PatsFans.com - Mobile
PatsFans.com
Search

OT: Official 2020 Tompa Bay Gronkaneers Thread

2021 Patriots Season:
Upcoming Opponent:
Next Up: vs Jets
Pick Results: NYJ: 0% at NE: 0%

Sun
Oct 24th

Current Patriots Twitter Feed:
Status
Not open for further replies.

Ice_Ice_Brady

Disputed Winner - Week 2 Predict the Score Contest
PatsFans.com Supporter
@Bleedthrough
@Deus Irae
@Tony2046

Here's an update on my QB rank project that I mentioned earlier. I saw some half-assed thing on Profootball-reference that was supposed to be a Hall of Fame monitor, and it was really bad, inconsistent, and made no sense; and not only that, I started reading the comments and saw it was the same morons from 2003 still pimping out Manning > Brady, so I decided to make my own.

The rankings are based entirely on input data with a few small exceptions of giving a few prorated seasons (see notes.) And I think the weight I've applied to the data is accurate as a HoF monitor because as you can see, the list has actual Hall of Famers in gold and has a very clear cutoff at #38. The first 37 are all either in the Hall of Fame or not eligible yet but likely. Tony Romo is the first guy to miss the cut (and makes me think people may be surprised when a debate begins about him...because it probably will in a few years). After that, if you look down the list, there's only a few scattered Hall of Famers who are in for special reasons (Moon basically being given full credit for all those CFL Championships, Namath's guarantee and NY media bias, some old timers.)

I have an adjuster on the left, which allowed/allows me to fine tune it. So, there are three major indexes: the Windex, the Trophydex, and the Ratedex. How much you want each to make an impact will change the rankings a bit to favor winning championships, accolades, passer rating, etc. But the thing is that most of these Hall of Famers have all of these things so it often just changes a few slots but nothing major. The way I have it now, I think it's pretty balanced. I worked on it this week and found some things didn't look right, some guys too high and others too low. I found a pretty weird solution to the problem, which was to bring Dan Marino (who was at the time ranked near #25) and Terry Bradshaw (who was ranked near #8) together as a compromise in the middle at 16 and 17. I think most people would have these two close to each other in their rankings, but they have pretty opposite resumes. And in fact both are pretty frustrating to rank because you're looking for objective stats/accolades to put Marino up the list and objective ways to put Bradshaw down the list, but you don't want to open the can of worms of subjectively modifying, awarding statistical milestones, or turning to counting stats (shuddering at the thought), etc. You can only work with what you have or you'll ruin the entire ranking system. Tarkenton and Aikman also wound up next to each other at 24/25, a similar contrast pair.

So, here are the data points that I put in:
  • WINDEX: League Championships, Championships Appearances, Playoff Wins & Franchise (a formula using winning percentage). The first three are self-explanatory, but I should mention that Playoff Wins are ideal because they bridge pre-merger and post-merger. They give the modern QBs more opportunity to rack up points because it's harder to win it all. Franchise is a stat I created which is based on winning percentage and years played; it gives some points just for playing (winning at .60 like most QBs) but really rewards guys for long-term dominance.
  • TROPHYDEX: MVP, All-Pro, All-Star, All-NFL (that's all-decade or NFL100 team.) These are self-explanatory and just weighted. All of the rankings here are based on the idea that they're connected, so an MVP isn't worth more than an All-Pro because if you win an MVP, you're also winning an All-Pro and and All-Star, so all together that's a lot of points (same with championships, appearances, and playoffs.)
  • RATEDEX: I used a simple passer rating for every player on this one. There are definiteliy better measures of QB skills, but this isn't looking to be precise but just to get a general level of play. It's important to note, for example, a major difference in passer rating between two guys playing in the same era (Young/Aikman, Staubach/Bradshaw.) I created a formula which does two things with passer rating...first, it assigns points for pure passer rating, in a vaccuum, so basically you're just saying how is this guy compared to the most average QB in NFL history playing in 1963 with a passer rating of 72.3. That leads to recency bias, so the other half of it, the heavier weighted half, is to compare passer rating to the average of that era. It was pretty easy to calculate this (on average, passer rating goes up about 0.5 per year). Both of these categories are adjustable for fine tuning. But please note that by adjustments, I mean must apply to all, so there's no selective changes. After tuning it, I I was satisfied that Eli Manning was no longer a top 30 QB, the Hall of Fame index was aligned, Rodgers and Brees were neck and neck, Mahomes was 26 (seems right), and of course, Montana was ahead of Manning. After all this seemed aligned, it was really interesting to look at the sub-Hall of Fame guys and how the formula works to separate them out. I'm not sure it's right or not. That's when you really notice how finely tuned it is; the Hall of Fame guys are easier to rank because they have such big resumes.
Anyway, let me know what you guys think and glad to get suggestions/input. I'll create a Googe sheet if you're interested.

1614505602822.png
1614505729483.png
1614508198356.png


Oh...one last thing that is really freaking cool...the rivalries between the guys at the top of the list who are all close to each other:

Brady-Manning (well, not close anymore, but you know)

Unitas-Starr (Starr won most of the head-to-heads; no doubt that changed the rankings)

Luckman-Baugh (this was another great rivalry; Luckman was also a defensive back, do-everything athlete credited for big innovations in the passing game...Baugh gets most of the legacy talk, but Luckman wins objectively in almost all categories; that rivalry included the 73-0 game btw)

Brees-Rodgers (kind of a sad rivalry...lol)...and of course, Favre-Rodgers.

Favre-Young

Bradshaw-Staubach


All in all, that's almost everyone in the top 15...pretty crazy.
 
Last edited:

Zarozzor

In the Starting Line-Up
@Bleedthrough
@Deus Irae
@Tony2046

Here's an update on my QB rank project that I mentioned earlier. I saw some moronic thing on Profootball-reference that was supposed to be a Hall of Fame monitor, and it was really bad, inconsistent, and made no sense; and not only that, I started reading the comments and saw it was the same morons from 2003 still pimping out Manning > Brady, so I decided to make my own.

The rankings are based entirely on input data with a few small exceptions of giving a few prorated seasons (see notes.) And I think the weight I've applied to the data is accurate as a HofF monitor because as you can see, the actual Hall of Fame has a very clear cutoff at #37. The first 37 are all either in the Hall of Fame or not eligible but likely. Tony Romo is the first guy to miss the cut (and makes me think people may be surprised when a debate begins about him...because it probably will in a few years). After that, if you look down the list, there's only a few scatered Hall of Famers who are in for special reasons (Moon basically being given full credit for all those CFL Championships, Namath's guarantee and NY media bias, some old timers.)

I have an adjuster on the left, which allowed/allows me to fine tune it. So, there are three major indexes: the Windex, the Trophydex, and the Ratedex. How much you want each to make an impact will change the rankings a bit to favor winning championships, accolades, passer rating, etc. But the thing is that most of these Hall of Famers have all of these things so it often just changes a few slots but nothing major. The way I have it now, I think it's pretty balanced. I worked on it this week and found some things didn't look right, some guys too high and others too low. I found a pretty weird solution to the problem, which was to bring Dan Marino (who was at the time ranked near #25) and Terry Bradshaw (who was ranked near #8) together as a compromise in the middle at 16 and 17. I think most people would have these two close to each other in their rankings, but they have pretty opposite resumes. And in fact both are pretty frustrating to rank because you're looking for objective stats/accolades to put Marino up the list and objective ways to put Bradshaw down the list, but you don't want to open the can of worms of subjective rankings, counting stats (shuddering at the thought), etc. You can only work with what you have or you'll ruin the entire ranking system.

So, here are the data points that I put in:
  • WINDEX: League Championships, Championships Appearances, Playoff Wins & Franchise (a formula using winning percentage). The first three are self-explanatory, but I should mention that Playoff Wins are ideal because they bridge pre-merger and post-merger. They give the modern QBs more opportunity to rack up points because it's harder to win it all. Franchise is a stat I created which is based on winning percentage and years played; it gives some points just for playing (winning at .60 like most QBs) but really rewards guys for long-term dominance.
  • TROPHYDEX: MVP, All-Pro, All-Star, All-NFL (that's all-decade or NFL100 team.) These are self-explanatory and just weighted. All of the rankings here are based on the idea that they're connected, so an MVP isn't worth more than an All-Pro because if you win an MVP, you're also winning an All-Pro and and All-Star, so all together that's a lot of points (same with championships, appearances, and playoffs.)
  • RATEDEX: I used a simple passer rating for every player on this one. There are definiteliy better measures of QB skills, but this isn't looking to be precise but just to get a general level of play. It's important to note, for example, a major difference in passer rating between two guys playing in the same era (Young/Aikman, Staubach/Bradshaw.) I created a formula which does two things with passer rating...first, it assigns points for pure passer rating, in a vaccuum, so basically you're just saying how is this guy compared to the most average QB in NFL history playing in 1963 with a passer rating of 72.3. That leads to recency bias, so the other half of it, the heavier weighted half, is to compare passer rating to the average of that era. It was pretty easy to calculate this (on average, passer rating goes up about 0.5 per year). Both of these categories are adjustable for fine tuning. But please note that by adjustments, I mean must apply to all, so there's no selective changes. After tuning it, I I was satisfied that Eli Manning was no longer a top 30 QB, the Hall of Fame index was aligned, Rodgers and Brees were neck and neck, Mahomes was 26 (seems right), and of course, Montana was ahead of Manning. After all this seemed aligned, it was really interesting to look at the sub-Hall of Fame guys and how the formula works to separate them out. I'm not sure it's right or not. That's when you really notice how finely tuned it is; the Hall of Fame guys are easier to rank because they have such big resumes.
Anyway, let me know what you guys think and glad to get suggestions/input. I'll create a Googe sheet if you're interested.

View attachment 30801
View attachment 30803
View attachment 30804
Amazing work. The top 10 list I came up with last year had your top 4 in the exact order. I had Manning, Starr, Unitas as my 4-6 and struggled with ranking them, but went with Manning at 4. I still flip Starr and Unitas between 5 and 6.

I also agree with the order of Rodgers, Favre, Brees.

I like Marino at 16 actually. I usually have him between 13-16. I personally think he's overrated, because people think he played at his 1984-1986 level most of his career. And say what you want about how bad his teams were, but I can't get past 0 Super Bowl wins and an 8-10 playoff record.
 

Ice_Ice_Brady

Disputed Winner - Week 2 Predict the Score Contest
PatsFans.com Supporter
Amazing work. The top 10 list I came up with last year had your top 4 in the exact order. I had Manning, Starr, Unitas as my 4-6 and struggled with ranking them, but went with Manning at 4. I still flip Starr and Unitas between 5 and 6.

I also agree with the order of Rodgers, Favre, Brees.

I like Marino at 16 actually. I usually have him between 13-16. I personally think he's overrated, because people think he played at his 1984-1986 level most of his career. And say what you want about how bad his teams were, but I can't get past 0 Super Bowl wins and an 8-10 playoff record.

Same here! I don't understand the obsession with Marino. I was a fan as a kid growing up and watching him, but I think because he didn't win in the postseason, people try to overcompensate for him. His season-to-season stats really aren't outliers at all. After that 1984 season, he's consistently very good but not not worth breaking your brain over creating an alternative category to capture his greatness.

I think I was trying to move him up mostly because I know how many people would see him at #24 and just dimiss the list due to that. The thing about him is there really isn't that much on paper to support his greatness, besides the eight all-pro selections. And I'm not saying that eight all-pro selections aren't impressive...it's just that you put a ton of weight on those and now Roger Staubach, who someone got screwed with just one, and a bunch of other guys who didn't get them, get pulled in the other direction. The Dan Marino/Terry Bradshaw problem is like one of those blackboard physics equations..lol

Yep...Starr and Unitas are interesting too. They are really close all the time and great for different reasons...it's like if Manning had been pretty good in the playoffs on his way to those SB wins and Brady had retired after the Seattle Super Bowl. I know a lot of people would probably get angry at seeing Unitas > Starr and saying that invalidates the rankings, but whatever...the two are within a point of each other whatever you do, so I think the overreaction there is similar to the Marino situation...you have people who just believe their perception no matter what, even if the numbers tell a different story. I like Unitas too...I would definitely be trying like hell to rank him over Starr if there were such a strong case, but it's a lot closer to a toss up. After early success, Unitas struggled a lot in big games for much of his career; Starr was money.
 
Last edited:

Zarozzor

In the Starting Line-Up
Same here! I don't understand the obsession with Marino. I was a fan as a kid growing up and watching him, but I think because he didn't win in the postseason, people try to overcompensate for him. His season-to-season stats really aren't outliers at all. After that 1984 season, he's consistently very good but not not worth breaking your brain over creating an alternative category to capture his greatness.

I think I was trying to move him up mostly because I know how many people would see him at #24 and just dimiss the list due to that. The thing about him is there really isn't that much on paper to support his greatness, besides the eight all-pro selections. And I'm not saying that eight all-pro selections aren't impressive...it's just that you put a ton of weight on those and now Roger Staubach, who someone got screwed with just one, and a bunch of other guys who didn't get them, get pulled in the other direction. The Dan Marino/Terry Bradshaw problem is like one of those blackboard physics equations..lol

Yep...Starr and Unitas are interesting too. They are really close all the time and great for different reasons...it's like if Manning had been pretty good in the playoffs on his way to those SB wins and Brady had retired after the Seattle Super Bowl. I know a lot of people would probably get angry at seeing Unitas > Starr and saying that invalidates the rankings, but whatever...the two are within a point of each other whatever you do, so I think the overreaction there is similar to the Marino situation...you have people who just believe their perception no matter what, even if the numbers tell a different story. I like Unitas too...I would definitely be trying like hell to rank him over Starr if there were such a strong case, but it's a lot closer to a toss up. After early success, Unitas struggled a lot in big games for much of his career; Starr was money.
Part of me really wants to keep Starr ahead of Unitas for good. He won 9 straight playoff games, something only Brady has passed. He was money in big time moments and a TD/INT ratio of 15-3 in the playoffs is crazy.

You have to choose your discussions wisely with this topic. I’ve ventured into the NFL Reddit and Facebook comments and some of the list I’ve seen, wow. A good majority of them have Marino, Elway, Rodgers, and Brees near the top. I’ve even seen some that have Mahomes top 10. That’s not even worth wasting the time arguing.

I have never ranked guys higher who have played more recently like most people seem to do. I don’t care when you played, I care how you dominated your own era and how good you were compared to your peers. It’s the reason I’m such a huge fan of Bill Russell and have Ty Cobb and Honus Wagner so high on my baseball list. You can’t control when you play, but you can control how well you played in your own era.
 
Last edited:

RIpats88

In the Starting Line-Up
Bucs will make it work to pay their guys, moving salary into the future. Brady is 43 and they need to maximize the time they have with him to try and win another one. They may hit a wall when the bill comes due like the Saints have this year in the future, but at that point who cares if you have won a couple super bowls?
 

PatsWSB47

Pro Bowl Player
Bucs will make it work to pay their guys, moving salary into the future. Brady is 43 and they need to maximize the time they have with him to try and win another one. They may hit a wall when the bill comes due like the Saints have this year in the future, but at that point who cares if you have won a couple super bowls?
Exactly and so if they hit a wall they'll revert back to the same no playoff team that they've were 2002 to 2020.
 

venecol

The FRG has a little ****
PatsFans.com Supporter
2021 Weekly NFL Picks Winner
@Bleedthrough
@Deus Irae
@Tony2046

Here's an update on my QB rank project that I mentioned earlier. I saw some half-assed thing on Profootball-reference that was supposed to be a Hall of Fame monitor, and it was really bad, inconsistent, and made no sense; and not only that, I started reading the comments and saw it was the same morons from 2003 still pimping out Manning > Brady, so I decided to make my own.

The rankings are based entirely on input data with a few small exceptions of giving a few prorated seasons (see notes.) And I think the weight I've applied to the data is accurate as a HoF monitor because as you can see, the list has actual Hall of Famers in gold and has a very clear cutoff at #38. The first 37 are all either in the Hall of Fame or not eligible yet but likely. Tony Romo is the first guy to miss the cut (and makes me think people may be surprised when a debate begins about him...because it probably will in a few years). After that, if you look down the list, there's only a few scattered Hall of Famers who are in for special reasons (Moon basically being given full credit for all those CFL Championships, Namath's guarantee and NY media bias, some old timers.)

I have an adjuster on the left, which allowed/allows me to fine tune it. So, there are three major indexes: the Windex, the Trophydex, and the Ratedex. How much you want each to make an impact will change the rankings a bit to favor winning championships, accolades, passer rating, etc. But the thing is that most of these Hall of Famers have all of these things so it often just changes a few slots but nothing major. The way I have it now, I think it's pretty balanced. I worked on it this week and found some things didn't look right, some guys too high and others too low. I found a pretty weird solution to the problem, which was to bring Dan Marino (who was at the time ranked near #25) and Terry Bradshaw (who was ranked near #8) together as a compromise in the middle at 16 and 17. I think most people would have these two close to each other in their rankings, but they have pretty opposite resumes. And in fact both are pretty frustrating to rank because you're looking for objective stats/accolades to put Marino up the list and objective ways to put Bradshaw down the list, but you don't want to open the can of worms of subjectively modifying, awarding statistical milestones, or turning to counting stats (shuddering at the thought), etc. You can only work with what you have or you'll ruin the entire ranking system. Tarkenton and Aikman also wound up next to each other at 24/25, a similar contrast pair.

So, here are the data points that I put in:
  • WINDEX: League Championships, Championships Appearances, Playoff Wins & Franchise (a formula using winning percentage). The first three are self-explanatory, but I should mention that Playoff Wins are ideal because they bridge pre-merger and post-merger. They give the modern QBs more opportunity to rack up points because it's harder to win it all. Franchise is a stat I created which is based on winning percentage and years played; it gives some points just for playing (winning at .60 like most QBs) but really rewards guys for long-term dominance.
  • TROPHYDEX: MVP, All-Pro, All-Star, All-NFL (that's all-decade or NFL100 team.) These are self-explanatory and just weighted. All of the rankings here are based on the idea that they're connected, so an MVP isn't worth more than an All-Pro because if you win an MVP, you're also winning an All-Pro and and All-Star, so all together that's a lot of points (same with championships, appearances, and playoffs.)
  • RATEDEX: I used a simple passer rating for every player on this one. There are definiteliy better measures of QB skills, but this isn't looking to be precise but just to get a general level of play. It's important to note, for example, a major difference in passer rating between two guys playing in the same era (Young/Aikman, Staubach/Bradshaw.) I created a formula which does two things with passer rating...first, it assigns points for pure passer rating, in a vaccuum, so basically you're just saying how is this guy compared to the most average QB in NFL history playing in 1963 with a passer rating of 72.3. That leads to recency bias, so the other half of it, the heavier weighted half, is to compare passer rating to the average of that era. It was pretty easy to calculate this (on average, passer rating goes up about 0.5 per year). Both of these categories are adjustable for fine tuning. But please note that by adjustments, I mean must apply to all, so there's no selective changes. After tuning it, I I was satisfied that Eli Manning was no longer a top 30 QB, the Hall of Fame index was aligned, Rodgers and Brees were neck and neck, Mahomes was 26 (seems right), and of course, Montana was ahead of Manning. After all this seemed aligned, it was really interesting to look at the sub-Hall of Fame guys and how the formula works to separate them out. I'm not sure it's right or not. That's when you really notice how finely tuned it is; the Hall of Fame guys are easier to rank because they have such big resumes.
Anyway, let me know what you guys think and glad to get suggestions/input. I'll create a Googe sheet if you're interested.

View attachment 30801
View attachment 30803
View attachment 30804


Oh...one last thing that is really freaking cool...the rivalries between the guys at the top of the list who are all close to each other:

Brady-Manning (well, not close anymore, but you know)

Unitas-Starr (Starr won most of the head-to-heads; no doubt that changed the rankings)

Luckman-Baugh (this was another great rivalry; Luckman was also a defensive back, do-everything athlete credited for big innovations in the passing game...Baugh gets most of the legacy talk, but Luckman wins objectively in almost all categories; that rivalry included the 73-0 game btw)

Brees-Rodgers (kind of a sad rivalry...lol)...and of course, Favre-Rodgers.

Favre-Young

Bradshaw-Staubach


All in all, that's almost everyone in the top 15...pretty crazy.
Wow, that's a lot of work. How'd you get away with not mowing the lawn for that long?

Is it possible to break out Brady's career into 3 different periods: 2000-2006, 2007-2013, 2014-2020, as if he was 3 different QBs (Tom1, Tom2, Tom3) and see where they fall in the chart?
 

Deus Irae

PatsFans.com Retired Jersey Club
PatsFans.com Supporter
@Bleedthrough
@Deus Irae
@Tony2046

Here's an update on my QB rank project that I mentioned earlier. I saw some half-assed thing on Profootball-reference that was supposed to be a Hall of Fame monitor, and it was really bad, inconsistent, and made no sense; and not only that, I started reading the comments and saw it was the same morons from 2003 still pimping out Manning > Brady, so I decided to make my own.

The rankings are based entirely on input data with a few small exceptions of giving a few prorated seasons (see notes.) And I think the weight I've applied to the data is accurate as a HoF monitor because as you can see, the list has actual Hall of Famers in gold and has a very clear cutoff at #38. The first 37 are all either in the Hall of Fame or not eligible yet but likely. Tony Romo is the first guy to miss the cut (and makes me think people may be surprised when a debate begins about him...because it probably will in a few years). After that, if you look down the list, there's only a few scattered Hall of Famers who are in for special reasons (Moon basically being given full credit for all those CFL Championships, Namath's guarantee and NY media bias, some old timers.)

I have an adjuster on the left, which allowed/allows me to fine tune it. So, there are three major indexes: the Windex, the Trophydex, and the Ratedex. How much you want each to make an impact will change the rankings a bit to favor winning championships, accolades, passer rating, etc. But the thing is that most of these Hall of Famers have all of these things so it often just changes a few slots but nothing major. The way I have it now, I think it's pretty balanced. I worked on it this week and found some things didn't look right, some guys too high and others too low. I found a pretty weird solution to the problem, which was to bring Dan Marino (who was at the time ranked near #25) and Terry Bradshaw (who was ranked near #8) together as a compromise in the middle at 16 and 17. I think most people would have these two close to each other in their rankings, but they have pretty opposite resumes. And in fact both are pretty frustrating to rank because you're looking for objective stats/accolades to put Marino up the list and objective ways to put Bradshaw down the list, but you don't want to open the can of worms of subjectively modifying, awarding statistical milestones, or turning to counting stats (shuddering at the thought), etc. You can only work with what you have or you'll ruin the entire ranking system. Tarkenton and Aikman also wound up next to each other at 24/25, a similar contrast pair.

So, here are the data points that I put in:
  • WINDEX: League Championships, Championships Appearances, Playoff Wins & Franchise (a formula using winning percentage). The first three are self-explanatory, but I should mention that Playoff Wins are ideal because they bridge pre-merger and post-merger. They give the modern QBs more opportunity to rack up points because it's harder to win it all. Franchise is a stat I created which is based on winning percentage and years played; it gives some points just for playing (winning at .60 like most QBs) but really rewards guys for long-term dominance.
  • TROPHYDEX: MVP, All-Pro, All-Star, All-NFL (that's all-decade or NFL100 team.) These are self-explanatory and just weighted. All of the rankings here are based on the idea that they're connected, so an MVP isn't worth more than an All-Pro because if you win an MVP, you're also winning an All-Pro and and All-Star, so all together that's a lot of points (same with championships, appearances, and playoffs.)
  • RATEDEX: I used a simple passer rating for every player on this one. There are definiteliy better measures of QB skills, but this isn't looking to be precise but just to get a general level of play. It's important to note, for example, a major difference in passer rating between two guys playing in the same era (Young/Aikman, Staubach/Bradshaw.) I created a formula which does two things with passer rating...first, it assigns points for pure passer rating, in a vaccuum, so basically you're just saying how is this guy compared to the most average QB in NFL history playing in 1963 with a passer rating of 72.3. That leads to recency bias, so the other half of it, the heavier weighted half, is to compare passer rating to the average of that era. It was pretty easy to calculate this (on average, passer rating goes up about 0.5 per year). Both of these categories are adjustable for fine tuning. But please note that by adjustments, I mean must apply to all, so there's no selective changes. After tuning it, I I was satisfied that Eli Manning was no longer a top 30 QB, the Hall of Fame index was aligned, Rodgers and Brees were neck and neck, Mahomes was 26 (seems right), and of course, Montana was ahead of Manning. After all this seemed aligned, it was really interesting to look at the sub-Hall of Fame guys and how the formula works to separate them out. I'm not sure it's right or not. That's when you really notice how finely tuned it is; the Hall of Fame guys are easier to rank because they have such big resumes.
Anyway, let me know what you guys think and glad to get suggestions/input. I'll create a Googe sheet if you're interested.

View attachment 30801
View attachment 30803
View attachment 30804


Oh...one last thing that is really freaking cool...the rivalries between the guys at the top of the list who are all close to each other:

Brady-Manning (well, not close anymore, but you know)

Unitas-Starr (Starr won most of the head-to-heads; no doubt that changed the rankings)

Luckman-Baugh (this was another great rivalry; Luckman was also a defensive back, do-everything athlete credited for big innovations in the passing game...Baugh gets most of the legacy talk, but Luckman wins objectively in almost all categories; that rivalry included the 73-0 game btw)

Brees-Rodgers (kind of a sad rivalry...lol)...and of course, Favre-Rodgers.

Favre-Young

Bradshaw-Staubach


All in all, that's almost everyone in the top 15...pretty crazy.
I think it's an interesting start, and I love the effort, but it's got some failures that are so obvious (IMO) that it undermines the totality, and leaves me thinking about the need for some adjustment(s):


Peyton above Unitas, Starr, AND Baugh?
Staubach below Young?
Romo ahead of Moon?
Mahomes already at 26?
Rodgers, Favre, and Brees above Elway and Marino (among others)?
etc...


I don't know how you'd fix those things, but they seem fatal to the process at this point. Maybe it's a recency bias built into the choices. That seems to be where the trend is leading.



Edit: Wow! I'd missed Wilson over Griese, Aikman, and Tarkenton, and Rivers over Stabler. Yeah, it's got to be an issue with some kind of recency bias.
 
Last edited:

Ice_Ice_Brady

Disputed Winner - Week 2 Predict the Score Contest
PatsFans.com Supporter
I think it's an interesting start, and I love the effort, but it's got some failures that are so obvious (IMO) that it undermines the totality, and leaves me thinking about the need for some adjustment(s):


Peyton above Unitas, Starr, AND Baugh?
Staubach below Young?
Romo ahead of Moon?
Mahomes already at 26?
Rodgers, Favre, and Brees above Elway and Marino (among others)?
etc...


I don't know how you'd fix those things, but they seem fatal to the process at this point. Maybe it's a recency bias built into the choices. That seems to be where the trend is leading.



Edit: Wow! I'd missed Wilson over Griese, Aikman, and Tarkenton, and Rivers over Stabler. Yeah, it's got to be an issue with some kind of recency bias.

Thanks for the feedback...I'm going to take a look at some of these issue a lot deeper. Will post an update later on. with more, but for now...I don't disagree with you on some of these rankings, and I'm looking for ways to adjust them, but it's all numerical, so change one thing and everyone shifts, which is why my Excel formulas and the data points keep growing. For example, I do have Staubach above Young now (below), based on my last adjustment...but that adjustment also put Marino below Bradshaw. I don't think it's possible to get every position to match exactly how you'd rank it subjectively, using this type of algorithm-based process.

The most difficult players to rank are: Marino, Bradshaw, Elway, Aikman, and Moon.

The problem with Peyton is that he doesn't have a real weakness on paper. You could very heavily weight championships, but if you do that you run into other issues because now how do you, for example, put Staubach over Bradshaw? And if you weigh those titles, Eli creeps into a level where he should't be. I mean, we know intuitively both of those are true (Peyton overrated, and yet some multi-winning champions are overrated too.)

I think Wilson over Griese/Aikman/Tarkenton makes sense though. or at least is arguable Wilson has nine seasons, which is almost as much as Aikman or Griese; Wilson really has championships/postseason/winning percentage and statistical prowess. Aikman has the three championshps, but his career just isn't that great overall; Griese basically has everything but watered down versions of them; and Tarkenton is lacking the winning. Wilson's passer rating is 101 and Aikman's is 81; now, I understand these are different eras and have made the system compare them with era adjustments, and it's still a huge gulf. So, that's why Aikman is going to lose a lot of ground; at the same time, that adjustment is why Young and Staubach make the top 10 while some other guys don't. The formula identifies Young and Staubach as statistical outliers and weighs that heavily because if it's based on just championships, playoff wins, and awards, neither would stand out enough.

That's also why Mahomes jumps up the rankings...due to his huge outlier passer rating. I did make an adjustment to limit the amount his passer rating would count, so not to let him jump into the top 10 or something crazy, and I reduced the amount of impact of his rating because he has had such a short career. But overall, his standing in the rankings is based on (a) that incredible passer rating, even if I flattened it, which is impressive against his era, not just the number itself, (b) he already has a championship, two appearances, six playoff wins, two MVPs (reg and SB), and he's winning at 81%, which has also been flattened to account for his lack of longevity but still is a big deal.

Not trying to be defensive, as I'm not married to the order of the rankings but just trying to explain the process/alegebra behind it....


1614548580449.png
 
Last edited:

Ice_Ice_Brady

Disputed Winner - Week 2 Predict the Score Contest
PatsFans.com Supporter
Wow, that's a lot of work. How'd you get away with not mowing the lawn for that long?

Is it possible to break out Brady's career into 3 different periods: 2000-2006, 2007-2013, 2014-2020, as if he was 3 different QBs (Tom1, Tom2, Tom3) and see where they fall in the chart?

I will post that...good idea.
 

Deus Irae

PatsFans.com Retired Jersey Club
PatsFans.com Supporter
Thanks for the feedback...I'm going to take a look at some of these issue a lot deeper. Will post an update later on. with more, but for now...I don't disagree with you on some of these rankings, and I'm looking for ways to adjust them, but it's all numerical, so change one thing and everyone shifts, which is why my Excel formulas and the data points keep growing. For example, I do have Staubach above Young now (below), based on my last adjustment...but that adjustment also put Marino below Bradshaw. I don't think it's possible to get every position to match exactly how you'd rank it subjectively, using this type of algorithm-based process.

The most difficult players to rank are: Marino, Bradshaw, Elway, Aikman, and Moon.

The problem with Peyton is that he doesn't have a real weakness on paper. You could very heavily weight championships, but if you do that you run into other issues because now how do you, for example, put Staubach over Bradshaw? And if you weigh those titles, Eli creeps into a level where he should't be. I mean, we know intuitively both of those are true (Peyton overrated, and yet some multi-winning champions are overrated too.)

I think Wilson over Griese/Aikman/Tarkenton makes sense though. or at least is arguable Wilson has nine seasons, which is almost as much as Aikman or Griese; Wilson really has championships/postseason/winning percentage and statistical prowess. Aikman has the three championshps, but his career just isn't that great overall; Griese basically has everything but watered down versions of them; and Tarkenton is lacking the winning. Wilson's passer rating is 101 and Aikman's is 81; now, I understand these are different eras and have made the system compare them with era adjustments, and it's still a huge gulf. So, that's why Aikman is going to lose a lot of ground; at the same time, that adjustment is why Young and Staubach make the top 10 while some other guys don't. The formula identifies Young and Staubach as statistical outliers and weighs that heavily because if it's based on just championships, playoff wins, and awards, neither would stand out enough.

That's also why Mahomes jumps up the rankings...due to his huge outlier passer rating. I did make an adjustment to limit the amount his passer rating would count, so not to let him jump into the top 10 or something crazy, and I reduced the amount of impact of his rating because he has had such a short career. But overall, his standing in the rankings is based on (a) that incredible passer rating, even if I flattened it, which is impressive against his era, not just the number itself, (b) he already has a championship, two appearances, six playoff wins, two MVPs (reg and SB), and he's winning at 81%, which has also been flattened to account for his lack of longevity but still is a big deal.

Not trying to be defensive, as I'm not married to the order of the rankings but just trying to explain the process/alegebra behind it....


View attachment 30812

No worries. You're essentially trying to do the impossible, which is to find a determinative QB ranking that works for everyone, when some aspects of QB play can't be measured , and when it's never going to happen that all people will be in line with all slottings. I do see what appears to be a recency bias, though, and that's not something that's at all surprising. The way stats, the rules, the coaching, and so many other things have changed make it a Herculean task to find the perfect formula to balance all of that. I mean, how does one find the exact and perfect way to convert QB skills of 1933 to QB skills of 2020, right?

As for slotting Manning, I absolutely understand that he has some nearly unique factors when it comes to ranking him. He got the breaks of playing under the most recent rules (which covered some of his weaknesses) specifically because his team whined about the old rules, being a much better regular season QB than postseason QB, getting gift MVP awards, playing in a dome with an offense specifically built to maximize both that home field advantage and the skillset of indoor Manning, and being a stat inflator in the mold of Rodgers and Brees.

But there's not a chance in hell you'll convince me about Wilson. :)
 

RobertWeathers

Author of the Port Huron Statement
PatsFans.com Supporter
@Bleedthrough
@Deus Irae
@Tony2046

Here's an update on my QB rank project that I mentioned earlier. I saw some half-assed thing on Profootball-reference that was supposed to be a Hall of Fame monitor, and it was really bad, inconsistent, and made no sense; and not only that, I started reading the comments and saw it was the same morons from 2003 still pimping out Manning > Brady, so I decided to make my own.

The rankings are based entirely on input data with a few small exceptions of giving a few prorated seasons (see notes.) And I think the weight I've applied to the data is accurate as a HoF monitor because as you can see, the list has actual Hall of Famers in gold and has a very clear cutoff at #38. The first 37 are all either in the Hall of Fame or not eligible yet but likely. Tony Romo is the first guy to miss the cut (and makes me think people may be surprised when a debate begins about him...because it probably will in a few years). After that, if you look down the list, there's only a few scattered Hall of Famers who are in for special reasons (Moon basically being given full credit for all those CFL Championships, Namath's guarantee and NY media bias, some old timers.)

I have an adjuster on the left, which allowed/allows me to fine tune it. So, there are three major indexes: the Windex, the Trophydex, and the Ratedex. How much you want each to make an impact will change the rankings a bit to favor winning championships, accolades, passer rating, etc. But the thing is that most of these Hall of Famers have all of these things so it often just changes a few slots but nothing major. The way I have it now, I think it's pretty balanced. I worked on it this week and found some things didn't look right, some guys too high and others too low. I found a pretty weird solution to the problem, which was to bring Dan Marino (who was at the time ranked near #25) and Terry Bradshaw (who was ranked near #8) together as a compromise in the middle at 16 and 17. I think most people would have these two close to each other in their rankings, but they have pretty opposite resumes. And in fact both are pretty frustrating to rank because you're looking for objective stats/accolades to put Marino up the list and objective ways to put Bradshaw down the list, but you don't want to open the can of worms of subjectively modifying, awarding statistical milestones, or turning to counting stats (shuddering at the thought), etc. You can only work with what you have or you'll ruin the entire ranking system. Tarkenton and Aikman also wound up next to each other at 24/25, a similar contrast pair.

So, here are the data points that I put in:
  • WINDEX: League Championships, Championships Appearances, Playoff Wins & Franchise (a formula using winning percentage). The first three are self-explanatory, but I should mention that Playoff Wins are ideal because they bridge pre-merger and post-merger. They give the modern QBs more opportunity to rack up points because it's harder to win it all. Franchise is a stat I created which is based on winning percentage and years played; it gives some points just for playing (winning at .60 like most QBs) but really rewards guys for long-term dominance.
  • TROPHYDEX: MVP, All-Pro, All-Star, All-NFL (that's all-decade or NFL100 team.) These are self-explanatory and just weighted. All of the rankings here are based on the idea that they're connected, so an MVP isn't worth more than an All-Pro because if you win an MVP, you're also winning an All-Pro and and All-Star, so all together that's a lot of points (same with championships, appearances, and playoffs.)
  • RATEDEX: I used a simple passer rating for every player on this one. There are definiteliy better measures of QB skills, but this isn't looking to be precise but just to get a general level of play. It's important to note, for example, a major difference in passer rating between two guys playing in the same era (Young/Aikman, Staubach/Bradshaw.) I created a formula which does two things with passer rating...first, it assigns points for pure passer rating, in a vaccuum, so basically you're just saying how is this guy compared to the most average QB in NFL history playing in 1963 with a passer rating of 72.3. That leads to recency bias, so the other half of it, the heavier weighted half, is to compare passer rating to the average of that era. It was pretty easy to calculate this (on average, passer rating goes up about 0.5 per year). Both of these categories are adjustable for fine tuning. But please note that by adjustments, I mean must apply to all, so there's no selective changes. After tuning it, I I was satisfied that Eli Manning was no longer a top 30 QB, the Hall of Fame index was aligned, Rodgers and Brees were neck and neck, Mahomes was 26 (seems right), and of course, Montana was ahead of Manning. After all this seemed aligned, it was really interesting to look at the sub-Hall of Fame guys and how the formula works to separate them out. I'm not sure it's right or not. That's when you really notice how finely tuned it is; the Hall of Fame guys are easier to rank because they have such big resumes.
Anyway, let me know what you guys think and glad to get suggestions/input. I'll create a Googe sheet if you're interested.

View attachment 30801
View attachment 30803
View attachment 30804


Oh...one last thing that is really freaking cool...the rivalries between the guys at the top of the list who are all close to each other:

Brady-Manning (well, not close anymore, but you know)

Unitas-Starr (Starr won most of the head-to-heads; no doubt that changed the rankings)

Luckman-Baugh (this was another great rivalry; Luckman was also a defensive back, do-everything athlete credited for big innovations in the passing game...Baugh gets most of the legacy talk, but Luckman wins objectively in almost all categories; that rivalry included the 73-0 game btw)

Brees-Rodgers (kind of a sad rivalry...lol)...and of course, Favre-Rodgers.

Favre-Young

Bradshaw-Staubach


All in all, that's almost everyone in the top 15...pretty crazy.
Well done.

I have some minor arguments but so what. You've done a great job quantifying winning which is the #1 responsibility of a QB.
 

Ice_Ice_Brady

Disputed Winner - Week 2 Predict the Score Contest
PatsFans.com Supporter
Well done.

I have some minor arguments but so what. You've done a great job quantifying winning which is the #1 responsibility of a QB.

Thanks! Let's hear them....then I can adjust for it and post what the rest of the list looks like based on the change in criteria.
 

Tony2046

PatsFans.com Supporter
PatsFans.com Supporter
2019 Weekly Picks Winner
2021 Weekly NFL Picks Winner
@Bleedthrough
@Deus Irae
@Tony2046

Here's an update on my QB rank project that I mentioned earlier. I saw some half-assed thing on Profootball-reference that was supposed to be a Hall of Fame monitor, and it was really bad, inconsistent, and made no sense; and not only that, I started reading the comments and saw it was the same morons from 2003 still pimping out Manning > Brady, so I decided to make my own.

The rankings are based entirely on input data with a few small exceptions of giving a few prorated seasons (see notes.) And I think the weight I've applied to the data is accurate as a HoF monitor because as you can see, the list has actual Hall of Famers in gold and has a very clear cutoff at #38. The first 37 are all either in the Hall of Fame or not eligible yet but likely. Tony Romo is the first guy to miss the cut (and makes me think people may be surprised when a debate begins about him...because it probably will in a few years). After that, if you look down the list, there's only a few scattered Hall of Famers who are in for special reasons (Moon basically being given full credit for all those CFL Championships, Namath's guarantee and NY media bias, some old timers.)

I have an adjuster on the left, which allowed/allows me to fine tune it. So, there are three major indexes: the Windex, the Trophydex, and the Ratedex. How much you want each to make an impact will change the rankings a bit to favor winning championships, accolades, passer rating, etc. But the thing is that most of these Hall of Famers have all of these things so it often just changes a few slots but nothing major. The way I have it now, I think it's pretty balanced. I worked on it this week and found some things didn't look right, some guys too high and others too low. I found a pretty weird solution to the problem, which was to bring Dan Marino (who was at the time ranked near #25) and Terry Bradshaw (who was ranked near #8) together as a compromise in the middle at 16 and 17. I think most people would have these two close to each other in their rankings, but they have pretty opposite resumes. And in fact both are pretty frustrating to rank because you're looking for objective stats/accolades to put Marino up the list and objective ways to put Bradshaw down the list, but you don't want to open the can of worms of subjectively modifying, awarding statistical milestones, or turning to counting stats (shuddering at the thought), etc. You can only work with what you have or you'll ruin the entire ranking system. Tarkenton and Aikman also wound up next to each other at 24/25, a similar contrast pair.

So, here are the data points that I put in:
  • WINDEX: League Championships, Championships Appearances, Playoff Wins & Franchise (a formula using winning percentage). The first three are self-explanatory, but I should mention that Playoff Wins are ideal because they bridge pre-merger and post-merger. They give the modern QBs more opportunity to rack up points because it's harder to win it all. Franchise is a stat I created which is based on winning percentage and years played; it gives some points just for playing (winning at .60 like most QBs) but really rewards guys for long-term dominance.
  • TROPHYDEX: MVP, All-Pro, All-Star, All-NFL (that's all-decade or NFL100 team.) These are self-explanatory and just weighted. All of the rankings here are based on the idea that they're connected, so an MVP isn't worth more than an All-Pro because if you win an MVP, you're also winning an All-Pro and and All-Star, so all together that's a lot of points (same with championships, appearances, and playoffs.)
  • RATEDEX: I used a simple passer rating for every player on this one. There are definiteliy better measures of QB skills, but this isn't looking to be precise but just to get a general level of play. It's important to note, for example, a major difference in passer rating between two guys playing in the same era (Young/Aikman, Staubach/Bradshaw.) I created a formula which does two things with passer rating...first, it assigns points for pure passer rating, in a vaccuum, so basically you're just saying how is this guy compared to the most average QB in NFL history playing in 1963 with a passer rating of 72.3. That leads to recency bias, so the other half of it, the heavier weighted half, is to compare passer rating to the average of that era. It was pretty easy to calculate this (on average, passer rating goes up about 0.5 per year). Both of these categories are adjustable for fine tuning. But please note that by adjustments, I mean must apply to all, so there's no selective changes. After tuning it, I I was satisfied that Eli Manning was no longer a top 30 QB, the Hall of Fame index was aligned, Rodgers and Brees were neck and neck, Mahomes was 26 (seems right), and of course, Montana was ahead of Manning. After all this seemed aligned, it was really interesting to look at the sub-Hall of Fame guys and how the formula works to separate them out. I'm not sure it's right or not. That's when you really notice how finely tuned it is; the Hall of Fame guys are easier to rank because they have such big resumes.
Anyway, let me know what you guys think and glad to get suggestions/input. I'll create a Googe sheet if you're interested.

View attachment 30801
View attachment 30803
View attachment 30804


Oh...one last thing that is really freaking cool...the rivalries between the guys at the top of the list who are all close to each other:

Brady-Manning (well, not close anymore, but you know)

Unitas-Starr (Starr won most of the head-to-heads; no doubt that changed the rankings)

Luckman-Baugh (this was another great rivalry; Luckman was also a defensive back, do-everything athlete credited for big innovations in the passing game...Baugh gets most of the legacy talk, but Luckman wins objectively in almost all categories; that rivalry included the 73-0 game btw)

Brees-Rodgers (kind of a sad rivalry...lol)...and of course, Favre-Rodgers.

Favre-Young

Bradshaw-Staubach


All in all, that's almost everyone in the top 15...pretty crazy.

This is really an incredible chart.

Where in the hell is Etling???
 

RobertWeathers

Author of the Port Huron Statement
PatsFans.com Supporter
Thanks! Let's hear them....then I can adjust for it and post what the rest of the list looks like based on the change in criteria.
Quantifying the greatness of NFL QBs is not an exact science but you are coming damn close.

Some thoughts.

Do they win? Reg season? Super Bowls? Playoff games? Record vs playoff teams?
Has their offense scored points?
Have they converted 3rd downs?
Have they managed the game (Qb rating is a good stat for game management but does not mean they win)?
Have their heroics (see GW drives opportunity success %) translated into victories?

My point is I believe there is a difference between the Tom Bradys, Otto Grahams and Sammy Baughs and Dan Marinos, Jim Kellys and Dan Fouts of the world.

I would take all of these stats and do a WAR/above league position average stat.
 

Ice_Ice_Brady

Disputed Winner - Week 2 Predict the Score Contest
PatsFans.com Supporter
This is really an incredible chart.

Where in the hell is Etling???

We can all be sure that he'll be in the top 20 when Ice Ice remembers to input him into the mix....

Pretty sure this is Etling territory...

1614562974726.png

It is good to see those loveable Jets make an appearance though.

This one is funny: nine seasons to confrim you're a net push.

1614563276678.png
 

Attachments

  • 1614562933180.png
    1614562933180.png
    45.6 KB · Views: 1
Last edited:

Ice_Ice_Brady

Disputed Winner - Week 2 Predict the Score Contest
PatsFans.com Supporter
Quantifying the greatness of NFL QBs is not an exact science but you are coming damn close.

Some thoughts.

Do they win? Reg season? Super Bowls? Playoff games? Record vs playoff teams?
Has their offense scored points?
Have they converted 3rd downs?
Have they managed the game (Qb rating is a good stat for game management but does not mean they win)?
Have their heroics (see GW drives opportunity success %) translated into victories?

My point is I believe there is a difference between the Tom Bradys, Otto Grahams and Sammy Baughs and Dan Marinos, Jim Kellys and Dan Fouts of the world.

I would take all of these stats and do a WAR/above league position average stat.

Good ideas here. I think passer rating can be very misleading too, but I think it's a good simple index to use for the right reasons. If you're looking to see if a guy is elite, it's usually pretty accurate. If you're looking to see if a guy sucks (maybe his team is carrying him), it usually tells you that too. Winning is by far the best measure of success; to me, the passer rating (or whatever stats being used) is more of way to separate within classes of players. We know that Russell Wilson is better than Joe Flacco; this gives us the statistical evidence. But a guy with a big passer rating by itself won't do so well in the eyes of history, and so yes, I think it's important to flatten out the impact when ranking QBs.

I think your last point about heroics is the one that makes it the biggest challenge because it's such a big deal and yet cannot be picked up in an algorithm. It's why, I think, a lot of us would have trouble ranking Manning ahead of Montana and even feeling dirty about ranking him in front of Starr; and same thing with Rodgers. These guys don't have the big moments to match their reputation and skill level.
 

RobertWeathers

Author of the Port Huron Statement
PatsFans.com Supporter
I think your last point about heroics is the one that makes it the biggest challenge because it's such a big deal and yet cannot be picked up in an algorithm. It's why, I think, a lot of us would have trouble ranking Manning ahead of Montana and even feeling dirty about ranking him in front of Starr; and same thing with Rodgers. These guys don't have the big moments to match their reputation and skill level.
Totally agree. QB heroics are an eye test at-best. If you look at NFL Fourth Quarter Comebacks Career Leaders (since 1960) | Pro-Football-Reference.com is has some HoF names but also has Matt Freaking Stafford. Hes prob there because he had the opportunities because his team sucked.
 
Status
Not open for further replies.

Top