Professional Football Researchers Association Forum

bachslunch wrote:I'm using Kiran Rasaretnam's system as the basis. Layne ranks 67th in "best 4," 50th in "best 7," and 35th in "best 10."

I remember something about his 'new' passer rating. Wasn't it simply completions minus interceptions adjusted for era? Layne ranking 50th in best 7 seasons wouldn't be a surprise with his high number of ints. Though football and particularly playing the position of quarterback isn't simply stats and especially not just two stats.

Does his 'system' take into account play-calling? Essentially on the fly in-game game-planning (who to use and how to use them)? Being a large part of why his team won?

Of course not. Bobby Layne can be evaluated by watching him play football, not by largely irrelevant - and possibly in this case made up - stats.

I will give credit for the best 4, best 7 and best 10. Stats aside, an interesting way to compare players and there's a certain "all things equal", in it. Though i would have chosen different years, the idea is good. Seems like it was something I saw in baseball though long before this guy? Though if he came up with it, good for him.

BD Sullivan wrote: I'm guessing that Blanda's two AFL titles as QB were a microscopic part of his HOF resume, considering they came in the league's first two years when the talent level was at its lowest,

I remember Bob Carroll observed that if it hadn't been for the American Football League, George Blanda would have been a footnote in NFL history. He never appeared in the NFL Pro Bowl, though he was in several AFL All-Star games. Unlike most pro athletes, I don't think he was famous until he was in his 30s.

Reaser wrote:I will give credit for the best 4, best 7 and best 10. Stats aside, an interesting way to compare players and there's a certain "all things equal", in it. Though i would have chosen different years, the idea is good. Seems like it was something I saw in baseball though long before this guy? Though if he came up with it, good for him.

You might be thinking of something Bill James did in his second Historical Abstract. I forget the exact details, but he evaluated players by their total careers, their three best seasons (possibly three best in a row), and five best seasons, trying to get a balance between peak value and career value. The concept is certainly an improvement on the one-size-fits-all use of career stats.

Bob Gill wrote:You might be thinking of something Bill James did in his second Historical Abstract. I forget the exact details, but he evaluated players by their total careers, their three best seasons (possibly three best in a row), and five best seasons, trying to get a balance between peak value and career value. The concept is certainly an improvement on the one-size-fits-all use of career stats.

Sounds about right.

Eli has a lot of black ink on his stat page, but it's mostly in the INT column.

I was surprised to see that he had 35 TDs this year, and that it was tied for 2nd in the league with all those other guys. It didn't seem like he had that good a season. If he can hang around for a few more years with Odell Beckham, I think that there's little doubt that he'll pump his stats up to looking much more Canton-worthy.

Jeremy Crowhurst wrote:I was surprised to see that he had 35 TDs this year, and that it was tied for 2nd in the league with all those other guys.

6 in one game, helps.

Reaser: I think that's the basis for the rankings, though there may be other aspects to it. And given that attempts at sabermetric type thinking is still pretty wild and wooly for football, it's likely that aspects of it could use some tweaking. But we don't have a lot available to choose from. And it's likely that Rasaretnam has passed away given what I saw online, so he certainly won't be able to refine it.

But there are things I like about it, too. Like you, I like the breakdown into small/medium/large chunks of career (4/7/10) in his case. And I really like that the results don't look like they came out of left field. For example, here's the C7 results (I think it only tracks through 2007 or so), first the top 16 names:

Montana, Young, K. Anderson, Baugh, Tarkenton, Dawson, P. Manning, Starr, Tittle, Aikman, Unitas, Graham, Staubach, Warner, Jurgensen, Favre.

Everyone on that list is a HoFer, a sure-fire HoFer, or Ken Anderson. Other HoFers on the list with ranking:

18. Marino. 19. Fouts. 24. Griese. 28. Moon. 30. Van Brocklin. 32. Jim Kelly. 33. Elway. 37. Luckman. 38. Stabler. 50. Layne. 51. Herber. 67. Namath. 70. Waterfield. 74. Bradshaw. 81. Blanda.

There might be some quibbling on exact ordering, but it seems to get a lot right. And the same seems to be true for the C10 and C4. Which likely means that Rasaretnam was either staggeringly lucky or onto something, and I think it could be the latter.

Could it be improved? Maybe so, and if it can be, it should. Sabermetric thinking in baseball undergoes an ongoing process of refinement. Should this be the only criterion for looking at QBs? Not necessarily, but I think it's appropriate to at the very least include it, especially if it produces a set of results that make reasonable sense. And I get that football is harder to do using sabermetric type approaches, but there's no reason not to try. Should film study be an important aspect to deciding HoF fitness? Absolutely, and I applaud what you and Ken have done thus far; I love the detail and clarity of breakdown as well as the attempts to try and quantify what you're seeing relative to other players. If you had a database of 200 potential HoF players, it would be a routinely important part of what I include when I talk about HoF player fitness. But that's a pipe dream given the amount of work it would take to put that together given that it's a labor of love. I'm glad for what's there, though, and it has changed my thinking on some players and confirmed it on others. I welcome anything I can get that seems reasonable and useful.

I also know that there's at least a little suspicion here generally towards the idea of applying stats to such discussions, even stats well considered. I guess I can be a thorn around here sometimes between that and the fact that I think postseason honors can be useful in considering HoF cases. I try to be a good sport about it, though, and hey, it never hurts to challenge things a little. I try to keep my mind open and welcome the jostles when they shake up my thinking.

Anyway. Hope that's all helpful.

bachslunch wrote:But there are things I like about it, too. Like you, I like the breakdown into small/medium/large chunks of career (4/7/10) in his case.

I would use 3 as the beginning, 7 and 10 are fine. Though if I started with 3, then 6 and 9 would probably make more sense.

And I really like that the results don't look like they came out of left field.

This is where I'm not as quick to jump in. Putting aside what I think about stats in football, in general.

It seems as if almost all 'advanced stats' are born out of the same foundation.

(at this point I'm not speaking specifically of the stat he came up with) ...

Play with, combine and manipulate numbers long enough until the final number you come up with produces a list that on the surface is reasonably "in order" (from what's obvious / e.g. a list with Montana at #1) ... The problem is (insert 'new' stat) doesn't predict future success, isn't applicable to much of anything and is solely backwards thinking - as in the result dictates the formula. Rather than someone (and I don't think this is possibly in football, because I don't believe stats do a very good job at all of explaining football) comes up with a brilliant 'advanced stat' THEN applies it to players/units/teams/etc. and lets the result be the result. That doesn't happen because there isn't a brilliant 'advanced player stat' and more importantly, the result wouldn't "make sense".

It's not how 'this' all works. It's toy with it until the order makes some sort of reasonable sense. Then present it to the public as something meaningful while explaining away the 'anomalies' ("why is Player A so high on the list?") as if they don't matter. The final step is the dumbing down of the collective.

Sabermetric thinking in baseball undergoes an ongoing process of refinement.

Which is fine, but the major difference is that baseball isn't football. One sport - being basic about it - has stats explain what happened fairly clearly. The other sport is football.

That said, the more people talking about and interested in football, the better. I have zero problem with 'trying'. My issue is always on the back end, when and how people use stats (out of context, with no context, as definitive statements, etc.) ...

Reaser, thanks for the reply. Couple things.

-my interest in sabermetric type thinking is limited to looking backwards, trying to quantify people's HoF worth in a fair and reasonable way. If I were a GM or a fantasy football type, I might wish for something that predicts forward, but I am neither, and thus this aspect holds zero interest to me. And my impression is that even baseball sabermetrics is not foolproof in predicting the future; as they say, that's why they play the games, and yes, baseball GMs still make problematic signings sometimes even when they have such info. Shin Soo Choo was supposed to be an excellent signing risk before the 2014 season looking sabermetrically, but he was no better than league average in the first year of his contract -- luckily for Texas, he was more in line with how he projected in 2015.

-we're going to have to agree to disagree on the idea of massaging data to try and come up with the best result as being a bad thing. For me, it's trying to "get it right" which I think of as a good thing.

bachslunch wrote:-my interest in sabermetric type thinking is limited to looking backwards.

If the 'advanced stats' and such to have no value moving forward they would also not have value looking backwards. Was my point there. It has to go left AND right. Otherwise you can't apply it across the board to all previous seasons if it can't be meaningfully applied to next season.

-we're going to have to agree to disagree on the idea of massaging data to try and come up with the best result as being a bad thing. For me, it's trying to "get it right" which I think of as a good thing.

It wasn't meant as saying "trying to get it right" is a bad thing. It was about why I'm not impressed by the process and view it as flawed. It's working backwards. Taking what is known (Players A, B and C are good) and manipulating numbers to reach that predetermined conclusion (A, B and C are good. "See, they rank high on the list") ... So move numbers around and voila!, "look guys, I invented something". It isn't impressive and it gets a whole lot worse when sure, A, B and C on the list "make sense", but then D doesn't fit and it's ignored because E and F "kind of" fit ... Doesn't work, for me.

And that is essentially all that current advanced football sites/blogs do. I'd argue that those types aren't even in the business of trying to get the best result and they're just profiting off the sport and the casual fans ignorance. But I don't want to insult any of the advanced stats guys, at least not intentionally. Good for them for finding a niche and as I said, the more people interested and talking about football, the better. What bugs me is they're purposely making people less educated about the sport. Telling them, taking advantage of their ignorance, and making them think that stats tell the entire story (which is more of a complaint off our forums, since most here know there's more to the sport than scrolling through a spreadsheet.)