Writers of Pro Football Prospectus 2008

03 Jan 2012

ESPN AGS: Chiefs over Broncos

Has the magic run out for Tim Tebow? There actually are some promising names on a list of QB seasons by similarity scores... until you realize all those names played 30 years ago and the game has changed an awful lot since then.

Posted by: Vincent Verhei on 03 Jan 2012

15 comments, Last at 07 Nov 2012, 12:16pm by click here

Comments

2
by Zheng :: Tue, 01/03/2012 - 8:58pm

Why is Tebow so inaccurate? Was he that inaccurate in college? Help me out here, I don't follow college ball.

5
by zlionsfan :: Wed, 01/04/2012 - 12:14pm

No, he was actually pretty accurate in college: from 64.4% as a junior to 67.8% as a senior, 66.4% for his career.

Now, that was with fewer attempts than some other well-known college QBs. He threw 298 passes as a junior, 314 as a senior, 350 as a sophomore. Compare that to Blaine Gabbert (445 and 475 in two years as a starter), Sam Bradford (341 and 483), and Andrew Luck (288, 372, and 373) ... but also Cam Newton (280), Denard Robinson (237 and 291), and Colin Kaepernick (247, 282, 359, and 383). Note that the latter group all had rushing as a significant aspect of their game, and the former group did not.

I don't know exactly why he's struggling so much in the NFL. Newton had basically the same completion percentage last year, and his NFL completion percentage is almost 15 points higher than Tebow's. Tebow's Comp%+ is 52 this season, 66 last season. To put that in perspective, Tony Banks was last among qualified QBs in 2001 in completion percentage, and his index was 81. In 1991, Vinny Testaverde posted a 78.

p-f-r's Index stats are based on a three-year average (one in each direction plus the current season, so only a two-year average for the current season). Every 15 points away from 100 represents one standard deviation ... which means Tebow's completion percentage is more than 3 SDs worse than average!

If you look at QBs with 200 or more attempts and a Comp%+ of 55 of lower, you get Akili Smith's 2000 season (44!), plus four guys from 53 to 55 (Gary Marangi, Mike McMahon, Ryan Leaf, and Rusty Hilger). Tebow's ANY/A+ is significantly higher than theirs, thanks to his TD percentage, but the rest of his indexes are comparable to theirs. Any QB can have a bad season, but if this is the type of year he's going to have in a hybrid system (one designed for a different type of QB but adjusted to fit him) ... well, he's going to have to get better.

7
by Eddo :: Wed, 01/04/2012 - 12:56pm

One argument for his high completion percentages in college is that the defenses were so concerned with Tebow's running ability, that they would peel off their receivers as he scrambled, leaving them wide open. That sounds like a legitimate argument, but I really didn't see enough Florida games to know if it's an accurate assessment.

------

Your list of other collegiate QBs intrigued me, so I pulled some stats for them. (I limited this to seasons in which they had 100+ attempts.)

Player ....| Year | School | Att | Comp%
----------------------------------------
Gabbert ...| 2009 | Mizzou | 445 | 58.9%
Gabbert ...| 2010 | Mizzou | 475 | 63.4%
Gabbert ...| 2011 | -NFL - | 413 | 50.8%
----------------------------------------
Bradford ..| 2007 | U.Okla | 341 | 69.5%
Bradford ..| 2008 | U.Okla | 328 | 67.9%
Bradford ..| 2010 | -NFL - | 590 | 60.0%
Bradford ..| 2011 | -NFL - | 191 | 53.5%
----------------------------------------
Luck ......| 2009 | Stanf. | 288 | 56.3%
Luck ......| 2010 | Stanf. | 372 | 70.7%
Luck ......| 2011 | Stanf. | 373 | 70.0%
----------------------------------------
Newton ....| 2010 | Auburn | 280 | 66.1%
Newton ....| 2011 | -NFL - | 517 | 60.0%
----------------------------------------
Robinson ..| 2010 | U.Mich | 291 | 62.5%
Robinson ..| 2011 | U.Mich | 237 | 56.1%
----------------------------------------
Kaepernick | 2007 | Nevada | 247 | 53.8%
Kaepernick | 2008 | Nevada | 383 | 54.3%
Kaepernick | 2009 | Nevada | 282 | 58.9%
Kaepernick | 2010 | Nevada | 359 | 64.9%
----------------------------------------
Tebow .....| 2007 | U.Flor | 350 | 66.9%
Tebow .....| 2008 | U.Flor | 298 | 64.4%
Tebow .....| 2009 | U.Flor | 314 | 67.8%
Tebow .....| 2011 | -NFL - | 271 | 46.5%

It's obviously common for completion percentages to drop once a QB reaches the NFL, so there's no surprise Tebow's not completing over 60% of passes any more. There's definitely something else happening, though.

Personally, I watch Newton play and Tebow play and it's night and day in terms of accuracy and throwing motion. Newton looks much more natural at QB, whereas Tebow looks like the classic "athlete playing quarterback". I see shades of Tebow when I watch Denard Robinson play, as well; he looks to run first, it seems, whenever I watch him, then go through his reads. I recall Newton (and Robert Griffin III this year) playing a much more thrower-then-runner role, as opposed to Tebow and Robinson's runner-then-thrower.

Of course, I'm by no means an expert, particularly on the college game.

12
by beargoggles :: Thu, 01/05/2012 - 1:54am

I've now seen Denard Robinson play 1 college game (last night), and he made Tebow look like Cam Newton. But I guess he "just wins"

3
by lester bangs (not verified) :: Tue, 01/03/2012 - 11:27pm

Tebow made headlines for leading the Broncos to seven wins in his first eight starts, but in hindsight, that streak is looking much less spectacular. Only one of those victories came by more than one score, three came in overtime and none came against a team that made the playoffs.

Not that I'm promoting the Broncos as any great team, but "who did you beat?" is a weak game. The Patriots didn't beat a team with a winning record en route to going 13-3. Denver beat all of their division opponents on the road, which means something.

We've gotten to the point that if you didn't beat the Packers, Saints, Niners, Steelers, Ravens or Patriots, some people try to throw all of your wins in the trash. It's silly. Simmons was calling Atlanta fraudulent on his Monday podcast and when the Falcons win at Detroit was mentioned, he laughed it off. Hey, Detroit did go 10-6. This isn't a league with six legitimate teams and 26 punching bags.

4
by justanothersteve :: Wed, 01/04/2012 - 9:40am

There's a big difference in the "who did you beat" argument when you have a 13-3 record like the Pats or 49ers, vs an 8-8 record. 13-3 means you had a winning record on the road. 13-3 means you had a winning record against your conference opponents. 13-3 means you had at worst a 3-3 record in your division. Beating all your division opponents on the road when none of those teams had a winning record is good, but nothing special when you went 8-8. An 8-8 division winner means you played in a weak division, even if all the teams essentially had a .500 record. Beating all your division opponents on the road begs the question why they only went 3-3 in their division as it means they lost every division home game.

6
by zlionsfan :: Wed, 01/04/2012 - 12:38pm

There is something to strength-of-victory, even if there are teams that don't seem to fit the standard argument well.

Yes, Detroit was 10-6, but they didn't beat a team with a winning record either: their three wins over .500 teams came over Good Chicago (which was probably their best win of the season, given where Good Chicago was heading prior to Cutler's injury), Dallas, and Denver, and of course the Dallas win required a large comeback. In addition to the Atlanta loss, the Lions also lost to Green Bay and San Francisco at home, as well as the Packers and Saints on the road. That doesn't necessarily make Atlanta's win worse, but it's also not the same as if the Lions had beaten San Francisco and New Orleans and lost to Atlanta.

The Patriots didn't beat a team with a winning record, but they only played two of them (Pittsburgh and the Giants) ... and they beat the crap out of a number of weaker teams they played. That's part of the Guts/Stomps thing: we want to look at more than just win or loss, but how the game played out. The Patriots are a good team even though they didn't have many opportunities to prove it against their peers; the Lions are not as good, in part because they showed that against their peers.

I wouldn't put too much stock into what Simmons says. His audience doesn't overlap much with FO's, and where it does, we look to Simmons for entertainment more than analysis.

8
by tally :: Wed, 01/04/2012 - 1:39pm

It's not just a matter of who did you beat but also how you beat them. The fact that most of those victories against mediocre teams came down to the 4th quarter or overtime means that it could have gone either way, so eking out those victories against mediocre teams really indicates that Denver is a mediocre team. A great team would routinely trounce those teams that Denver was squeaking by.

And to the topic of selection bias, how about counting game-winning drives in the 1st or 2nd quarter rather than just the 4th quarter?

9
by Will Allen :: Wed, 01/04/2012 - 3:23pm

I understand what you are saying, but to me it has always seemed much, much, sillier to think of a 1 point victory, in a game as close as the score suggests, to be much more indicative of the strength of a team than 1 point loss of the same type. What informs me much more regarding a team is when they win big or lose big with some frequency.

What I know about the Broncos is that they can't pass the ball against a team withy any quality of pass defense. To think that one can be confident that a team with such poor passing ability can win 9 games, with the current NFL rulebook, seems extremely foolish. Now, maybe they have some reason to think that Tebow will make gigantic strides, and become an average passer. I have not seen much reason to think so, but then I've seen him attempt about 80 passes, so what do I know?

10
by JonFrum :: Wed, 01/04/2012 - 4:35pm

Then again, the Patriots two games against the Jets are a significant reason why they Jets were not a team with a winning record. And the Patriots win against the Broncos kept them from a winning record as well. Point taken, but you do have an effect on your opponent's record.

And by the way - "Denver beat all of their division opponents on the road" is called cherry picking. The fact is that Denver had a series of fluke wins - play those games over and you can't believe the results wouldn't be different. By the eyeball test, the Broncos were a 5-11 team that piled four years of luck into one season.

13
by JonFrum :: Thu, 01/05/2012 - 7:05pm

I'll reply to myself to add that I just read the Patriots beat seven teams this year that ended up with 8-8 records. So the Patriots kept them from being winning teams. Of course, if they LOST a game to the Jets, then they'd have beaten another winning team.

14
by NYMike :: Fri, 01/06/2012 - 12:41am

I'm more concerned that the only playoff team they beat was Denver. Lost to Pittsburgh and the Giants. And compiled a 12-1 record against mediocrity. By contrast, Baltimore beat six playoff teams, including Pittsburgh twice.

11
by Jim Glass (not verified) :: Wed, 01/04/2012 - 5:14pm

Not that I'm promoting the Broncos as any great team, but "who did you beat?" is a weak game. The Patriots didn't beat a team with a winning record en route to going 13-3.

But "how did you beat them" is a very strong game.

All wins count the same in the standings, but they *don't* count the same in determining how strong a team is. OT wins don't indicate the winner is any better than the loser -- they are a tie. In fact all close games, one score or less, have only the value of ties when determining team strength.

Close game outcomes are luck. Random chance. Vince Lombardi and Bart Starr were 50% in close games. Bill Walsh and Joe Montana, and Chuck Noll and Terry Bradshaw had *losing* career records in close games. So Denver fans who believed "Tebow knows how to win close games!" were making a very strong claim indeed, saying that he can do what they couldn't. (In fact, putting him in a class with Dick Jauron/Jim Miller, and Vince Tobin/Jake Plummer, who won a whole lot of close games while losing everything else in sight.)

Here's a quick, easy and effective way of judging how strong a team is: Figure it's W-L% counting all its one-score decisions as ties. Simple as that. It works.

In 2+score decisions this year New England was 8-0, Denver was 1-4. By "Big Win%", counting one-score games as ties and adjusting for strength of schedule, New England finished a .743 team and Denver a .399 team.

During Tebow's 11 starts Denver was 0-3 in 2+score games, counting the other eight as ties they were 4-7 against opponents of only an average .476 strength.

15
by click here (not verified) :: Wed, 11/07/2012 - 12:16pm

wonderful publish, very informative. I wonder why the opposite specialists of this
sector do not realize this. You must continue your writing.
I'm sure, you have a great readers' base already!