do that, others yell and bark at their receivers for whatever reason.
There are a lot of misguided ways to assemble criteria for drawing a conclusion. I'm not sure what the frequency with which they occur has to do with making them any less fallible.