Saturday, December 09, 2006

Troy Smith vs. Chris Leak

It's a foregone conclusion that Troy Smith is going to win the Heisman Trophy tonight. Reading many predictions about what to expect on January 8th I decided to simply put Smith's stats up next to Chris Leak's. No in-depth analysis or extrapolation necessary here just a look at the obvious numbers.

Troy Smith
199-297, 67%

Chis Leak
207-329, 62.9%

What we see here is that we have two pretty accurate passers facing off against each other. Leak has completed and attempted more passes (He played one more game this season) than Troy Smith but has a lower completion percentage.

Troy Smith
2507 Passing Yards, 8.4 YPA

Chis Leak
2729 Passing Yards, 8.3 YPA

Again, pretty close.

Troy Smith
30 Passing TD, 5 Int

Chis Leak
22 Passing TD, 13 Int

Here we see the real differences between these two quarterbacks. Leak occasionally has a brain lock that results in a disastrous interception.

Troy Smith
1 Rushing TD

Chis Leak
3 Rushing TD

This is surprising considering that Leak is not a running QB. Smith isn't either but he is certainly considered more athletic.

Troy Smith
167.9 QB Rating

Chis Leak
146.8 QB Rating

Here we see the how the interceptions and lower completion percentage affect Leak's overall rating. Smith is 6th in Division 1A while Leak is 33rd.

My conclusion is that Troy Smith obviously has the better resume beating Leak in every important statistical category, but not by that much. OSU gets the nod in the starting QB match-up, but Florida always has a wildcard named Tim Tebow. Tebow has been used sparingly but in crucial situations.

I can't wait until January 8th.

6 comments:

Anonymous said...

This is all interesting, but the way to compare QBs is to watch them play.

Troy makes amazing plays, and his mistakes are remarkably rare. A similar analysis might suggest that Troy and Brady Quinn are roughly equal, or perhaps a slight edge would go to Quinn.

But how did they play in the biggest games? Did you see Smith dissect Michigan's defense, which had been destroying everything in it's path?

There is a reason that some Buckeye fans are thinking that Troy Smith is among the greatest players to ever play for OSU, and it isn't "passer efficiency".

EW

Henry Louis Gomez said...

EW,

This site is about mainly statistical analysis. If it's not quantifiable then it might as well be voodoo. Good players are good statistically. Great players are great statistically. Troy Smith is one of the top QBs in the game because his statistics say its so. Chris Leak is also a QB that about 85-95 programs would like to have since he ranks better than their QBs.

Anonymous said...

"This site is about mainly statistical analysis. If it's not quantifiable then it might as well be voodoo."

I guess you can do what you want with your site, but I have a hard time believing that you believe the second statement. I'm guessing Tressel and Meyer are busy these days with the voodoo of studying game film, gleaning non-quantifiable bits of information for use in preparing game plans.

Is it not obvious that the statistics available provide very little information about what actually happens on the field?

I understand a desire for objectivity, but since we are not trying to publish in an academic journal, there seems to be no need to jettison all non-quantifyable-but-useful observations.

EW

Henry Louis Gomez said...

Wow,

We couldn't disagree more. First of all looking at film is studying tendencies. That's an empirical process of a more or less scientific nature. Not like the typical fan that watches a game on TV following only what the cameraman shows him (the ball carrier).

What Troy Smith did to Michigan is tangible and can be analyzed. But one game does not a season make. I mean is that the real Troy Smith or is it the one that was limited to 108 yards and no TDs against Illinois. Probably neither. The stats for the season give better portrait of the complete Troy Smith than any one game. (And don't think the coaches at Florida aren't looking that Illinois tape too).

I could put together a highlight reel of Chris Leak that would have you convinced that he's every bit the player that Troy Smith is.

As far as what "we are trying to do" I think that's up to us. For the record what we ARE trying to do is look at things a little deeper than the average casual observer. You may not agree with the conclusions and that's fine but we're never going to back our arguments with "he always come through in the clutch" or some other non-quantifiable descriptor.

Read the book Money Ball and tell me if it changes your mind about the value of statistical analysis as a tool for predicting success. A tool that is not 100% accurate (that's why they play the games, after all) but that can be quite effective when used correctly.

Anonymous said...

What you are doing is different than the average casual observer, but I'm not convinced that it's deeper.

The following is my idea of deeper:

http://mgoblog.blogspot.com/2006_11_01_mgoblog_archive.html

(find the posts titled "Interesting Plays From OSU-Texas" and " Interesting Plays From OSU-Iowa".)

There you will see analysis of aspects of offense and defense that matter.

If you would want to make some observations about how Florida could do what Illinois did on defense, that would be interesting. (Though part of the plan would involve convincing Tressel to call about five passing plays in the second half.)

If what you do seems helpful to you, great. But it doesn't seem to give the slightest idea of why Smith may be better, and what types of plays Florida needs to be concerned about. (How often does he throw deep? What percentage of passes were on rollouts? How many of the incompletions were smart throwaways? What type of plays are run in the red zone?)

There is something possible in between trying to be a computer and relying on ideas like "he comes through in the clutch."

EW

Henry Louis Gomez said...

Different strokes for different folks.