As you have probably heard (if you follow college football even passively) it is a big advantage to have extra time to prepare for GT’s offense.
What we do is pretty unique, but if a team has more than the standard 1 week to get ready, we tend to struggle.
In fact, many fans go so far as to say that is all a team needs to beat us.
So, how much truth is there to this claim.
Well, here is a look at Coach Johnson’s record at GT
| | Overall | Opponent has 1 week | Opponent has Extra Time | BCS Conference Opponent has Extra Time (Reg. Season only) |
| W/L Record | 33-16 | 22-5 | 11-11 | 7-8 |
| Win Percentage | 67.35% | 81.48% | 50% | 46.66% |
So, that settles it right?
When the opponent only has 1 week, we play much better. We win 81% of the time. When the opponent has more than one week to prepare, we win only 50% of the time.
Hmm... Well, maybe, maybe not.
Could it be just a coincidence? Is it possible that the games where opponents have extra time just so happen to have fallen against the better teams on our schedule?
I guess its possible.
So I decided to take a look. I compared the results of the games where opponents had extra time with the expected result of that game, according to Sagarin’s end of year ratings. For 2011 games, I used the most current Sagarin ratings.
Basically, I figure as follows. Sagarin’s poll simply compares all games played by all teams, based upon who won, who lost, and the margin. Your ranking is the summation of what you have done, weighted against how good your opponents are, as calculated by what they have done against their opponents, based upon how good they are, etc.
Sagarin does not factor in anything else. There is no adjustment to his poll for having extra prep time, or any other factor besides who you played and what the score was.
So, for example, our 2008 game against UNC was a bad loss by this metric. According to Sagarin’s poll, we should have lost by only 8, but instead we lost by 21. UNC had a bye week before this game, so this supports the argument that extra time allows the defense to “solve” our offense.
However, UGA in 2010 had also had a bye week. According to Sagarin, they were supposed to win by 12. They only won by 8. We still lost, but this was a better loss. Even though UGA had a bye week, we performed better than the computer expected us to perform, based upon our results for the rest of the season. That cuts against this extra prep time argument.
The 2008 UGA game is even worse. They had a bye week there, and according to Sagarin they also should have won by 12 (just like 2010... strange) but instead, they lost by 3. That really cuts against this extra time argument.
So, here is a look at all the data for all 22 games where an opponent had any kind of extra time (in this, I am including the first game of the season, as well potential “extra time” where a team plays an “Easy” game the week before us, and might use extra practice time that week to prepare for GT):
| Expected GT Won/Loss According to Sagarin | Actual Won/Loss | GT Record in Games Where Losing Team Was the Team That Sagarin Expected to Win |
| 10 Wins, 12 Losses | 11 Wins, 11 Losses | 4 Wins, 3 Losses |
Well, now we see something interesting.
According to Sagarin, we were supposed to be 10-12 in these games. So, maybe it is true that, just by chance, the tougher teams on our schedule have been the ones getting extra time.
We have actually performed slightly better in those games than we should have.
Here is a look at just the games against BCS conference opponents, where the “Extra Time” is literally extra time. I’ll use regular season only for this table, so we are looking at games where opponents have 9, 10, or 14 days instead of the usual 7 (the 30+ days for a bowl game would make this a little misleading I think, I have given the bowl games their own chart at the end).
| Expected GT Won/Loss According to Sagarin (Regular Season, BCS conference opponents Only) | Actual Won/Loss | GT Record in Games Where Losing Team Was Team Sagarin Expected to Win |
| 5 Wins, 10 Losses | 7 Wins, 8 Losses | 4 Wins, 2 Losses |
So, does the opponent having extra time actually help us??
Kidding. I don’t think that is possible. More likely, the fact that we have done slightly better in those games than we should have is a matter of mere chance. The difference is only 1 or 2 games.
And in fact, I suspect the extra time does give the opponent an edge. But I think its only a slight edge. Our offense is unique, and so having a little extra time logically should help. But I don’t think it helps very much.
In short, I think this is just another myth trumped up by Paul Johnson’s many detractors. They want to find something to use against this offense, because they don’t like it. This makes logical sense, and in fact our record is actually much worse (see the first table above). If you don’t bother to look any deeper, this seems to be true.
I think another big part of this myth is the effect of perception. A lot of things in college football are easy to perceive the way you want to.
A great example is the NC State game in 2010. At the time we were favored by 9 points. Then we lost the game by 17. Many ACC fans used that game as further proof to argue the effects of “extra time”. NC State played on Thursday the week before, so they had two extra days. This must have been why they managed to beat us by 17 when we were supposed to win by 9.
Of course, Sagarin’s poll at the end of the season, based upon how the two teams played the rest of the year, concluded that NC State should have beaten us by 13 points on the road (and 19 had the game been in Raleigh). They were actually the better team. Its just that, 3 games into the season, nobody knew that. And, most people did not take the time to go back and look either.
Here is a look at the average Sagarin rank of teams we have played, beaten, and lost to, in various circumstances during Coach Johnson’s tenure:
| | GT Wins | GT Losses |
| Average Sagarin Rank - No Extra Time | 61.01 (21 Wins) | 37.2 (5 losses) |
| Average Sagarin Rank - Opponent off a Bye week | 59.2 (5 wins) | 28.5 (6 losses - only 1 team ranked worse than 34th) |
So you can see, our record is actually fairly similar in terms of how good the teams are that we beat, on average, and how good the teams are that beat us, on average. Apparently, a higher percentage of the teams that get extra time before us just so happen to be good teams. Strange.
Additionally, 9 of the 11 losses we have suffered against teams that had extra time to prepare have come with a team off a bye week or in a bowl game.
Here is a look at those games:
| Average Sagarin Rank of GT Opponents that Have a Bye Week Before GT, or play us in a Bowl Game | Average Sagarin Rank of All GT Opponents that play GT with no Extra Time |
| 39.14 ( GT 5 wins, 9 losses) | 56.46 ( GT 21 wins, 5 losses) |
In summary, I would say that extra prep time probably helps our opponents, but the effect does not appear to be terribly significant. In fact, over a sample size of 22 games, the effect appears to be helping GT. That does not make any sense, but if the effect is small enough that after 22 such games, fluctuations due to random chance can make it appear to be working in the other direction, then I don’t think its something we should worry about too much.
The one exception to this may be the bowl games. Which does make logical sense. “Extra time” during the season just allows a few extra hours of repetitions. In season, teams still spend time practicing their base formations, and doing what they normally do. However, for a bowl game, you have 5 weeks to play 1 game, and then you get spring practice and preseason practice to prep for next year. So, in preparing for a bowl game, there is no reason for a team to waste any time practicing their base defense. Plus, 5 weeks is much more “extra time” than they would get for any other game.
So, Bowl Games could well be a unique case. And we certainly have not played well in bowl games. (understatement) Then again, a sample size of 3 does not prove a whole lot. And there are other factors that could explain our flops. We did not matchup well with LSU or Iowa (both had tremendous D-Lines while we had average at best O-Lines), plus there was reason to think we might have been a little flat against LSU. Everybody thought we would win easily. We had just beaten UGA, who beat LSU that year pretty easily, in Baton Rouge. And finally, we lost the Air Force game largely due to a bunch of special teams fumbles. I don’t know how “Extra time” could have affected that.
I would say this extra time phenomenon is largely a myth during the season, but may be a serious factor for the Bowl Game. Hard to say with any certainty concerning the bowl games with only 3 data points, but we have 19 regular season games to evaluate, and there not much support that extra time has a significant effect.
Below are charts showing all of the raw data, for every conceivable “extra time” scenario - 1) season opening game 2) opponent played a very weak team the week before 3) opponent had a bye the week before 4) opponent played a mid week game the week before 5) and the bowl game.
The charts show what result the end of season Sagarin ratings would expect in the game, what the actual result was, and the difference.
You can see some games make it look like extra time helps, some make it appear it hurts, but overall, there is not much of an effect.
| GT Opponent | Extra Time (Season Opener) | What Result “Should Have” Been According to End of Year Sagarin | Actual Result with Extra Preparation Time | Difference |
| 2008 Jax St. | Season Opener | GT by 24 | GT by 27 | GT +3 |
| 2009 Jax. St | Season Opener | GT by 25 | GT by 20 | GT -5 |
| 2010 SC State | Season Opener | GT by 21 | GT by 31 | GT +10 |
| 2011 Western Carolina | Season Opener | GT by 39 | GT by 42 | GT +3 |
| | | | | Total = GT +11 |
| GT Opponent | Extra Time (Bowl Game) | What Result “Should Have” Been According to End of Year Sagarin | Actual Result with Extra Preparation Time | Difference |
| 2008 LSU | Bowl Game | LSU by 8 | LSU by 35 | GT -27 |
| 2009 Iowa | Bowl Game | GT by 1 | Iowa by 10 | GT -11 |
| 2010 Air Force | Bowl Game | Air Force by 7 | Air Force by 7 | 0 |
| | | | | Total = GT -38 |
| GT Opponent | Extra Time (Easy game the week before) | What Result “Should Have” Been According to End of Year Sagarin | Actual Result with Extra Preparation Time | Difference |
| 2008 Boston College | Kent. St. | BC by 7 | GT by 3 | GT +10 |
| 2008 VT | Furman | VT by 10 | VT by 3 | GT +7 |
| 2010 Kansas | N. Dakota St (although Kansas lost this game...) | GT by 12 | Kansas by 3 | GT -15 |
| 2011 Maryland | Towson | GT by 14 | GT by 5 | GT -9 |
| | | | | Total = GT -7 |
| GT Opponent | Extra Time (Mid Week Game week before) | What Result “Should Have” Been According to End of Year Sagarin | Actual Result with Extra Preparation Time | Difference |
| 2008 Clemson | Wake Forest on a Thursday | Clemson by 9 | GT by 4 | GT +13 |
| 2009 Miami | FSU on Labor Day before GT on Thurs | Miami by 1 | Miami by 16 | GT -15 |
| 2010 NC State | Cincinnati the Thursday before | NC St. by 13 | NC State by 17 | GT -4 |
| 2010 MTSU | Troy the Tuesday before | GT by 19 | GT by 28 | GT +9 |
| 2011 NC State | Cincinnati the Thursday before | GT by 5 | GT by 10 | GT +5 |
| | | | | Total = GT +8 |
| GT Opponent | Extra Time (Full Bye Week) | What Result “Should Have” Been According to End of Year Sagarin | Actual Result with Extra Preparation Time | Difference |
| 2008 UNC | Bye Week | UNC by 8 | UNC by 21 | GT -13 |
| 2008 UGA | Bye Week | UGA by 12 | GT by 3 | GT +15 |
| 2010 UNC | Bye Week | UNC by 10 | GT by 6 | GT +16 |
| 2010 VT | Bye Week (but game was on a Thursday, so not a full 2 weeks) | VT by 19 | VT by 7 | GT +12 |
| 2010 UGA | Bye Week | UGA by 12 | UGA by 8 | GT +4 |
| 2011 UVA | Bye Week | GT by 2 | UVA by 3 | GT -5 |
| | | | | Total = GT +29 |
No comments:
Post a Comment