Walkthrough: Colt Dilemma

Walkthrough: Colt Dilemma
Walkthrough: Colt Dilemma
Photo: USA Today Sports Images

by Mike Tanier

Somewhere near Baltimore…

SALESPERSON: Good morning gentlemen, and welcome to Crate & Barrel! What can I do for you?

LITTLE RAY: We are looking for a belated wedding gift for one of our coworkers.

SALESPERSON: Oh yes, the Joe and Dana registry. That really did sneak up on us, didn’t it? I will assume you are friends of the groom. What is your price range?

BIG RAY: We had to tighten our belts a little because of the lockout. We should keep it under, say, $10,000.

SALESPERSON: T…t…ten thousand? For that price you can purchase that aisle over there!

LITTLE RAY: Anything on that aisle?

SALESPERSON: No, the whole aisle.

BIG RAY: I don’t know about this place, Little Ray. I don’t see a single crate or barrel for sale. Joe wouldn’t have signed up here unless he really wanted some sort of barrel. I think this place is a front or something.

SALESPERSON: No sir! The name of the store is just a clever way of saying that we sell a variety of home furnishings and supplies. Now you gentlemen probably want to purchase something that suggests strength and masculinity. May I recommend a set of high-quality, German-crafted knives?

BIG RAY: I guess that’s supposed to be some kind of joke. Why don’t I just…

LITTLE RAY: Woah, woah. Easy there, Big Ray. He wasn’t trying to be funny. He doesn’t know who we are. Knives might not be the right gift for us. What’s on the registry that is still available?

SALESPERSON: Dish towels, damp mops, plain flatware, brown napkin rings.

LITTLE RAY: Wow, all of those are perfect gifts for Joe.

SALESPERSON: Sorry, I made a mistake. All of those were recently purchased.

BIG RAY: Really? All of that bland, boring stuff? What kind of imagination-less milk toast of a human being would string all of those dull, ordinary things together into one package?

SALESPERSON: They were bought by a ... Mr. Cameron.

BIG RAY: I’m telling you, Little Ray, I don’t like this place. Let’s go next door to Barnes & Noble. I can get an iced coffee, and there’s a whole section of books by or about Michael Oher.

SALESPERSON: Ooh, this may be the perfect gift: a stepladder! Nothing says "marital bliss" like a good stepladder!

LITTLE RAY: Let me try that. Wow! I can almost reach the top shelf with this thing! You're going bald, Big Ray!

BIG RAY: Don’t make me pound you until you are two inches shorter, Little Ray.

SALESPERSON: So you guys are Big Ray and Little Ray? That’s cute. It reminds me of a Dr. Seuss alphabet book I read to my nephew sometimes. “Big Ray, Little Ray, what begins with Ray? Ravens lose a playoff game, Ray! Ray! Ray!”

BIG RAY: Oh man! Hold me back Little Ray!

LITTLE RAY: Easy, easy big fella! Look, Mr. Crate & Barrel guy, we aren’t the best wedding shoppers. Can you give us any advice?

SALESPERSON: Well, men of your means can afford just about anything they want. That makes the feeling behind the sentiment all the more important. The best things you can offer your coworker are sincere congratulations, continued support, and friendship.

BIG RAY: I like that. Best of all, it’s free, so DeMaurice won’t call us and yell at us for spending too much money.

LITTLE RAY: I agree. But now we have other wedding shopping to do. We still haven’t gotten anything for Big Ben, and he’s not registered here.

BIG RAY: Then let’s go to where he is registered.

LITTLE RAY: Right! Hooters it is!

Quarterback Top Fives

Let’s build the suspense by starting with the mighty Texans.

Houston Texans

1 Matt Schaub. Schaub turned 30 last week. He is a solid quarterback, but he is firmly in his prime and will probably start declining in the next few years. One thing these Top Fives have taught me is that most quarterbacks’ peaks are relatively short: four or five years, in most cases. Schaub is already two years in.

2. David Carr. On September 28, 2003, Carr scored a last-second game-winning touchdown on a quarterback sneak against the Jaguars. Dom Capers could have ordered a chipshot field goal to force a 20-20 tie but ordered the sneak instead. It was an incredibly exciting moment, and I thought at the time that it marked the start of a productive career for Carr. Instead, it became his greatest moment as a professional. At least it was unforgettable.

3. Sage Rosenfels. Rosenfels was always interesting. He had a four-touchdown and two three-touchdown games for the Texans, but he also had a four-interception and two three-interception games. He packed a lot of living into ten starts.

4. Tony Banks.

5. Dave Ragone. These happen to be the only five quarterbacks to throw more than 10 passes in Texans history. Rex Grossman was 3-of-9 in Houston as their only other quarterback. Grossman’s efficiency rating was 5.6, lower than those of Jabar Gaffney and Matt Turk. There is an NFL team seriously considering him as a starting quarterback this year. It’s not the Texans.

Baltimore/Indianapolis Colts

1. Peyton Manning.

2. Johnny Unitas.

Manning won four MVP awards and had five other extremely valuable seasons. He has been an MVP-caliber performer every year since 2003, as well as in 2000, when he led the NFL in passing yards and touchdowns for a 10-6 team. He earned a Pro Bowl berth in every season except his rookie year and 2001, the only year of his career that can be considered “off” by any meaningful standard. Sustained excellence of that magnitude for five years is rare, let alone ten. It's an accomplishment no quarterback, and frankly few athletes in any team sport, can match.

Counting MVP-worthy seasons is a good way to differentiate among all-time greats, isn’t it? Counting Pro Bowl berths is great, but the sixth best quarterback in the league can get a Pro Bowl spot, and that's assuming nobody comes up with any phantom injuries they need to rest. Working with actual statistics is fraught with problems once you cross eras. Counting actual MVP awards leaves you with a bunch of players who won two or three, so you do not get the separation you need when comparing Hall of Famers among themselves. We all know the problem with counting Super Bowl rings. Without hunting down MVP votes, we can call an MVP-worthy season one in which the player accumulates outstanding statistics while making a major contribution for a playoff-caliber team. That’s what Manning did in 2000, and in every year from 2003 on.

We lose sight of just how amazing Manning’s accomplishments are because we are living through them, we enjoy laughing at his infrequent slipups, and his consistency makes us take what he does for granted. The last five seasons have been variations on a 4,000-yard, 30-touchdown, 15-interception theme, with his team winning at least 10 games and reaching the playoffs every year. Even for the great quarterbacks in history, a five-year run of playoff appearances, statistical excellence, and perennial Pro Bowl berths would be the centerpiece of a career. While this five-year run does include Manning’s two Super Bowl appearances, it comes after two MVP awards, a record-breaking 49-touchdown season, two passing yardage crowns, a 14-2 season, a 13-3 season, and two 12-4 seasons. If Peyton Manning’s career had started in 2006 -– if everything before that happened in the USFL -– he would be a Hall of Famer and one of the 15 or so best quarterbacks in history (his record would be similar to Steve Young’s in many ways). But his career started in 1998, and that’s why he is arguably the greatest quarterback ever, and unquestionably one of the top five.

Johnny Unitas won MVP awards in 1959, 1964, and 1967. The 1964 award was a little odd –- Jim Brown and Bart Starr may have deserved it more –- but MVP ballots are delicate things. Heck, Manning may not have been the best choice in 2008. Anyway, Unitas had true MVP-type seasons in 1957, 1958, and 1960. Let’s give him 1965 as another one. That’s seven MVP-type seasons to Peyton’s nine. Both players were first-team All Pros five times, though Unitas did it in a much smaller league. Manning has one more Pro Bowl appearance. Manning has won 23 more regular season games with only three more losses.

Careers tend to be much longer now, which gives Peyton an edge, but it’s important to recognize that he has not really entered his decline yet. Unitas’ decline started with an injury in 1968, when he was about the same age as Peyton will be this year. His decline phase included two Super Bowl appearances, one win, and over 6,000 passing yards during a defense-dominated era. It also includes most of the memories many of us have of Unitas, either from life or from NFL Films. Peyton could have all manner of adventures ahead of him: more Super Bowl appearances (possible), another Pro Bowl berth or three (likely), even a Unitas-like stint as the old gunslinger on a defense-oriented team. The fact is that he doesn’t need them to create a very strong case that he is better than Unitas.

One thing that often happens in these arguments is that we adjust the contemporary player down because careers are longer and stats are more prolific, but we never adjust the old timer down for the fact that media coverage has changed. In Unitas’ heyday, sportswriters were almost uniformly fawning and invested in mythmaking. The football media was also rather primitive, compared to the modern football mass media and to baseball media of the time, so players weren’t scrutinized heavily or scouted minutely. We don’t have detailed scouting reports documenting every minor Unitas mistake, long columns explaining how Unitas lacks the courage or gumption to defeat Bart Starr, or bloggers making fun of Unitas’ post-interception facial expressions. We had a hero-champion-warrior king. You cannot compare that bronze bust to the guy who will take the field in September (I type this with ever-increasing confidence) and possibly lose or throw two interceptions. You have to make harsh judgments when comparing old legends to new. You have to notice the fact that from 1961-63 Unitas was the third-to-fifth best quarterback in a 14-team league. You have to remember that Colts did just fine without him the year they reached Super Bowl III, a sign that his “leadership” was not all that important to a team that did darn well with his backup at the helm.

There are other things to adjust for. Unitas won three titles. Winning an NFL championship in 1958 and 1959 meant winning one championship game, no playoffs. The other title came in 1971, when the NFL Colts got to share a division with four newly-arrived AFL teams, three of them terrible. The Jets, Bills, and Patriots combined to win seven games the year the Colts won Super Bowl V. This sounds like I am picking away at Unitas, and I don’t mean to do that to one of the best quarterbacks ever. I am just explaining that “adjustment” is a two-way street. Unitas’ 1950s Colts won 12-team leagues. It was an accomplishment, and a sign of excellence, but not a cudgel that can be used to beat another player over the head for winning a 32-team league “only” once.

There’s a reflexive need to argue against Manning’s greatness, and I noticed it when looking through the message boards in the last few weeks when some of you were anticipating this Colts list. For Manning, we have amazing stats, wins, and longevity. Leadership that in any other era would be universally lauded. Uniqueness and durability at a position where a missed game is potentially a disaster. We toss them all away and point to a handful of playoff losses. The problem isn’t bad here at Football Outsiders, where you guys really delve into the evidence and come away with measured conclusions. In other places, it is pathologically nutty, and some of them aren't even Patriots fansites.

There are many reasons to pick Unitas over Peyton, and I would not go to war with someone who wants to make that choice. But I think some people want to kick Peyton out of the Top 10 of all time, or even the Top 20. To them, I say this: anyone who does not consider Peyton Manning one of the ten best quarterbacks of all time either hasn’t studied the issue at all or is arguing from some kind of goofy agenda. I don’t see any reason to keep him out of the Top Five.

While doing this project, I have spent hours and hours staring at quarterback records: the stat lines, their playoff records, John Maxymuk’s Quarterback Abstract, old encyclopedias, game logs. I had to study Steve Young and Joe Montana, all the Cowboys greats, the 1950s legends and AFL guys. There is nothing, absolutely nothing, in any of the records like Manning’s body of work over the last decade. There is no run of 12-4 seasons and statistical dominance that lasts anywhere near as long, period. Terry Bradshaw and Tom Brady cannot touch it. John Elway does not come close. Neither does Dan Marino, who had a lot of great statistical years for 8-8 teams. Joe Montana has something more checkered, although it is brilliant enough that it is arguably greater. Unitas has 1957-60 and several punctuation marks, so he has a case. Brett Favre has his four-year run and lots of (very good) stuffing. Otto Graham has something that looks better if you can stomach AAFC statistics. But really, Manning’s statistical record is completely on its own, and to write it all off is to write this whole decade off.

There’s the selection, feel free to criticize it.

3. Bert Jones. The Colts ran a proto-West Coast Offense under Ted Marchibroda in the mid-1970s, and Jones had three excellent seasons during the deadest of the Dead Ball Era. Jones then suffered several major injuries, and by the time he returned to health in 1980, owner Robert Irsay was engaging in drunken tantrums in the locker room, taking headsets off coordinators’ heads during games, and alienating just about every productive player the team had.

As we have written about elsewhere, Jones’ 1976 season (3104-24-9, 60.3 completion percentage, some rushing value) is one of the best quarterback seasons ever once you account for offensive levels and season length. That season may be highlighted on an upcoming edition of NFL’s Top 10, so keep your eyes open.

4. Jim Harbaugh. Harbaugh had one of the longest careers as a custodial starter ever. His specialties were avoiding interceptions and making plays on the run, and he was the perfect quarterback for a team that wanted to hand off 30 times per game because he didn’t have any delusions that he was a star. Because he was smallish, spunky, hard working, of European descent, and had much of his success in his mid-30s, he became the kind of quarterback who gets undue credit for his team’s success. It's a phenomenon I am starting to call Sportswriter Wish Fulfillment Syndrome.

5. Earl Morrall. Morrall led the Colts to Super Bowl III, leading the NFL with 26 touchdowns. He won spot starts in 1969 and 1970, then posted a 7-2 record for a very good Colts team in 1971. There aren’t really any worthy honorable mentions, as Unitas, Manning, and Jones ate up much of Colts history. Jeff George did have some productive years. Guys like Mike Pagel and Jack Trudeau played hot potato with the starting job for many years but weren’t ever very good.

And Finally

I was supposed to do some statistical research but fell behind last week; sorry for that.

Obviously, many of us are holding our breath for good news on the NFL labor front. Every setback (like Thursday’s) comes with a heavy shudder, every nugget of good news has us bracing for some around-the-clock work to get things like the Football Outsiders Almanac done. A lot of us are moving family vacations up to, well, now, which is why I was at Dutch Wonderland instead of in the data mines in the first half of the week.

Some promotional materials for The Philly Fan’s Code arrived this week, and I just edited the galleys. Take a look at the ads on this page and you will see that the book is available for pre-order at Amazon. There will be public appearances and signings in the Delaware Valley. Details will land here first. I hope to have some news about other exciting new projects in a few weeks. I also hope to be preparing for training camp.

Happy Independence Day!

Comments

272 comments, Last at 22 Jul 2011, 7:30pm

102 Re: Walkthrough: Colt Dilemma

You speak as if Harbaugh never had a chance to develop his potential. But he played 14 years and started 140 games. He was 66-74 in those games, 59% completions, 129 TDs and 117 INTs. In addition to his insane numbers in Canada, Flutie still played 12 years in the NFL and went 38-28 as a starter, 55% completions, 86 TDs and 68 INTs.

103 Re: Walkthrough: Colt Dilemma

You speak as if Harbaugh never had a chance to develop his potential. But he played 14 years and started 140 games. He was 66-74 in those games, 59% completions, 129 TDs and 117 INTs. In addition to his insane numbers in Canada, Flutie still played 12 years in the NFL and went 38-28 as a starter, 55% completions, 86 TDs and 68 INTs.

117 Re: Walkthrough: Colt Dilemma

I think I'd go with Harbaugh, as I think he could have been a legitimately good QB if he had the right development environment. He could have been Jeff Garcia-esque. While I think Flutie did eventually reach his natural peak, and it wasn't that good.

Yes, but the Flutie 'peak' you're talking about came at age 36; it's plausible that his ceiling was really much higher.

136 Re: Walkthrough: Colt Dilemma

Rather than compare Unitas and Manning, I'm going to take a minute to grade Tanier's analysis by FO standards, point by point.

Preferring "MVP-caliber" to "Pro-Bowl" as a "stat". -1 point.
The made-up stat "MVP-caliber season" seems to allow at least as much extra room for QBs as Pro-Bowl does. Manning won 3 solo MVPs, but seems to be credited with 12 MVP-like seasons. Sure sounds like MVP-caliber is similar to Pro-Bowl back-up quality to me.

Quoting cumulative stats and ignoring rate stats. -2 points.
This is a classic FO no-no. It's especially bad when comparing seasons of different lengths. Tanier quotes only seasonal cumulative stats. This is a red flag for bad analysis.

Altering stats to suit his thesis. -1 point.
Normally, this would disqualify the analysis as biased and bogus entirely. Tanier denigrates Unitas' accomplishments several times (e.g. "1964 was a little odd") but backs off a bit each time. It's as if he knows he's cheating, and wants to pass it off as a joke. Sorry. Still a point against.

Using made up stuff. -3 points
"Peyton could have all manner of adventures". Including a hypothetical future with specific accomplishments is just bogus. Simply comparing the first 13 seasons of each would have been valid.

Acknowledging the need to adjust for era. +1 point
Failing to actually adjust for era. -3 points
Even cumulative seasonal stats can be adjusted, but not here. Why not just say "Unitas was so lazy he didn't even play as many games each season!" and be done with it. Comparing across eras without era-adjustments is naive if you don't know better, and unforgivably sloppy if you do. Tanier knows better.

Adjusting accomplishments based on "media" and "mythology". -2 points
Instead of era-adjustments, Tanier gives us media/mythology adjustments, calling 1960's media "fawning" while implying that Manning has been unfairly characterized in the 2000's. Media coverage is not relevant. Or it wouldn't be, except that Tanier is trying to convince us that his gushing, fawning, subjective evaluation of Manning is somehow more balanced than the 1960's media. Methinks the lady doth protest too much.

Bad adjustments for league size. -1 point
League size can be used to adjust for some things, but not others. Being the best of the top 12 QBs in the world is pretty much the same as being the best of the top 32 QBs in the world.

Ad hominem attacks. -1 point
Calling anyone who disagrees "pathologically nutty" is insulting. Dismissing them as "Patriots fansites" is an admission that your arguments are so weak that you need to dismiss the messengers rather than address the facts.

Dependence on the "Wins" statistic. -1 point
Sometimes you have to include "wins" in a discussion. But in comparing two QBs, it's a weak stat to build an argument on. Coupled with made-up stats (MVP-caliber season, forsooth!) and misuse of cumulative seasonal stats, it's pretty much a classic botched analysis.

I'll give Tanier credit for being readable and passionate. But for someone who claims to have looked at a lot of stats to form his opinion, he sure did a rotten job of making his case.

116 Re: Walkthrough: Colt Dilemma

It has to be Unitas, right? I mean, he has a haircut you can set your watch to. That's enough of a tiebreaker for me.

120 Re: Walkthrough: Colt Dilemma

While that's true, and clearly bests Joe Namath's long flowing sideburns which open whole new worlds and push frustrated housewives to lives of abandoning their families and living on the lam, it also puts to bed another argument:

Peyton Manning's haircut is clearly a haircut you can set your watch to when compared with Tom Brady's flowing locks.

143 Re: Walkthrough: Colt Dilemma

At last, a compelling argument for Manning over Brady. Given their faces, shape of head, etc, I can't even make a "strength of supporting cast" case for Brady. Manning overcomes adversity for the win!

119 Re: Walkthrough: Colt Dilemma

My theory on Jim Harbaugh being dubbed "Captain Comeback", even though he only had a merely above-average 3 comebacks in 1995:

1. He had a very good season for a team that wasn't expected to do anything.
2. He led two comebacks in dramatic fashion early in the season against the Jets and Dolphins.
3. He had a few losses (the Colts lost 8 games that year including playoffs) where he made a decent attempt at coming back (tied the game several times), but ultimately didn't complete the job and Indy lost.
4. He's probably remembered best for the almost-comeback; the dropped hail mary in the AFC-C in Pittsburgh.

Even though he has just as many comeback wins as Roger Staubach (15) for his career, Harbaugh was not deserving of the title "Captain Comeback". Especially odd when you have two other QBs for the Colts, Unitas and Manning, that will be record setters for comebacks.

127 Re: Walkthrough: Colt Dilemma

Not counting the dropped hail mary, he had four other games that season where he lead drives to tie or take the lead in the 4th quarter, only to have his defense lose the game.

So he had 1 GWD, 3 comebacks, and 4 games were he came back, but lost anyway. Indy played a *lot* of close games that season.

126 Re: Walkthrough: Colt Dilemma

"Schaub turned 30 last week. He is a solid quarterback, but he is firmly in his prime and will probably start declining in the next few years. One thing these Top Fives have taught me is that most quarterbacks’ peaks are relatively short: four or five years, in most cases. Schaub is already two years in."

Nnamdi then clicked to the last slide in the PowerPoint: 5yrs, $75mil. Knowing he had been defeated, Bob McNair shed a single tear.

133 Re: Walkthrough: Colt Dilemma

I was really amused by one claim the author made "we adjust the contemporary player down." I think just the opposite is the case. We all want to think we are watching the greatest players now so we tend to give current players the benefit of the doubt. Only a few people posting here ever watched Unitas and most of us who did saw the tale end of his career. People just don't want to believe that there were amazing athletes back then. Everyone looks so slow on the old black and white films. I think that makes people discount players from previous eras. Athletes are not inherently better now. Nutrition is better, training is better and equipment is better, so performance has improved, but we haven't evolved as a species since 1957. That is why you have to compare people to the peers of their era. How does a guy measure up against the people who played when he did?

145 Re: Walkthrough: Colt Dilemma

To wrap up Unitas v. Manning (sure, this'll work):
Era-adjusted TD rate: tie
Era-adjusted Comp%: Unitas
Era-adjusted Yards/Attempt: Unitas
Era-adjusted Interception rate: Unitas, by a mile.
Era-adjusted sack rate: ?
Era-adjusted fumble rate: ?
Era-adjusted playoff stats: ?

Easier conditions (relative to era): Manning (worth 6 inches per attempt, see PFR link elsewhere in discussion)
Easier division: Manning
Better supporting cast (relative to league): ?
Easier defenses faced: Unitas (worth 2 inches per attempt, according to PFR)

Every QB-specific rate stat we have points to Unitas as the better QB, and the one adjustment that works against that conclusion is too small to matter. There are ways to fudge the numbers (e.g. compare Manning to the 16th best QB and Unitas to the 8th best; use counting stats rather than rate stats) but without such fudging, Manning comes in second. Still a great QB, but not as great as Johnny U.

"Almost as good as Johnny Unitas" is pretty damned good, if you ask me.

146 Re: Walkthrough: Colt Dilemma

Those weren't the only criteria.

Durability: Manning
Consistency: Manning
Volume (Which does matter): Manning
Longevity: Manning

All of these criteria Manning destroys Unitas in.

I don't grant you 'easier division' is even a category, because PFR showed Manning played tougher defenses overall, so I'm not sure how 'easier division' correlates to anything other than consistency of playoff berths, and even so the division gap is not so great as to account for all of Indy's success over the decade.

Rate stats are vitally important, but they don't tell the whole story and to pretend that small advantages in rate stats over come massive advantages in volume stats is disingenuous at best.

149 Re: Walkthrough: Colt Dilemma

Durability: needs to be era-adjusted. Plus, why care? They both have had long enough careers. Unitas overcame adversity that Manning never had to face.

Consistency: Are you proposing a "low variance bonus"? Why? Of the four QB-rating stats, only completion % has a standard deviation difference worth talking about. Unitas has five seasons where his era-adjusted completion percentage is better than Manning's best year, and one year that was lower than Manning's worst. That's why his std deviation is so high. He was better than Manning's best five times.

Volume: needs to be era-adjusted. Also, what would a high era-adjusted volume mean? That your team threw the ball a lot? So what?

Longevity: needs to be era-adjusted. It's much easier to have be a healthy QB in 2010 than in 1968. Also, aren't you giving Manning credit for the imagined future? Shouldn't you also give him credit for the expected late-career stats decline? Hmmmm? Or better yet, stick with reality.

Easier division: this came up in the context of QB wins, where you must consider the whole team, not just the defense. Defense was also covered in another point. Credit Manning with an additional 2 inches per attempt, if it makes you happy. Don't forget to ding him 6 inches for playing in a dome, too, to use the full results of that PFR analysis.

"Volume" is just a word for relying on Manning's pass-heavy era, pass-wacky coaching, and counting stats to make your case.

Nope: All you got is counting stats and ignoring different eras. That ain't gonna do it.

151 Re: Walkthrough: Colt Dilemma

Durability needs to be era adjusted? Sorry, I get the argument, but I don't buy it at all. Manning missed one play his whole career. You can't adjust that for era. He was available to play more often than Unitas was. Might that be a product of modern training?

Sure, but it doesn't change the reality that Manning did things Unitas was physically not capable of.

Consistency: Awful argument. Unitas had awful seasons on his record. 1961, 1962 1966, 1969. You can't pretend those whole seasons didn't exist. Manning's run of unbroken greatness is unparallelled. He simply hasn't had a truly 'bad' season since his rookie year. Certainly nothing approaching the badness of several of Unitas's years.

Volume: certainly needs to be era adjusted, but my point is that slight dips in efficiency (only era adjusted BTW, not real)are way over come by massive jumps in volume by Manning. Completing 65% of 350 passes (era adjusted) is great, but completing 64.5% of 600 passes is even better. That's the point with volume. Manning was nearly as efficient (era adjusted) over a FAR greater volume.

Longevity: Nope. Not buying era adjustment here. Manning will have a longer career. That matters. More great seasons is more great seasons regardless of the era comparison. Era adjustment doesn't make it that whole seasons never happened. If the two QBs are truly neck and neck, I'd rather have the one who played 5-7 more prime seasons than the other. Hands down. In fact, I'd say the gap between the two players would have to be quite wide before I'd take the player with fewer prime seasons. He doesn't get credit for seasons he didn't play.

Frankly, you are using era adjustment as a massive crutch to make up for the fact that Unitas had many whole seasons where he not only wasn't one of the three best QBs in the league, but wasn't even very good by any measure.

I'm not buying it. Your arguments strike me as poor rationalizations. No one is ignoring era, but relying on rate adjustments gives players too much credit for what they didn't actually accomplish.

153 Re: Walkthrough: Colt Dilemma

Durability needs to be era adjusted? Sorry, I get the argument, but I don't buy it at all. Manning missed one play his whole career. You can't adjust that for era. He was available to play more often than Unitas was. Might that be a product of modern training?

Not just modern training, but modern medicine as well. Manning has had several surgeries and has recovered from them in time to not miss any time. Do you think that would have been possible in the 1960s?

166 Re: Walkthrough: Colt Dilemma

No one said that Unitas was more durable than Manning. The contention was that avoiding injury may be easier or harder to do in different eras. Was missing a game very unusual in Unitas' era? A season?

Without an era baseline, we can't say how unusual Manning's health run is. The equipment was different then. The training and medical practices are much improved now. On the other hand, QBs are now sitting out games for concussions, where they used to go back in. Perhaps the next ten years will make sitting a few games out the rule rather than the exception.

168 Re: Walkthrough: Colt Dilemma

Without a baseline we can't know how unusual Manning's run of health is?

Please. He's missed one play his whole career.

That's pretty freaking unusual. If you need a statistical baseline to tell you that, then you aren't very observant.

174 Re: Walkthrough: Colt Dilemma

I'm not saying Manning doesn't have freakish durability, I just think it's likely that if he had been born in 1945 he probably would have missed a couple games by this point in his career.

Actually though, one of the knee surgeries he had could have ended his career in about 3 years. Just look what happened to Gale Sayers.

177 Re: Walkthrough: Colt Dilemma

He had a bursa sac burst. It's seriously unlikely he would have had his career ended by it.

Obviously, modern medicine helps. It also keeps elite defenders elite too. My point is that by any standard, what Manning has done in terms of durability and ability to stay on the field is virtually peerless. Even Favre missed plays and partial games over the years.

Saying that we need some kind of statistical measure to verify Manning's durability is just being argumentative.

252 Re: Walkthrough: Colt Dilemma

It's about the inability in Unitas' time to have a knee cleaned up arthroscopically and getting back to a full workload within a month. (If I need to say this, it's not a shot at anyone, just an acknowledgement that Manning is, like everyone now, able to take advantage of advances in medicine.)

152 Re: Walkthrough: Colt Dilemma

Unitas has five seasons where his era-adjusted completion percentage is better than Manning's best year, and one year that was lower than Manning's worst. That's why his std deviation is so high. He was better than Manning's best five times.

Talk about using some downright nutty math. How is it even possible to arrive at that conclusion?

Unitas led the league in comp. % one time. He completed 58.5% in 1967, which is also his career high for any season. The league-average was 51.0% (+7.5%).

Manning has led the league in comp. % one time. He completed 67% in 2003. The league-average was 58.8% (+8.2%).

You mind explaining how Unitas can have five seasons better than Manning's best?

159 Re: Walkthrough: Colt Dilemma

I used the era averages. And I used (an approximation of) the median of the top 16 teams, rather than the whole league average.

Why? Because it's not relevant whether Manning is farther above the 16th best QB than Unitas is above the 8th or 6th best. When a league expands, it doesn't add 16 new QBs who are as good as the first 16. It adds 16 worse ones. Doh! If we expanded the league to 200 teams, would you still want to use league average to determine the era-adjustment? Of course not.

Manning's era average for completion % was about 60%. I approximated the bump up to get to that top 16 team median at 2%. I choose that 2% bump from looking at 2010 data. It's approximate. Feel free to build an exact era-average for the top 16 QBs. But I used 62%. I was and am open about that, and why. Short answer: I'm comparing apples to apples.

Manning's best was 68.8%, which is 6.8 above the "top 16 median". As I noted before, using a median HELPS Manning's case. It would be fairer to use the average, which would make Unitas look even better.

Unitas's Manning-era-equivalent (adjusting for his era's 51.3% median for the top 16 teams) is an average of 66.5%, less than 1% above Manning's average. But his peak was 70.7%. That was the year he was 7.2 unadjusted points above league average - and this was in a league that did not have 16 also-rans to bring down that average.

How did he do it? He did it by being really good. How could he beat Manning's best five times? Because Manning didn't have the COMP% peaks (or valleys) that Unitas had. Doh.

Unitas was farther above the best 12-16 starting QBs in his world than Manning is today. Not by much (except in INT%). But in every category except a tie in TD%.

Oh, and "nutty" is not the right word for doing era-adjustments. It's simply what you have to do to compare across eras.

171 Re: Walkthrough: Colt Dilemma

No offense but your methods are pretty much bunk. You also fall into the common trap of pretending the league just expanded to twice its size in a short period of time. That's so far from the truth. When the merger happened in 1970, the league didn't just inherit 10 new teams. They absorbed 10 teams that had been developing in the AFL, most of them for as long as a decade (1960). Those guys knew how to play football and there were many talented players from the AFL. Then they added the Seahawks and Bucs in 1976. There wouldn't be any more teams added until 1995, when Carolina and Jacksonville made it 30. Adding two more teams since then hasn't diluted anything. There are more people playing football than ever before, and the competition is stronger than ever. Along with free agency, that's why we're seeing more upsets, more Wild Card teams winning road playoff games, and more quick turnaround stories. The gap between best and worst is smaller than it was in Unitas' days. Just because there were fewer teams doesn't mean those teams were loaded. They followed the same normal distribution you get today: some teams were elite, a lot were average, and some were terrible.

Call me crazy, but consistently being near the top of a league of 32 teams is more impressive than a league with 12-16 teams. You seem to think Manning ranking 6th or higher with 32 teams is somehow not as good when Unitas had 6 seasons ranking 7th or higher in a league with half the teams. It doesn't make any sense unless you fudge the math to make for Unitas like you've done.

182 Re: Walkthrough: Colt Dilemma

Okay, I'll call you crazy. You're missing the point entirely.

The average of the best 16 of anything is always higher than the average of the best 32 of the same thing.

Being one standard deviation above the average of the top 16 QBs is more impressive than being 1 standard deviation above the average of the top 32 QBs. The average is higher (obviously!) and the standard deviation is larger (assuming the pool of people willing to play professional QB is much larger than the league).

But that's exactly what PFR's + stats are measuring.

It's like comparing Celsius and Fahrenheit. The zeroes are different and the degrees are different. You have to convert to the same scale and zero-point before you decide who's hotter.

183 Re: Walkthrough: Colt Dilemma

"The average of the best 16 of anything is always higher than the average of the best 32 of the same thing."

You have to realize that's ridiculous right?

You are assuming 1950s/60s QBs are the same as 2000s QBs and that the 'best' are equally distributed in both eras.

There's simply no basis for that conclusion.

1950s/60s era QB was of a vastly inferior quality overall to modern QB play. It's not just about a 'dead ball era'. It's about actual quality of play. We aren't tweaking for defensive adjustments or quality of equipment. We are comparing 32 professional coached up quarterbacks all making millions of dollars to a group that largely consisted of part time players who not only had offseason jobs, but many of whom PLAYED OTHER POSITIONS during the actual games.

185 Re: Walkthrough: Colt Dilemma

I call bullshit.

I am not assuming the QBs are the same. I'm adjusting for era to deal with the differences, such as training, nutrition, medical care, injury-preventing equipment, rules, schemes, officiating, coaching, pay, even use of performance enhancing drugs.

If you don't believe in adjusting for eras, just say so. You'd be alone, but you'd be honest.

But if you believe in adjusting for eras, then you have no statistical basis for saying the 1960's QBs - adjusted for era - were inferior. What does that even mean? The best in each era is still the best. The tenth best is still the tenth best. The average of the best 16 is still the average of the best 16. And the average of the best 16 is still not the average of the best 32.

186 Re: Walkthrough: Colt Dilemma

You are adjusting for era in a completely non-standard way that assumes that the top 16 QBs in 1958 are inherently equal to the top 16 in 2008.

By your method, the WORST QB in the 1958 NFL is equivalent to the 16th best QB in 2008, just because both were 16th in the league.

I believe in adjusting for eras. It provides a sense of context. I don't believe in adjusting for eras in the manner in which you are doing it. You aren't attempting to figure out what the average QB looked like in 1958 and compare Unitas to that. You are assuming the 16th best QB in 1958 is analogous to the 16th best in 2008 and comparing Manning to the top half of QBs in his day to the entire pool of QBs in Unitas's day.

That simply makes no sense, and relies on very dubious assumptions about the quality of play in 1958.

192 Re: Walkthrough: Colt Dilemma

The average of the best 16 of anything is always higher than the average of the best 32 of the same thing.

However, QBs in the NFL in 1960 and QBs in the NFL in 2010 are not the same thing.

Just look at the population growth of the country since then, not to mention further integration (think McNabb would be a starting QB in 1960?). Plus other things like better scouting, better preparation in college, better coaching, better deliniated responsibilities of the job.

194 Re: Walkthrough: Colt Dilemma

Two sensible concepts. Integration and population growth could in fact increase the size of the talent pool. Population is up by about 50% between the eras. That doesn't mean we expect the average QB to 50% better. It might mean that we should be comparing against the top 24 QBs (2010) vs the top 16 QBs (1970) to get the same part of the talent pool. But that's essentially assuming that the modern players are better rather than proving it, which runs counter to the idea of era-adjustments. Regardless, the population of starting NFL QBs has grown much faster then the US population, so using league averages still results in grade inflation.

Integration is a tricky issue. On the face of it, you'd expect the percentage of non-white QBs to approximate the population at large. Unless you subscribe to ethnic superiority theories, which I don't. That might be 2-3 QBs who deserved to start in the NFL in 1965. But if you look at various leagues in various sports, you'll see that it's a lot more complicated than that.

As for scouting, prep, coaching etc, that's all part of the eras that should be adjusted away, along with rules, medicine, drugs, and the ability to study film of Joe Montana to improve your game. That's why we compare each QB to his contemporaries rather than using unadjusted stats. We already know that Manning has had many unearned advantages over Unitas due simply to being born in a different era. We're trying to compare what each of them did with what they were given by their era.

I think everyone is already on board with the idea of era adjustments. The discussion is about how to do them without introducing grade inflation.

196 Re: Walkthrough: Colt Dilemma

Your methodology is unorthodox, so you can't expect people to accept it on face value. You are treating this as simple math problem, but it isn't. Statistics are just descriptions of history, and we have to be careful to make sure that we use them in a way that's intellectually honest with what really happened. The NFL didn't just add 16 QBs worse than the 16 that came before, so while that reasoning makes a rudimentary kind of sense, it's simply not descriptive of what really happened to the league.

The philosophical issue is this: is grade inflation the biggest worry in figuring out era adjustments?

I don't think it is. In fact, just in looking over the numbers from those years, I strongly suspect most era adjustments will naturally be too kind to the older players. That is to say, the PFR method is already over-generous to Unitas, but I think it's probably the best that can be done, and I support it.

Your method paints a wholly inaccurate view of the league, however, that just adds 16 new QBs to the league that are all worse than the previous 16.

Essentially this debate comes down to this question:

is the general state of quarterbacking better today or in the 1950s and by how much?

I say, Yes. It's a lot better by a wide margin.

If nothing else, I'm willing to grant that the league is distributed proportionally between great, good, average, and bad QBs similarly to in the 50s and 60s, though I suspect that's over generous to the past.

You say QB play was proportionally better in the past, hence even averages need to be adjusted to reflect that.

The public will decide, I guess.

As for Manning's 'advantages', you can't 'stat away' all of them. You can adjust for scoring rates, even percentages of times passing the ball, and obviously games played per season, but there's some degree in which you can only judge players by what they really did on the field. Hypothetical arguments about what Unitas would have done with modern surgery are fun, and you can wonder if he wouldn't have been better with modern coaching and game film, but none of that changes the fact that there were things he didn't accomplish that Manning did.

Manning has done far more than Unitas did on the field. At some point trying to explain that away becomes what Tanier called 'justification'.

197 Re: Walkthrough: Colt Dilemma

In the end, you back away from the idea of comparing Unitas and Manning to their peers. That's fine, although you should ask yourself honestly why you only took up that position once Manning didn't come out on top. At the beginning of this discussion, did you come out against era adjustments? Really?

200 Re: Walkthrough: Colt Dilemma

I don't back away from comparing Manning and Unitas to their peers. I supported it.

I don't support your odd methodology which compares Unitas to all QBs and Manning only to the top half of QBs.

There's a massive difference.

Manning comes out on top in a more reasonable system of comparison.

206 Re: Walkthrough: Colt Dilemma

Pop growth: Unitas was born in 1933 when the US population was 125.6 million, Manning in 1976 when it was 218.0 million, an increase of 73%. Furthermore, the dramatic increase of NFL popularity, a goodly portion due to Unitas' play (which is an "effect on the game" argument, not "best QB") has to have resulted in a greater portion of the total population playing football for 1976-born than for 1933-born, though I consider that latter difference non-quantifiable or nearly so. Today's much larger pool will affect the very top (and very bottom) the least, but will affect both extremes.

Integration: How many black starting QB has been the average since expansion to 32 teams? Or perhaps since the increase to 30, for a larger sample size. Whatever the number, Nat's calculations ought to subtract it from 32 to make a reasonable comparison, as those athletes were not in Unitas' era QB pool.

If, say, there have been an average of 5 black starters in the 32-team era, the comparison begins at 27 QB vs 16. Adjusting that 27 by 125.6 mm and 218.0 mm (no allowance for greater football interest) leaves 15.5. Too many assumptions to take this last paragraph very seriously, but the top two should carry some weight.

Edit: DisplacedPackersFan made the same arguments earlier in post 203, only better.

262 Re: Walkthrough: Colt Dilemma

Sorry, nutty is exactly the (polite) word for what you are calling era adjustments. Statistics FAIL would be the modern internet way.

If you use your eyes you'll see that Unitas was a modern quarterback in a time when the league was in transition. Thus, instead of speculating that QB16 in his era was the equivalent of QB16 in the modern era, it makes more sense to compare Unitas only to the 5-10 guys who were actually doing something equivalent (i.e. focusing on modern offensive concepts, working out timing full time with receivers, not playing two ways or kicking). He was the best out of five or ten actual peers.

Manning is the best among 32 starters today. Population takes care of deepening the competitive pool to match expansion all by itself. The growing popularity and pay for NFL QBs make the competitive pool even deeper. And integration does as well (consider the percentage of the NFL, or say, or high school footballers, which is of African-American descent, not the total U.S. population). And greater player movement means a greater chance that the best QBs hold starting positions or even roster spots. Guys like Brett Favre, Matt Schaub, Matt Hasselbeck, Rich Gannon, Brad Johnson, and Kevin Kolb would have lost multiple seasons as starters (and some might never have seen the field) in the old system. I would guesstimate that it's more like the 65th best QB in today's league who is equivalent to QB starters 12-16 fifty years ago. My personal attempt at era adjustments would compare Unitas to the top 10 of his league as a baseline and Manning to the top 32 of his, and I think the adjustment is probably overgenerous to Unitas.

In any event, mere understanding of statistical distribution would reveal the fallacy of comparing Unitas to his top 5 and Manning to his top 5. If you have 32 QBs, the top 5 averages will be much higher than if you have 16 because flukes like injuries to a top guy won't pull down the top 5 as much and because you have a greater chance for average guys to have career years simply because there are twice as many rolls of the dice. So even if you wanted to do, as Deshawn did, a top QBs of era comparison, the pool against which you compare Manning should be larger. Do Unitas vs. Top 3 and Manning vs. Top 5, 6, or even 8.

Finally, your comments about NFL failures is laughably hypocritical given that Unitas couldn't even get a training camp snap on one of the worst teams in the NFL and was subsequently playing for $6 a game.

150 Re: Walkthrough: Colt Dilemma

If you're coming up with Unitas leading in era-adjusted comp. %, then you're not calculating things right and I wonder if any of your stats are right.

Manning has ranked in the top 6 in comp. % for 12 straight years. He's been in the top three 8 times. And this is in a league where he's competing with 31-35 players a year for efficiency stats.

Unitas has 10 seasons where he ranked in the top 6 in comp. %. The difference is he often did this in a league where he only had to compete with 11-18 other players for efficiency stats. He only has four seasons in the top 3, half of Manning's total.

154 Re: Walkthrough: Colt Dilemma

TD%+: Manning 119, Unitas 105
Comp%+: Manning 119, Unitas 111
Y/A+: Manning 114, Unitas 110
Int%+: Manning 106, Unitas 108
Sack%+*: Manning 123, Unitas 108

Manning: http://www.pro-football-reference.com/players/M/MannPe00.htm#passing_advanced::none
Unitas: http://www.pro-football-reference.com/players/U/UnitJo00.htm#passing_advanced::none

*Only covers the last 5 years of Unitas's career, but it is 5 years of data. It's probably still fair to ignore it.

Is there a reason why you are completely dismissing the p-f-r "advanced passing" metrics?

160 Re: Walkthrough: Colt Dilemma

PFR normalizes to league average, not to top-16-team average. As a result, it is biased against the smaller, earlier leagues. Manning is, I agree, farther above league average than Unitas was. But that's just grade inflation from diluting the average with 16 second-rate (by definition) starting QBs.

In other words: scoring 100 on Comp%+ makes you the 16th best starting QB in Manning's era, and a much better 8th best in Unitas' time.

They also scale based on standard deviation. 15 points = 1 std dev. That's great to put everything on a ~60-140 point scale. But it screws up comparing values across eras, since each era gets its own scaling factor. I'm pretty sure this problem gets worse when you compare a 16 team league with a 32 team league. When you add 16 also-rans to the mix, you're adding QBs from the flatter part of the talent bell curve, which reduces the variance, in addition to lowering the average (as noted before).

In other words, boosting your score from 100 to 115 in Manning's era probably doesn't take as much additional skill as moving from 100 to 115 in Unitas' time.

I really did think this out, you know.

It's cool that they have a sack%. When they get it for his whole career, we should look at that, again comparing 16 team averages and skipping the sack%+ std. deviation scaling to avoid grade inflation.

164 Re: Walkthrough: Colt Dilemma

That sucks. Sack rate does indeed matter. Of course, it's also highly dependent on the OL, not just the QB. And since it's also very scheme-related, the eras are going to be quite different too as rules and schemes have evolved.

It's probably less important than COMP%, Y/A, TD%, and INT%. But it would be nice to have for the complete picture. Ditto for Fumble rate, which would need to cover muffed handoffs - probably as a separate rate? I'm not quite sure.

(edit: all the stats are dependent on the OL, obviously. Sack rate is the most dependent, or so I would suppose)

165 Re: Walkthrough: Colt Dilemma

Sacks are definitely important. I'm fond of pointing out, because it's an extreme example, that sack yardage alone accounts for a 1 yard per dropback difference between Doug Flutie and Rob Johnson in 1998.

There are simple fumble totals available, so it is possible to have a rough idea of that one, though it would be a bit of a blunt instrument. It also could be used as a roundabout way to estimate Unitas's sack rate, though it would also only be a rough estimate.

170 Re: Walkthrough: Colt Dilemma

That's precisely backwards. A top-16 metric punishes a current player and rewards a player from a smaller league. Top-16 is currently "above-media". Top-16 in 1958 would be all the starters and 4 scrubs. That's a below median starter sample.

So if you thought this out, all you were thinking of was how to artificially boost Unitas' scores.

269 Re: Walkthrough: Colt Dilemma

Exactly. The 12th best QB in 1958 was putting up a QB rating somewhere between 40 and 50, averaging under 120 yards per game, and throwing 50% more INTs than TDs. The 12th best QB last year was either Kyle Orton (by DYAR) or Drew Brees (by DVOA). Orton threw for 3600/20/9 in 13 games, while Brees had 4600 yards, 33 TDs, and 22 INTs. Does anyone truly, honestly believe that either of those guys is even remotely comparable, talent-wise, to the worst QB in the NFL in 1958? It's nonsense of the highest order. Ludicrous, indefensible nonsense.

I think it's far more realistic to assume that the QBs that got added to the league where not, in fact, the 16 worst QBs in the NFL... but that they were instead of middling quality. The top 3 in a 12 team league were probably equivalent to the top 6 in a 32 team league, the bottom 3 in a 12 team league were probably equivalent to the bottom 6 in a 32 team league, and so on down the line. To pretend that there were no brutally awful QBs in Unitas's time is to be blindly (and possibly even willfully) ignorant of history.

173 Re: Walkthrough: Colt Dilemma

"PFR normalizes to league average, not to top-16-team average. As a result, it is biased against the smaller, earlier leagues. Manning is, I agree, farther above league average than Unitas was. But that's just grade inflation from diluting the average with 16 second-rate (by definition) starting QBs."

Total fallacy. In fact, as was described earlier in the thread the overall level of QB play in Unitas' era was far worse than today. Teams hoarded QBs, so the 16 starters were by no means the 16 most talented, and in many cases Unitas was competing against players who didn't even play QB full time.

If anything, your method of adjustment is giving FAR too much credit to the overall quality of play in Unitas's era.

He dominated an era, but it was a terrible era. That doesn't mean he was better than Manning, just perhaps incrementally more efficient using non-standard rate adjustments that you invented that assume the 8th best QB in 1958 was as good as the 16th best QB in 2008.

I'm sorry, but that really stretches credulity. Some of the numbers from QBs even in the top ten from that era are so ungodly awful that I simply can't accept your assumption. In fact, given segregation, coaching methods and roster construction, it's far more plausible that there are more better QBs playing now as a percentage of the whole than there ever were in the 1950s.

Take 1958 as a random example. The 8th best passer rating was Van Brocklin who threw 15 TDs 20 INTs for 2400 yards completing just 52% of his passes for a rating of 64.1 for an Eagles team that won 2 games. AND HE WENT TO THE PRO BOWL.

I'm sorry, but era adjustment or not, I cannot accept that that level of play was anywhere near 16th best in the NFL in 2008. It's a complete fabrication.

Unitas was an Everest in Kansas to be sure, but your reasoning for reworking era adjustments doesn't hold water.

184 Re: Walkthrough: Colt Dilemma

"PFR normalizes to league average, not to top-16-team average" - that's indisputable.
"Manning is, I agree, farther above league average than Unitas was." - fact which you agree with.
"But that's just grade inflation from diluting the average with 16 second-rate (by definition) starting QBs." - fact that you dispute???

But seriously, are you arguing that the 17-32 QBs in today's NFL are just as good as the top 16? Because if you aren't then it's indisputable that including them in the average brings the average down, and biases the analysis.

Of your other points, just one is worthy of consideration: segregation. The talent pool of people allowed to and motivated to train and develop into professional quarterbacks has grown faster than the population, and not just by including more mediocre talent. It's one aspect of the changing eras. Should we try to adjust for segregation effects when comparing eras? How big would that adjustment be?

That's a worthwhile topic, but I think it's too serious (and possibly political) to cover now.

187 Re: Walkthrough: Colt Dilemma

"But seriously, are you arguing that the 17-32 QBs in today's NFL are just as good as the top 16? Because if you aren't then it's indisputable that including them in the average brings the average down, and biases the analysis."

No, I hold it's more of a bell curve and the sharp drop off doesn't occur toward the bottom (say bottom 5 or 6) if it occurs at all. I'm not sure there's massive difference between Freeman, Flacco, and Cassel and guys like Eli Manning, Bradford, Cutler, and Mark Sanchez.

Look at the chart for passer rating leaders from last year. There's just not a huge variance in skill level over the bottom half of the league. Eli Manning was 17th. You are telling me that an Eli Manning comp in 1958 wouldn't have a job in the NFL? I'm sorry but I don't accept that.

http://www.pro-football-reference.com/years/2010/passing.htm

The distribution of talent in the NFL isn't linear. There's a much bigger middle class of QBs now than there were in 1958. Taking the average performance is a much truer way of providing a baseline than taking the top 16 in each era.

If you believe Eli Manning would be a backup QB in 1960 because he's the 17th best QB in 2010, then fine. I completely disagree. The worst QB in 1960 is much more comparable to the worst QB in 2010 than he is to the 16th best QB in 2010.

190 Re: Walkthrough: Colt Dilemma

I was in the middle of writing a long post disagreeing with you. Then I looked up the census data. The population of the United States has nearly doubled since 1960. Throw in segregation, the AFL, the vastly increased remuneration of modern players (making it more likely that those with the ability will choose to pursue football as a career), the increased popularity of football over other sports . . .

If anything, I suspect that using league average is slightly over-generous to Unitas.

198 Re: Walkthrough: Colt Dilemma

The population is up about 50% from 1970 to 2010. Comparing the start of Unitas' period to the end of Manning's is wrong. It's an easy mistake to make. You initial impression was closer to correct.

Segregation - in theory - reduced the pool by about 10%. Unless you want to argue black genetic superiority, which I assume you don't.

AFL and competing sports is more interesting. I think by 1960 NFL was the premier football league, and that if you were offered equivalent positions in both leagues, you went to the NFL. I am not aware of any NFL-level QBs who left for a career in baseball, but I have no idea of the situation at the college level. In 1960, the AFL's top QB was Frank Tripucka, who, after a promising start, had ended up as a washout in the NFL, went to the CFL, and joined the AFL at its founding. The second best - Jack Kemp - had already failed to advance beyond the taxi-squad on 4 NFL teams and had been cut from a CFL team. The third best - Al Dorow - had actually started several seasons, but like the others washed out and headed to the CFL.

Yup, it's pretty clear that the AFL was where you ended up if you didn't make it the NFL. Granted, some AFL QBs ended up as pretty good later, and a few were lured into the AFL by big contracts (e.g. Joe Namath) But the effect on the quality of NFL QBs was not large. The AFL was always the smaller league and always the weaker one.

201 Re: Walkthrough: Colt Dilemma

"The population is up about 50% from 1970 to 2010. Comparing the start of Unitas' period to the end of Manning's is wrong. It's an easy mistake to make. You initial impression was closer to correct."

Why would influx of population toward the end of his career affect the pool of available QBs?

No, the correct figures to use are the ones toward the beginning of his career. Unless you are arguing that his competition at QB was affected by the influx of new immigrants or from children born after 1950, the population of the US around the time Unitas entered the league is the more correct standard.

You just don't want to use it because it nicely syncs between a doubled US population and a doubled size of the NFL.

The population growth after Unitas started his career didn't affect his competition at QB at all, of course neither did the growth after the start of Manning's career. There's still just shy of double the population.

1950-151 million
2000-281 million

203 Re: Walkthrough: Colt Dilemma

I agree with you, I'm just going to use a slightly different method to show that agreement.

I think better numbers to use would be the average population centered around the year of their birth. That is where the majority of the competition pool would come from. Yes you could expand that range because they both had longer careers than that, but an 11 year average is close enough, someone who doesn't agree with us can feel free to find the youngest born QB at the end of Unitas's career and run the years out to that birth year and also the oldest QB when he started and the run the start down to that birth year, I'm satisfied with these. Unitas was born in 1933, Manning in 1976. The numbers of the 5 year cloud around those dates works out pretty close to the birth year numbers (which is expected). You should also look at ratio of male to female in the total population since that has changed a bit, but we are being relatively rough with the numbers here anyway, especially since all non census years are estimates anyway.

1928 - 1938 average is 125,467,143, the 1933 population was 125,578,763.
1971 - 1981 average is 218,354,330, the 1976 population was 218,035,164.
(numbers taken from http://www.demographia.com/db-uspop1900.htm but those seem to match with other sources on a quick glance)

So that is 1.74 times more people that would be in the pool to compete with Manning.

Segregation played a part as well. In 1933, 88.7% of the population was white so that would reduce the Unitas pool from 125,467,143 to 111,289,356.1. That changes the ratio to 1.96. Yes, the first black player was drafted in the 19th round in 1949 but the effects of segregation, on the QB position lasted much longer. I believe that Marlin Briscoe was the first black NFL quarterback, starting a game in 1968 so at the tail end of Unitas's career. It seems pretty safe to use that 1.96 number.

So if you assume a normal distribution of talent in the population it's pretty safe to say that a segregated 16 team league, made up of players born between 28 and 38 (the league that Unitas played much of his career in) would be fair to compare to a non segregated league made up of players born between 1971 and 1981 (the league that Manning has played much of his career in).

As mentioned I think there are other effects that actually make the "average" QB now better than the average back then. But I'll ignore that. I think the worry about "grade inflation" is pretty well covered by population growth and integration at least when looking at Unitas vs Manning. Era adjustment by comparison to average seems good to me.

207 Re: Walkthrough: Colt Dilemma

We seem to part ways on this: I believe the relevant comparison is to the same number of top QBs in each era. You believe that to be unfair to recent QBs, because their top QB peers are now taken from a larger pool of appropriately aged men. I consider using more players in the baseline to be grade inflation, essentially letting Manning be compared to the 16th best QB, while Unitas has to go up against the 8th best. Reasonable minds can disagree, it seems.

Statistical facts: Manning is farther above league average of the era. Unitas stands out more among the top ten (or 12 or 16 or 8...) QBs of the era. Both are valid and clear definitions of what it means to be "better" when adjusted for the conditions of each era.

I still go with Unitas. That seems to be because I want to compare great QBs to the top N quarterbacks in their years. I can understand and respect someone who goes with Manning, preferring to compare to a league average in a larger league in a more populous country. At least we're all looking at the right kind of stats, and avoiding the "myths vs. fawning vs. unadjusted season cumulative stats vs. win counts vs. greatest games" garbage that started this discussion.

Regardless, it's a very near thing either way you look at it. I said earlier that being second to Johnny Unitas isn't a bad thing. The same goes for being second to Peyton Manning.

210 Re: Walkthrough: Colt Dilemma

There's something to be said for either method of comparison, though I favor that of DPF. Due to the ends of a typical bell curve, it may not be statistically sound to expect #16 of 2X population to be as good as #8 of X population. However, I'd also think it unsound to expect #8 of 2X to be no better than #8 of X. Surely in that near-doubled population, one in which a higher proportion are likely to be seriously interested in football (my assumption, based on the huge increase in NFL popularilty and salaries), one may expect more than 8 QBs as good or better than #8 from the smaller group.

However, I wholly agree with that final sentence - it's a good summary of this whole thread.

216 Re: Walkthrough: Colt Dilemma

Actually, it'd be pretty interesting to JUST compare Manning to the top 5 in the league and Unitas to his top five.

I think that would actually be much more fair to Manning than using Unitas's 16 (which included a lot of real dogs).

I'll have to get to work on that.

217 Re: Walkthrough: Colt Dilemma

Actually, it'd be pretty interesting to JUST compare Manning to the top 5 in the league and Unitas to his top five.

I think that would actually be much more fair to Manning than using Unitas's 16 (which included a lot of real dogs).

I'll have to get to work on that.

223 Re: Walkthrough: Colt Dilemma

I'm currently going through year by year to compile the stats for the 5 best quarterbacks other than Unitas and Manning for each year of their career in each of the following stats.

Comp%
TD%
INT%
Y/A
Rating

I'm then going to average those numbers together over the course of their careers and compare the resulting averages to the career numbers for each player.

If anyone has any objections or corrections to that methodology, please speak now.

225 Re: Walkthrough: Colt Dilemma

I'm going to write this up on my site (18to88.com), but here's what I did.

I took the stats (NOT the percentages) for the top 5 QBs not named Unitas in the league in each category every year from 1956-1973. Starting in 1966 (the Super Bowl era), I included AFL quarterbacks as well (didn't affect much).

I then did the same thing from 1998-2010 taking Manning's own stats out of the equation. That means these averages are often 'top 6' in the league, but we obviously don't want the subjects polluting the data.

Comp %: Average Top 5 for Unitas: 56.7%
Unitas career: 54.6%
For career he was within 96.3% of top 5

Average top 5 for Manning: 65.5%
Manning career: 64.9%
For career he was within 99.1% of the top 5
Advantage Manning

TD %: Unitas: 7.0% to 5.6% (80% of top 5)
Manning: 6.2% to 5.5% (88.7% of top 5)
Advantage Manning

Int %: Unitas: 3.8% to 4.9% (128.9% over top 5)
Manning: 1.7% to 2.7% (158.8% over top 5)
Advantage Unitas

YPA: Unitas 8.3 to 7.8 (93.97% of top 5)
Manning 8.1 to 7.6 (93.8% of top 5)
Advantage Unitas (but essentially tied)

Passer rating: Unitas 84.2 to 78.2 (92.9% of top 5)
Manning 105.2 to 94.9 (90.2% of top 5)
Advantage Unitas

That looks like a pretty split verdict to me. Manning was better compared to the best of the best in completion % and TD%. Unitas was better in interception % and passer rating, and they were basically tied in YPA (though you can give the tie breaker to Unitas on decimal points if it makes you feel better).

The rate stat adjustments just don't help Unitas nearly as much was claimed. Against the league as a whole, Manning was clearly better, and against top 5 competition they split.

When you take that data and add the weight of many more productive seasons by Manning (not to mention the massive difference in volume stats beyond just playing extra games), and I think Tanier's position looks pretty well validated.

226 Re: Walkthrough: Colt Dilemma

Sorry, caught a spreadsheet error. (I had a QB throwing for 40,000 yards in a season).

The average passer rating for the top 5 in Manning's era is 99.2. That means that Manning is within 95.7% of the Top five for his career vaulting him ahead of Unitas.

227 Re: Walkthrough: Colt Dilemma

Hum. I sense a pattern here. You run a Manning fan site, that's named after him, even. I wonder whether your analysis will be balanced....

Did you....
Include Unitas' injury year in the total?
Did you add non-NFL stats into the mix to make Unitas look weaker?
Did compare Unitas' complete career, including 'victory lap' year to Manning's early and peak years?
Did you include stats while not a Colt? (Interesting choice in this context)

It's obvious why you kept these particular method choices obfuscated when you asked for comments and objections.

Try again. Compare NFL stats for the same number of years as the Colt's starter. Or just be a fanboy.

It's your choice.

229 Re: Walkthrough: Colt Dilemma

1. What difference does it make if I include Unitas's injury year? I used his career stats. Do you want me to pick and chose what years of his to include now?

2. I used AFL players from Super Bowl era simply because it makes sense. Frankly, I could have used them from the whole AFL era. It did very very slightly raise the averages, but not enough to change any of the outcomes. I'm surprised you would object to comparing Unitas to Len Dawson who is in the Hall of Fame. Their net effect on the data was negligible.

3. If you want to play the 'peak years' game, then tell me which years to chose. That's a different argument. It won't help Unitas though. Manning's peak years crush his in this category. I'd be happy to chose any set of years you want. When you take Manning's 1998 and even his 2001 off the board, his numbers skyrocket. If you want Unitas's 10 best years verses Manning's 10 best, I'd be happy to give them to you.

Tell me what will satisfy you.

4. He played one year in San Diego and played in 4 games threw 76 passes. You are correct, however, that season is not germane to the debate. Adjusting for those 76 throws means the following:
Comp % 54.7 (96.4%-still below Manning)
TD % 5.6 (actually lower)
INT % 4.8% (126.3% to top 5)
YPA no change
Rating 78.8 (93.6% still below Manning)

You can nitpick the data all you want. Tell me the parameters you want, and Manning will still be ahead. I've been staring at the numbers all day, and there's no way to play with them to make Unitas better unless you start comparing unequal groupings like you did.

I'm not being fan boy. I spent all day running numbers. You want different numbers tell me which ones.

231 Re: Walkthrough: Colt Dilemma

The only comparison that might work for Unitas is the best 6 years (1957-59 and 1963-65). I would have already done it with a method I came up with to adjust for era and opponent, but I can't do it for pre-1960 without the gamelogs.

253 Re: J. Uniatas 1950s gamelogs

Write worods wrong sometines but not numbers, so thehse are good. Put 0s in front of single digits in dcomps and stuff so for code looks nicer. For exmaple, first game played Unitas 0 comps, 2 att, 0 yards.


1956
det 00 02 000 0 1
chi 09 19 131 1 1
bay 08 16 128 2 1
cle 05 14 021 0 0
det 17 30 314 0 1
los 18 24 293 3 1
san 18 30 179 0 2
los 14 29 147 1 1
san 11 16 124 1 1
was 10 18 161 1 1


1957
det 14 23 241 4 3
chi 17 26 184 2 1
bay 07 17 130 2 2
det 16 21 239 4 1
bay 16 31 188 2 2
pit 03 09 056 1 3
was 17 30 247 2 0
chi 11 23 245 1 0
san 16 25 230 1 0
los 18 30 271 3 1
san 23 37 296 1 2
los 14 29 223 1 2


1958
dte 23 43 250 2 1
chi 10 23 198 4 1
bay 16 35 238 1 1
det 11 17 221 1 0
was 08 15 183 2 0
bay 05 16 099 2 0
los 12 18 218 2 0
san 17 33 229 1 1
los 23 38 214 3 3
san 11 25 157 1 0

playofgame
nyg 26 40 361 1 1

1959

det 13 30 230 2 0
chi 17 38 221 3 3
det 13 25 257 3 2
chi 16 30 233 2 2
bay 19 29 206 3 0
cle 23 41 397 4 3
was 15 35 265 2 2
bay 19 33 324 3 0
san 10 19 141 2 1
los 14 24 242 2 1
san 21 36 273 3 0
los 13 27 110 3 0

playoff
nyg 18 29 265 2 0

264 Re: J. Uniatas 1950s gamelogs

Ok, here's my era & opponent adjusted numbers on Unitas and Manning. I'd love to go into detail, but I need to write and present this before I do that.

Reg. season career
Manning - 208 games, +206.1, +.99/game
Unitas - 207 games, +153.0, +.74/game

If you just did games started
Manning - 208 games, +206.1, +.99/game
Unitas - 185 games, +163.3, +.88/game

If you took their first 13 seasons
Manning - 208 games, +206.1, +.99/game
Unitas - 155 games, +171.9, +1.11/game

Playoffs
Manning - 19 games, +22.0, +1.16/game
Unitas - 9 games, +7.7, +.85/game

267 Re: J. Uniatas 1950s gamelogs

I wonder if you could apply the formula pfr uses for weighted career approximate value. Which is 100% of their best year, plus 95% of their second best year, and so on.

This rewards players for having a excellent peak and excelling for a long time.

270 Re: J. Uniatas 1950s gamelogs

no. habve random games bnut not full gamelogs.

Like 12-13-59 at Kezar Stadium, Starr go 20 of 25, 249 yards, 2 TD pass, 0 Int.

have others somewhere but not relally going to hel,p if you dont have all

233 Re: Walkthrough: Colt Dilemma

Here are the numbers run other ways:

1. 10 best seasons of each guy

Comp%: Unitas: 55.8% to 55.8% (100.0%)
Manning: 66.0% to 65.9% (99.8%)

TD%:Unitas: 7.0% to 6.4% (91.4%)
Manning: 6.2% to 5.9% (95.2%)

Int%: Unitas 4.1% to 4.3% (104.9%)
Manning 1.6% to 2.3% (144%)

YPA: Unitas 8.3 to 8.2% (98.8%)
Manning 8.2 to 7.8 (95.1%)

Rating: Unitas 85 to 86.1 (101.3%)
Manning 100.0 to 99.1 (99.1%)

So that's 3 wins for Unitas, 1 for Manning, and one virtual tie

2. Years as the primary starter for Colts
This is every year Unitas started at least 7 games for the Colts. That leaves off 1968, 1971-73. Leaving off '71 is a tough call because he played 13 games but started just 5. It's favorable to Unitas to lose it, so I will. That's 14 total seasons as a starter for Unitas compared with 13 for Manning. If you want to quibble I'll drop 1956 too (just let me know).

That results in:
Comp%: Unitas 56.2% to 54.9% (97.7 %)
Manning 65.5% to 64.9% (99.1%)

TD%: Unitas 7.0% to 5.9% (84.3%)
Manning 6.2% to 5.5% (88.7%)

INT%: Unitas 4.0% to 4.8% (120%)
Manning: 1.7% to 2.7% (158.8%)

YPA: Unitas 8.3 to 7.9 (95.2%)
Manning: 8.1 to 7.6 (93.8%)

Rating: Unitas 84.3 to 80.4 (95.4%)
Manning 99.2 to 94.9 (95.7)
So as starting QB, it's a virtual tie. Manning takes Comp% and TDs, Unitas gets Int% and YPA. Manning edges him rating, but it's basically even.

Again, all this shows is that there is no real era adjusted rate stat gap between Manning and Unitas. They are neck and neck and the question has to be decided on different grounds (quality seasons, volume, ect.).

Anything else you want me to run?

234 Re: Walkthrough: Colt Dilemma

Dropping 1956 (which makes it 13 years as a starter for both men) does this to Unitas:

Comp: 56.3 to 54.9 (97.5%)
TD%: 7.1 to 5.9 (83%)
INT%: 3.9 to 4.7 (120.5%)
YPA: 8.3 to 7.9 (95.2%)
Rating: 86.1 to 80.7 (93.7%)

That doesn't do anything but help Manning in the rating comparison.

If we are dropping a starter season by Unitas, then 1956 is clearly the one to chose. Unitas started 7 games and played in 12. It was his fewest attempts of the remaining years as well.

238 Re: Walkthrough: Colt Dilemma

The first at least appears to make an honest attempt to compare apples to apples. The second is honest but compares Manning's youngest and healthiest years (we agree he's still near his prime, right?) to Unitas both in young health and in late career post-injury decline. The aging Unitas was never the same after the injury year, matching his own career average in the major rate stats only 2 out of 20 times, despite moving to the weaker AFC with the merger in 1970.

So, Unitas young-and-healthy is a bit ahead of Manning young-and-healthy, while Manning young-and-healthy is ahead of Unitas old-and-hurting (who'd'a thunk it?).

Stick to the apples-to-apples comparisons and it sure looks like Unitas was a bit better, compared to the best QBs of the time. Bias the analysis to favor Manning, and sure, he comes out on top, too.

Look, I'm not some die hard Unitas fan with a Unitas website. I hardly ever worry about what people think about Unitas. In fact, I'm more concerned about Billy Kilmer, of all people.

I'm interested in the challenge of using statistics to compare two great QBs from different eras. To do that well, I have to compare era-adjusted rate stats in ways that don't depend on the size of the league. It's frustrating to deal with someone who skews every analysis choice to favor his guy.

If Manning plays into his decline to whatever age is currently equivalent being 39 in 1972, we can compare whole careers and talk about longevity, too. Until then, the best we can do is compare Manning and Unitas up to the same stage in their career. On that balanced and sensible basis, the two are quite close, with Unitas coming out on top.

240 Re: Walkthrough: Colt Dilemma

How does the second list reflect an 'old and hurting' Unitas? It's a list of his years as a full time starter! It compares 14 seasons of Unitas to 13 seasons of Manning. It doesn't get much more apples to apples than that!

I haven't skewed anything to 'favor my guy', and I can't see any instance where I have. You've offered no bench mark for how to do the comparison other than an oddball way of looking at league averages that no one else buys. I've tried to be as even-handed as I can. Frankly, I believe AFL numbers from the early 60s should be included, but I knew you'd object, so I left them out to suit you and you alone.

I find your manner insulting, to be frank. You act like I've tried to play games with this data, which is something I never do. I'm doing my best to make reasonable comparisons, and at no point have you specifically asked me for anything. You've only criticized my choices after the fact. That's BS.

Do you want a comparison of them that goes from age 23 to age 34? I'll do that. That would chop off Manning's rookie year and take Unitas through age 34. That way it's strictly apples to apples by age.

Your system is the only skewed thing in this entire debate. Go back and look at the league leader charts from the 50s, and then try to make your argument about how 'inflated' Manning's comparison verses league average is with a straight face. It's frankly embarrassing. The QB in the 50s bears almost no similarity to modern play beyond a couple of guys (Unitas being one). The league leader charts don't even go 16 deep, because so few guys qualified for inclusion in the rate stats leaders.

I think you are the one who isn't honestly looking at era adjustments.

The point of the debate was to show their quality as players for the Colts. The second list does that far better than the first does.

Beyond that, these comparisons don't encapsulate the totality of the debate. They show that on a few rate stats compared with league average Unitas has some small advantages, but they ignore all the other facets of their careers, especially the question of number of elite seasons. The fact is that Unitas doesn't have 10 elite seasons.

It can be argued that his absolute peak was slightly higher compared to league average than Manning's absolute peak, but then you have to deal with the fact that Manning's peak lasted 5-6 more seasons than Unitas's did. That's why when the comparison stretches out to 11-13 seasons, Manning passes Unitas in the rate stats compared to the top 5 players. He was elite far far longer.

That cannot and should not be adjusted away. It's a shame Unitas got hurt, but HE DID GET HURT. You can't adjust away full seasons.

242 Re: Walkthrough: Colt Dilemma

From ages 23-34 for both QBs:

Comp% Unitas: 56.0% to 55.2% (99.1%)
Manning: 65.8% to 65.7% (99.8%)

TD%: Unitas: 7.1% to 6.2% (87.3%)
Manning: 6.1% to 5.6% (91.2%)

Int%: Unitas 4.2% to 4.6% (109.5%)
Manning 1.7% to 2.6% (152.9%)

YPA: Unitas: 8.4 to 8.1 (96.4%)
Manning: 8.1 to 7.7 (95.1%)

Rating: Unitas: 83.9 to 82.9 (98.8%)
Manning: 99.1 to 96.9 (97.8%)

Unitas gets Manning using this standard in three of the five metrics.

243 Re: Walkthrough: Colt Dilemma

By the way, for just this comparison, taking the AFL stats out doesn't flip any of the results. It brings Unitas closer in TD%.

The overall effect of using AFL players on this study is negligible. However, because the entire debate centers on the weird notion that the modern NFL is too watered down to use league averages (a stance no one agrees with), it seems strange to me that you would want to over Unitas by excluding a large sample of quarterbacks.

I'm steadfast in my belief that to not include AFL stats from the Super Bowl era would be huge mistake and would not accomplish your so called goal of exploring ways to compare eras.

I don't hold that position because it changes the results. It does not. I hold that position because I believe the QB play in the NFL in that era was not deep or strong. The more quality players that can be brought into the comparison, the better.

I do firmly believe QB play today is vastly superior to the level of play in the 50s and 60s. After doing these studies and seeing how thin the play was league wide, I actually believe in the entire concept of era adjustments before the Super Bowl era much less than I did before.

Combing through the league leaders from the 50s paints a nasty picture of the league.

239 Re: Walkthrough: Colt Dilemma

Regarding the AFL...

The AFL was mostly populated with players who failed in the NFL, or would not have started in the NFL. It's quite interesting to read the wikipedia bios of the most famous AFL QBs, for example.

Len Dawson: After encountering similar problems in battling Browns' quarterback Milt Plum, Dawson was released, having completed only 21 passes for 204 yards and two touchdowns in his five seasons of NFL play.

Jack Kemp: After being drafted by the Detroit Lions in the 17th round of the 1957 NFL Draft, Kemp was cut from the team before the 1957 NFL season began. He spent 1957 with the Pittsburgh Steelers and 1958 on the taxi squads of the San Francisco 49ers and New York Giants. The 1958 New York Giants played in that year's NFL Championship Game, known as the "Greatest Game Ever Played" and the first overtime NFL playoff game,but, as a third-string quarterback member of the taxi squad, Kemp did not take the field.

Daryle Lamonica: He was drafted by the Buffalo Bills in the 24th round of the 1963 AFL draft. He was also drafted by the Green Bay Packers in the 12th round of the 1963 NFL draft. Lamonica played with Buffalo for four seasons, backing up Jack Kemp on a team that won back-to-back AFL championships in 1964 and 1965.

George Blanda: It would not be until 1953 that Blanda would emerge as the Bears' top signal caller, but an injury the following year effectively ended his first-string status. For the next four years, he was used mostly in a kicking capacity.

Other top QBs left the NFL for the CFL, choose the AFL because it was clear they would not start in the NFL, etc.

There are a few, notably Joe Namath and Bob Griese who would have been considered top prospects in the NFL at the start of their AFL careers. But even they were playing in a league of also-rans, throwing against also-ran defenses.

That makes their stats "oranges" to compare with the NFL's "apples". It's like comparing college stats to pro stats. Sure, there are great college QBs, but the stats just aren't something you can usefully combine.

241 Re: Walkthrough: Colt Dilemma

I absolutely reject your reasoning here. That may have been true of the AFL in the early years, but by 1966 the league was far more competitive. Including AFL stats for the Super Bowl era is completely appropriate, and your efforts to exclude them seriously call into question your objectivity.

It's absolutely credible to include AFL players in the list. When you look at the lists of NFL leaders, you realize how flimsy your base argument is to begin with.

244 Re: Walkthrough: Colt Dilemma

This post kind of lost me when it quoted Wiki in disparaging Len Dawson, because his initial NFL non-success sounds so much like Wiki's description of Unitas' experience with the Steelers. And while Dawson couldn't do all that much in SB1 (though 16-of-27 for 210 yd and one pick against the Pack is decent), he was MVP of SB4 as the Chiefs upset the Vikings. (IMO, the Chiefs' defense was the real MVP though Dawson completed 71% of his modest 17 attempts, but my point is that he was a good NFL QB in the SB era, thus a silly choice to use in knocking the AFL.)

255 Re: Walkthrough: Colt Dilemma

It's not that Dawson was good or bad. (Hint: he was good, although his stats naturally dropped as he aged and faced better competition once the Commom Draft started)

It's that the NFL (two coaches, plus all the ones who saw him play and didn't snap him up when he was released) thought he wasn't good. With a few exceptions (e.g. Namath), quarterbacks who were thought to be the best went to the NFL with the AFL getting the leftovers, until the merger agreement and start of the Common Draft in 1967.

From 1967 to 1969, good players were drafted into both leagues equally. That's when Bob Griese and Larry Czonka went to the Dolphins, for example. In 1969, the AFL had three years of draftees of NFL quality, making the talent imbalance less than it had been in 1966 and earlier. It still wasn't up to NFL standards, but it had younger players of equal perceived talent and was getting closer overall.

Meanwhile, Unitas was cut by a Steelers coach who never let him take a snap, even in practice. I can only assume he was deemed "uncoachable". He landed a backup job the next season, and earned the starting job on the field when the Colt's starter broke his leg. He didn't have to leave for the AFL after five years, because when he was given the chance to start he proved to be good and quickly developed into a great QB.

Does anyone have an example of a QB washing out of the AFL and then landing a starting QB job in the NFL? Anyone?

256 Re: Walkthrough: Colt Dilemma

Pete Liske
http://www.pro-football-reference.com/players/L/LiskPe00.htm

Washed out of the AFL, went to Canada. Wound up starting for the Eagles in 71.

Also note, that he was originally selected by the Eagles, but chose to play in the AFL.

258 Re: Walkthrough: Colt Dilemma

Cool. What a strange career. AFL washout, CFL all-star and passing record setter, AFL-AFC part-time starter, then the NFC part-time starter. In his AFC/NFC comeback, he started roughly half the games for his teams.

Not exactly an AFL washout that moved to start in the NFL, since he got his second chance in the AFL, and only moved to the NFC after proving himself viable at Denver. But an interesting find nonetheless. Well done. Thanks.

How the heck did you find him?

260 Re: Walkthrough: Colt Dilemma

I did the research. I went through the careers of all the QBs of the 60s and examined what they did. There were actually very few cross overs. What was interesting was that his career showed that the AFL was better than Canada, and that a player drafted by both leagues chose the AFL.

I appreciate the props. I've been taking this seriously on principle, and did resent your implication that I've been playing games with the numbers. I actually don't believe in ranking QBs at all (I'm a 'tier' guy). I just have a personality defect and have to research points I find interesting. My agenda hasn't been to prove Manning better than Unitas, but I do admit to believing that the overall level of play in the 50s and 60s wasn't great.

Frankly, while I don't think it truly shows up statistically (because comp v league average is the best method), the single best (and undeniable) argument for Unitas is that he was essentially a modern QB in a time when most guys were flat awful.

261 Re: Walkthrough: Colt Dilemma

I did the research. I went through the careers of all the QBs of the 60s and examined what they did. There were actually very few cross overs. What was interesting was that his career showed that the AFL was better than Canada, and that a player drafted by both leagues chose the AFL.

I appreciate the props. I've been taking this seriously on principle, and did resent your implication that I've been playing games with the numbers. I actually don't believe in ranking QBs at all (I'm a 'tier' guy). I just have a personality defect and have to research points I find interesting. My agenda hasn't been to prove Manning better than Unitas, but I do admit to believing that the overall level of play in the 50s and 60s wasn't great.

Frankly, while I don't think it truly shows up statistically (because comp v league average is the best method), the single best (and undeniable) argument for Unitas is that he was essentially a modern QB in a time when most guys were flat awful.

232 Re: Walkthrough: Colt Dilemma

Agreed, and thanks again. One thing in the lengthy critique might be useful, though it wasn't stated quite the way I'm trying to.

Comparing Manning vs his "top 5" right now offers him the benefit of little/no decline phase. Would using just Unitas' 1st 13 seasons, beginning with his 1st as a fulltimer, make sense? Since your comparison uses just rate stats, the "injury season" would only be a factor if Unitas had a bad statistical year because he was playing hurt that year. Missing time wouldn't affect those rate stats any more than 2008 affects Brady's.

155 Re: Walkthrough: Colt Dilemma

Heck, Manning may not have been the best choice in 2008

Or 2009 (Brees).

180 Re: Walkthrough: Colt Dilemma

Anon,

Just curious but who exactly was more deserving in 2008? Consider that Manning's 08 team was nearly as flawed and injured as his 2010 team was and had to carry an awful run game and suspect defense to a 12-4 record despite starting the year injured. He was the best player in 08 and had the best season too.

In 09, the only argument you could make to merit brees would be his superior stats but even this is flawed since manning sat out the half the jets game and nearly all of the bills game. Brees sat out one game in tampa, but prorating their stats, manning would've finished ahead of brees in nearly every meaningful stat. He deserved it in 09 as well, especially considering the saints near dominant run game and manning's paltry one.

214 Re: Walkthrough: Colt Dilemma

In 2008, I really felt that Chad Pennington was more deserving. He had a higher qb rating than Manning while leading the Dolphins to an 11-5 record and the AFC East crown one season after the Fish finished 1-15.

In 2009, Brees' qb rating was 111 or so. I don't think Manning's was that high.

219 Re: Walkthrough: Colt Dilemma

Pennington was so important to his team that they took him off the field for long stretches to run the wildcat.

His play was vastly overrated that year in terms of quality and impact.

157 Re: Walkthrough: Colt Dilemma

You know what the difference between Manning's 2010 season and his 2003-09 seasons is?

He didn't pull a couple of rabbits out of the hat last year like he did in the previous 7 seasons.

Against the Patriots, you would have expected him to throw the 24 yard GW TD pass rather than the game-ending INT. Make that stat adjustment and you're looking at a 420 yard day, 5 TDs, 2 INTs and a 17 point comeback in the 4th QT. Then against Dallas, you expect him to complete some passes (let's say 3/4 for 40 yards) to get them in FG position and win that game, rather than another INT.

Make that stat adjustment and you're looking at the following line:
454/682 (66.6%), 4764 yards, 6.99 YPA, 34 TDs, 15 INTs, 94.1 PR, 12-4 record

Maybe one long completion instead of an INT against the Eagles, and they win that game with a FG at the end too.

Now you wouldn't even blink at that line (except maybe the YPA) in conjunction with the seven seasons prior to it. Fits right in. But since those two drives didn't end well, he drops to 10-6 and a rating just under 92.

178 Re: Walkthrough: Colt Dilemma

Gettinh some Tommyknocker Butthead tonigt. Goimh.to drink.it read these new comments. Fun friday night of foogball and beer

179 Re: Walkthrough: Colt Dilemma

To guy who said p. Manning like ripekn. Yes it true. Manning plays all the games. Is luke gerigh. The Iron Clot.
Defintely gets mentioned. Very impressive to play all the time. Does get noted bywriters and broafcasters.

To guywho say to imagine if p. Manning had Beiber hair. Yes would be annoying becsusse many clots fsns would copy hairdo and would be embarrasing. Thank goodness manning likes Kentucky farmboy look insteasd of Jasom Bieber dork/teenyboopper rat's nest- Captain Kanggaroo mix

181 Re: Walkthrough: Colt Dilemma

Manning never played for the Colts, he plays for the "Colts". There is no connection other than a uniform between one of the great football franchises of all time, and the pretenders playing in that suburb of Chicago. Comparing Manning to Unitas is like comparing Tom Brady to Bart Starr, or Aaron Rodgers to John Elway. Interesting to compare I guess, but none of these people played for the same teams, and Manning doesn't belong on a list with Bert Jones, Unitas and Morral. He is simply the best Indianapolis QB ever.

247 Re: Walkthrough: Colt Dilemma

Baltimore lost their rights to bitching about the Colts when that carpetbagger Modell fled to Baltimore from Cleveland.

Or are you mad that would mean that now Otto Graham is Baltimore's best QB instead of Unitas?

189 Re: Walkthrough: Colt Dilemma

From 1961 -1963, Unitas was the third to fifth-best QB in a 14-team league....

Not on any planet that supports human life -- who was better? Starr? Maybe -- but even with a better supporting cast, he never put up the results Unitas did. Tarkenton, Meredith? Way too young if ever. Wade, Ryan, LeBaron? Please. Tittle? Jurgensen in Philly? Ed Brown? Milt Plum? Who was under center in SF or LA back then? That's far too flat a statement to accept on face value, Mike. And to make it a determining factor in your argument -- well I'm sorry.

My thoughts about the Colts line up with Jon's -- but that has nothing to do with this. If you're going to make a statement like that -- I'd like to know the two to four in those given years you think were better - -and why.

Oh, and Peyton needs to cut down on his supplements -- or Raiderjoe's nickname for him will become real...

191 Re: Walkthrough: Colt Dilemma

Well from 1960-63 Unitas had the following ranks among NFL QBs (no AFL):

Completions - 1st
Attempts - 1st
Comp. % - 5th
Yards - 1st
YPA - 6th
TDs - 2nd
Rating - 8th

That's out of 18 players with a qualified number of attempts. http://pfref.com/tiny/xiKcC

When one tries to make the argument that Unitas was twice as dominant as Manning, it doesn't serve that argument well when he has those numbers over a 4 year period. 8th out of 18? That's not to say guys ranked ahead of him were actually better, as Unitas was the volume passer of the bunch and had the most attempts by a wide margin, but clearly he wasn't dominating the league.

1999-2002 is not thought of as a dominant stretch for Manning, but here are his NFL ranks in those seasons:

Completions - 1st
Attempts - 1st
Comp. % - 2nd
Yards - 1st
YPA - 3rd
TDs - 1st
Rating - 4th

That's out of 30 qualified passers
http://pfref.com/tiny/lV3rO

Pretty good for someone that was a year away from hitting their prime.

199 Re: Walkthrough: Colt Dilemma

well those are all interesting -- but I notice you pick the outlier rating stat -- one almost universally derided and discredited -- as the focus for a response that didn't even answer the question I raised.

SO I'm not exactly sure why you attached it to my post -- unless you actually think there were SEVEN QB's superior to Unitas in that period, and would care to mention them ....

208 Re: Walkthrough: Colt Dilemma

Not sure how his rating could be an outlier when his other efficiency stats are not that elite. 7 players had a better TD:INT ratio.

Tittle, Starr and Jurgensen can definitely lay claim to being better than Unitas over that period of time.

195 Re: Walkthrough: Colt Dilemma

I think the one thing this column has proved is that the season can't start soon enough. We NEED NFL football.

236 Re: Walkthrough: Colt Dilemma

There is still a major international American football tournament going on, and the US is kicking butt!

http://www.americanfootball2011.com/

202 Re: Walkthrough: Colt Dilemma

Johnny Unitas was responsible for: "The Greatest Game Ever Played" in the NFL.

Peyton Manning was responsible for: "The Greatest Game In The History Of The New Orleans Saints." Which is nice, too. Even his father couldn't claim that.

211 Re: Walkthrough: Colt Dilemma

Sorry I lost the numbers here, but I did some research on Manning's numbers that put his greatness into a little perspective. Feel free to repeat it if you have the free time.

During the last game of the 2009 season, NBC (I think) put up a graphic of the statistical leaders for the decade. (Yes, we know, nobody cares.) Manning of course was way out in front in yards and TDs, something like 42K yards and 310-ish TDs. For those in the back, that's over 4200 yards and 31 TD per year for ten straight years. Sweet merciful crap.

Now, being a Browns fan, I of course wondered if they had ever, even once, had a quarterback hit either of those numbers. The answer - of course not, although Brian Sipe came close in 1980, but somehow forgot how to THROW THE BALL INTO THE FREAKING LAKE AND KICK THE FIELD GOAL!

Uh, sorry about that. Where was I? Oh yeah. Manning's averages over a 10-year span would have been a franchise record for the Browns in both yards and touchdowns. Wow. I mean, just, wow.

So that of course got me to wondering how many other franchise records his 10-year average would have. And the answer is.... lost somewhere in my notes. But I'm sure someone with spare time can figure it out for you! It's a lot, though.

Anyway, that's a little something to put Manning's greatness in perspective. What he accomplishes year after year after year, many franchises have never seen even once.

212 Re: Walkthrough: Colt Dilemma

Even if you just dropped those numbers down to more common milestones like 4000 yards and 30 TDs, that is the kind of season that rarely happens.

http://pfref.com/tiny/q31cI - 38 times by 20 QBs, and Manning has 6 of them. Granted, you're not going to put those numbers up without 16 games, so it's more of a "since 1978" thing, but still. 6 times for something that only 7 other QBs can say they've done more than once. And that could easily be 7 times without all the rest in 2005.

221 Re: Walkthrough: Colt Dilemma

Manning is one of the all time greats, but it does get a bit silly when you compare across eras using unadjusted stats, and bulk numbers are even worse. As a device for pointing out just how big Manning's stats have been, that's okay, but it goes awry when used as a proxy for a good season. Specifically you point to Cleveland as never having any QB even have 1 season as impressive as the average of Manning's last 10, but Cleveland has arguably (and in my opinion) the GOAT in Otto Graham, who had a pretty long career himself.

215 Re: Walkthrough: Colt Dilemma

Wow. There's a lot of discussion about this. Some really great stuff.

I've gotta agree with Mike. Johnny U was a not-doubt-about-it HOFer and the best of his era, but Peyton has had a longer run of qb dominance.

235 World Championship of American Football

Come on guys. Our national team just beat Mexico 17-7 and we are going to the final to be potentially "the best American football team in the world. Yet Football Outsiders continues to ignore this.

http://www.americanfootball2011.com/

USA!