70 Years of GT Football – Strength of Schedule

Discussion in 'Georgia Tech Football' started by GTNavyNuke, May 5, 2014.

  1. GTNavyNuke

    GTNavyNuke Helluva Engineer Featured Member

    Messages:
    1,608

    One of the things we often talk about is how good a certain coach is or how good a certain GT team was and whether we ought to get a new coach or not. Putting aside the PR and money, I thought I’d look at just performance of the team.

    To do that, I have a three part post. The first talks about Power Rankings and Strength of Schedule over the last 70 years, the second, what team performance has been over the last 70 years under coaches and third, what are the odds of getting better performance under another coach.

    It is very difficult to compare teams since the sport of football has evolved so much. The typical way has been to compare win/loss records in a season. But that depends a lot on how good your opponents are in a certain year.

    One way to compare teams over many different years is to look at their Power Rankings. Power Rankings take into account not only relative strength of the other team, but where the game was played and margin of victory. The home field advantage is worth about 6 to 7 percent increase in the likelihood of a win (about 3.9 points). So who and where you play as well as how convincingly you win make a difference. The J Howell Power Rankings ones are the ones I use since the database goes back to 1869 (http://www.jhowell.net/cf/cfindex.htm ).

    A Power Ranking is a percentage which is the best estimate of how likely a team is to beat an average team that year on a neutral field. So last year, GT ended up with a Power Ranking of .642 which means that we had a 64.2% chance of beating last year’s average team on a neutral field. After the 2013 season, we were ranked 40th, so 39 teams had a higher Power Ranking. FSU had a 92.7% chance of beating that average team. So you get relative rankings that allow comparing teams in a certain year, and also the margin by which they dominated.

    Calculating SOS is similar. It considers the Power Ranking and where you play. Every year, the average of all teams SOS is about .5 (50%) Here’s the chart of our SOS from the beginning of Bobby Dodd’s time till now. Our SOS has gotten progressively weaker. It was highest in the SEC, then went down as we became independent and finally decreased even more in the ACC.

    [​IMG]


    Now realize that the SOS includes championship games (if any) and bowl games (if any). So if you win the ACCCG and go to the Orange bowl, your SOS is going to be higher. So this isn’t a pure reflection of our regular season lineup, but regular season and post season. That’s why 2009 was higher. And also why the Dodd teams of the 50’s where we were top 10 most of the time, resulted in high SOSs. Dodd played the toughest teams in the regular season and then had a tough bowl. For 2013, our SOS increased about 4% with the bowl game.

    Another interesting thing is that in our 1990 National Championship year, we had a weak SOS (68.5%) for a national champion. It helped that others lost and there wasn’t the BCS. But it also gives hope to teams with weak SOS to be able to win the NC.

    Finally here are the average SOS for our coaches in the last 70 years in the far right column where you can see how the average SOS has decreased steadily over the years with the exception of Curry and Gailey. [​IMG]
    We are going to get a tougher SOS next year, since we basically replace Georgia Southern with Notre Dame. (Next year, the D2 cupcake hasn’t been determined.) Here’s our current (and likely to change) out of conference schedule

    2014 08/30 vs Wofford, 09/06 at Tulane, 09/13 vs Georgia Southern, 11/29 at Georgia
    2015 TBD, 09/12 vs Tulane, TBA at Notre Dame, 11/28 vs Georgia
    2016 TDB, 09/03 vs Mercer, 09/17 vs Vanderbilt, 11/26 at Georgia

    What are people's thoughts on the SOS we play?
     
  2. iceeater1969

    iceeater1969 Ramblin' Wreck

    Messages:
    322
    Can you multiply the SOS % and winning % and get a meaningful number --- if so bobby Dodd 52 and paul Johnson - 36. This seems about
    Now for IMO.
    I hate cup cakes - the hurt recruiting. We are in Atlanta and should play tough out of conference teams as they don't count in the conference ranking. Was at tech right when Dodd retired. Clemson , virginia and duke Miami were the cupcakes between Tennessee and auburn and uga. I would rather be 1 for two against real teams in stead of beat down on 6a high school team.
     
  3. GTNavyNuke

    GTNavyNuke Helluva Engineer Featured Member

    Messages:
    1,608
    I agree there is a benefit in playing better teams for recruiting. But you have to win or be competitive, not be a doormat. So if you can compete and win 1 of 2 against great teams, then go for it. But against the group of really good teams we play (UGAg, Clemson, power bowl teams), we have not been competitive overall with the exception of Clemson in the receding past.

    But the more I think about it, the more sense it makes to have 3 of the 4 non-conference games be really tough ones. The rest of the ACC games are not that tough with the exception of Clemson and occasional FSU game or Miami team.

    I'd still put the cupcake in the week before the UGAg game if we had the luxury, but I think we get the ACC schedule and then work around it.

    As for comparing the program under different coaches, I'll post that Wednesday and Friday. :brb:;)
     
  4. Boomergump

    Boomergump Moderator Staff Member

    Messages:
    1,094
    The first thought that comes to mind for me is "isn't that what has happened across all of college football?" I feel sure the trend for all teams, even mighty Special Ed Conf teams is the same. The entirety of CFB is trying to avoid getting knocked out of the championship run through scheduling.
     
  5. 70Jacket

    70Jacket Jolly Good Fellow

    Messages:
    79
    Excellent observation!
     
  6. IronJacket7

    IronJacket7 Helluva Engineer

    Messages:
    1,236
    I know we don't want to be a doormat for the tougher opponents... However, if we played a lot tougher schedule wouldn't that theoretically get players more excited about playing those particular games. We may lose some but we might win some also and in turn it could help recruiting just by facing them period. I don't know... Just a thought. Not saying my logic is correct.

    Just seems to me that players "get up" for the bigger teams.
     
  7. AE 87

    AE 87 Helluva Engineer

    Messages:
    2,814
    I like having 2 or 3 easier games. It helps us get the kinks out against people not us and gives game time to back-ups.
     
    augustabuzz likes this.
  8. 4shotB

    4shotB Ramblin' Wreck

    Messages:
    116
    That supports what I have been thinking/saying for several years. Our records are camouflaged by playing in the ACC. That and the bowl streak glosses over things imo. I was at GT for the Pepper and Curry years and I'm really not sure that the program is really much better (in terms of on the field performance) than they were back then (excluding Curry's first several years of course) when all things are considered. I miss the excitement at BDS when Auburn, Tn, etc. came into town.
     
  9. Supersizethatorder-mutt

    Supersizethatorder-mutt Ramblin' Wreck

    Messages:
    361

    Being "competitive" does not always mean winning; it means being in a position to win. I think to say that we have not been competitive overall is not at all true. Gailey never beat Georgia, be we were certainly "competitive" with them in all but his first trip to Athens. The same can be said of Johnson's teams at Tech. There have been very few games in which we were not at least competitive. But for a botched play, or (as in 2012 particularly) a lousy defense, we very well could have won many of the games that we ended up losing.
     
  10. danny daniel

    danny daniel Ramblin' Wreck

    Messages:
    489
    Yes, we have a long history from Rogers, O'Learly, Gailey, to Johnson of letting the other team make a go-ahead fourth q drive while our attempted fourth q go-ahead drive ends in an interception. Is this QB skill differential, defensive recruiting, depth? it seems to cross several coaches.
     
  11. Supersizethatorder-mutt

    Supersizethatorder-mutt Ramblin' Wreck

    Messages:
    361
    I guess if anybody knew the answer to that question, they would be worthy of a head-coaching job somewhere. You are right: it has gone on at Tech for years, and it is frustrating as hell.
     
  12. augustabuzz

    augustabuzz Ramblin' Wreck

    Messages:
    402
    Some of that is perception. I noticed Howell didn't rate our 1990 schedule as strong even though Sagarin rated the ACC as the strongest conference that year.
     
  13. AE 87

    AE 87 Helluva Engineer

    Messages:
    2,814
    I think Howell uses a matrix solve based on margin of victory, but I think he also sets some initial conditions which may bias the final results some.

    I've never taken the time to think through how I would come up with a power ranking based only on final scores and home-field. It's obviously not a trivial problem.
     
  14. GTNavyNuke

    GTNavyNuke Helluva Engineer Featured Member

    Messages:
    1,608
    If you give me three teams, I'll check their SOS for the last 5 years. I think the top tier SEC has a very high SOS, but would like to check the "facts".

    If I get time, I'll check the NC team SOSs. I'm sure our 1990 win was one of the lowest in 20 years.
     
  15. GTNavyNuke

    GTNavyNuke Helluva Engineer Featured Member

    Messages:
    1,608
    Here are some notes I wrote up after e-mails with J Howell and copied from the J Howell site. http://www.mediafire.com/download/jvpmis9wcwld9ba/J_Howell_Methodology.pdf

    There is no perfect solution since at their core all ranking systems rely on the transitive property. As we know, that doesn't hold in many (10-30%??) games. So you are left with a error minimization algorithm. Many people have spent a lot of time on this given the financial incentive of predicting the outcome of games........ not trivial.
     
  16. GTNavyNuke

    GTNavyNuke Helluva Engineer Featured Member

    Messages:
    1,608
    J Howell ranked the ACC as the strongest conference in 1990 - they had the highest Power Ranking. It's at the bottom of http://www.jhowell.net/cf/cf1990.htm

    We got lucky in that all the other top contenders lost that year and we didn't (tied one).
     
  17. AE 87

    AE 87 Helluva Engineer

    Messages:
    2,814
    Thanks, but as far as I can tell those links don't tell me how they calculate P, i.e. the actual power rating, based on the historical results.
     
  18. AE 87

    AE 87 Helluva Engineer

    Messages:
    2,814
    I also think that a true SOS would include location of bye weeks, location of consecutive away games and short weeks. For example, last year's VPI on a short week followed by @d'ohU and @BYu is rougher than if we faced those same teams in a different order or with a bye in between. This year's schedule is much better for us in that regard. We get a bye before d'ohU and U[sic]Ga, Southern before VPI and NCState before CU. We really can't complain about scheduling this year, imo.
     
  19. augustabuzz

    augustabuzz Ramblin' Wreck

    Messages:
    402
    We played everyone in the conference. No one OOC played a healthy UVa and Clemson. Those two were easily top 10 teams.
     
  20. GTNavyNuke

    GTNavyNuke Helluva Engineer Featured Member

    Messages:
    1,608
    Sorry, when I said top contenders was thinking out of conference. (Big 8, Miami, Pac 10, SEC). Clemson was the next ACC team at 5th. Beating a top 10 team has to help with the credibility of giving us the NC.
     

Share This Page