Geoff Collins as a HC - Full Season Data

[GCdaJuiceMan]

Where did the previous Temple coach go?

Baylor - started off his first season with a loss to FCS liberty.

 

[KCJackets]

Rhule was also at Temple from 2006-2011 as an assistant coach. Left for one year to be OL coach for the NYG in 2013 before returning to Temple as HC.

 

[CuseJacket]

Temple was 10 and 2 for 2015 and 2016 when the previous coach left.

ESPN shows a record of 10-4 for 2015 and 2016... is that incorrect / am I missing something?

Edit: looks like Rhule left before 2016 bowl game, making him 10-3* in 2016.

 

[CuseJacket]

We were 11-3 the year before 2015 with the same coaching staff. Theres to much that can happen between years compare with. Lost players, system change, injuries etc.

Thanks for the info! I think you would need to understand strength of schedule over the years as well and what players did not return his first year. If we are just looking at overall wins and loses I would rather have UVAs schedule than what we get annually with Clemson and UGA.

Temple lost their Senior QB - Phillip Walker - after 2016 at the same time Rhule left.

Here were some of Walker's accolades in 2016 from his Temple profile:

• American Athletic Conference Championship Game Most Outstanding Player
• CFPA National Player of the Year Watch List (top offensive player)
• Davey O'Brien Award Watch List (top quarterback)
• Johnny Unitas Golden Arm Award Watch List (top quarterback)
• Manning Award Watch List (Nations Most Outstanding QB)
• Maxwell Award Watch List (top offensive player)

I'm fairly ignorant on roster mix of Temple, so I didn't recognize Walker's name. A quick glance at stats show that he started all 14 games as a Junior and Senior. He played in 12 games as a Soph and 9 as a freshman.

https://owlsports.com/roster.aspx?rp_id=7150

 

[jandrews]

Thanks Cuse! I looked up SoS. Nothing really popped out. Temple floats around the high 60s to low 70s. CGC faced about the same.

First 2016 depth chart shows 12 senior starters. I would say that would cause some issues CGCs first year.

 

[Pointer]

Thanks Cuse! I looked up SoS. Nothing really popped out. Temple floats around the high 60s to low 70s. CGC faced about the same.

First 2016 depth chart shows 12 senior starters. I would say that would cause some issues CGCs first year.

I don't know if this is what you're implying or not, but I don't think it's fair to say someone performed better or worse because the level they were coaching was more difficult or easier respectfully. If someone is coaching at a non P5 school and does well, you can't discredit their success with the reasoning that the competition is easier. It's not like they are rolling out Bama level players and playing they're lower division. I think a coach who follows what makes them successful and doesn't try to change it because they go to a high level of playing (Chip Kelly comes to mind) then they have a chance at continuing their success. This is the main thing that has me nervous about CGC his previous HC records have not been great. What makes me optimistic is his defensive coordinating record, because he's put out elite level defenses. So maybe if he gets a good OC or if CDP wakes up, maybe his DC success can translate into his head coaching.

Again not sure if you were implying this, but I've seen allot of people bringing up this point recently. But I do agree losing 12 senior starters is tough, but the next year the record was worse so idk. We shall wait and see.

 

[jandrews]

I don't know if this is what you're implying or not, but I don't think it's fair to say someone performed better or worse because the level they were coaching was more difficult or easier respectfully. If someone is coaching at a non P5 school and does well, you can't discredit their success with the reasoning that the competition is easier. It's not like they are rolling out Bama level players and playing they're lower division. I think a coach who follows what makes them successful and doesn't try to change it because they go to a high level of playing (Chip Kelly comes to mind) then they have a chance at continuing their success. This is the main thing that has me nervous about CGC his previous HC records have not been great. What makes me optimistic is his defensive coordinating record, because he's put out elite level defenses. So maybe if he gets a good OC or if CDP wakes up, maybe his DC success can translate into his head coaching.

Again not sure if you were implying this, but I've seen allot of people bringing up this point recently. But I do agree losing 12 senior starters is tough, but the next year the record was worse so idk. We shall wait and see.

Sorry I was relating it back to my first post. But that is not what I was implying. I was just trying to say if we keep the variable of school the same, Rhule may have had an easier schedule than CGC had while he was at Temple. Now YoY, rosters change so the 12 seniors was could of caused his first year to be a little lower. Rhule got out after having a senior laden team. CGC had a mini rebuild on his hands at Temple. I've seen a lot on twitter of overall record. But you gotta look at the SoS as well to understand the complete picture. UCF went 12-1 lat year but their SoS was 65th. Nobody cares about the SoS after the overall record is mentioned. I would of rather had UVA's schedule last year and switched Clemson for NCST and UGA for Indiana. We would of had a completely different record.

 

[TheFlyest]

To be honest a lot of the Temple fans criticism centered around the offense and how it took them a year and two games to figure out who should be the starting qb, and how Collins brought in a spread coach for a team built to run a grinder pro style offense. Very sloppy and messy initially.

Any of this sounds familiar?

 

[jandrews]

To be honest a lot of the Temple fans criticism centered around the offense and how it took them a year and two games to figure out who should be the starting qb, and how Collins brought in a spread coach for a team built to run a grinder pro style offense. Very sloppy and messy initially.

Any of this sounds familiar?

Nope...

 

[tech_wreck47]

Temple lost their Senior QB - Phillip Walker - after 2016 at the same time Rhule left.

Here were some of Walker's accolades in 2016 from his Temple profile:

• American Athletic Conference Championship Game Most Outstanding Player
• CFPA National Player of the Year Watch List (top offensive player)
• Davey O'Brien Award Watch List (top quarterback)
• Johnny Unitas Golden Arm Award Watch List (top quarterback)
• Manning Award Watch List (Nations Most Outstanding QB)
• Maxwell Award Watch List (top offensive player)

I'm fairly ignorant on roster mix of Temple, so I didn't recognize Walker's name. A quick glance at stats show that he started all 14 games as a Junior and Senior. He played in 12 games as a Soph and 9 as a freshman.

https://owlsports.com/roster.aspx?rp_id=7150

Thanks for this! This is my point, that’s not easy to replace so picking record from those years really aren’t fair imo.

 

[grandpa jacket]

For those unfamiliar, Football Outsiders (FO) publishes a popular efficiency ranking that many have referenced in posts throughout the years. FO adjusts for strength of opponent and also removes garbage time, among other things, in determining each team's relative strength against another. In other words, the premise is this get team comparisons as close to "apples to apples" as possible.

I decided to revisit the rankings for Collins' two years at Temple, to put his Head Coach track record into greater context than 3 games at GT. This has been posted before in various capacities, however I thought it useful during a bye week, particularly in the lead-up to the game @ Temple.

Temple

1st year - 2017

• Regular season
  • 7-6 overall (includes bowl win)
  • 4-4 in conference
• Football Outsiders Rank
  • Overall - 61
  • OFEI - 94
  • DFEI - 36
  • SFEI - 39

2nd year - 2018

• Regular season
  • 8-5 overall (includes bowl loss)
  • 7-1 in conference
• Football Outsiders Rank
  • Overall - 45
  • OFEI - 58
  • DFEI - 42
  • SFEI - 54

Note: Football Outsiders excludes games against FCS opponents in their rankings. In 2018, Temple lost to FCS Villanova by a final score of 19-17 in the first week of the season.

Link to FEI: https://www.footballoutsiders.com/stats/fei/2018
Link to Temple box scores: https://www.espn.com/college-football/team/schedule/_/id/218/season/2018

How do you interpret this data? How much weight should we put on it?

Certainly there is additional context pre- and post-Collins, and also relative comparisons to GT, but I didn't want to get too detailed in the OP.

 

[grandpa jacket]

Thanks for consolidating. Hard to gauge since he was a first time HC and didn’t get a chance to do a full recruiting cycle. Also tough because I don’t know if those numbers are “good” for temple historically. When I get home I will look up historical data to try to compare.

3 Games into a season is unlikely to tell us anything, i do not agree that it will take 3 years to return to being a bowl team.

 

[Longestday]

So are you saying the data is un-reliable or can’t be used to make a judgment due to changes from year to year

I would say the data does not predict the future and has nuances.

I would say the data has similarities to the start of this season.

I would say the data shows we did not hire a coach with a proven record of a high level of achievement or turnaround ability. Please don’t miss understand... the data shows that we did not hire a coach with no level of achievement.

The only data we will have is 2017 through the current date.

 

[ibeattetris]

I’ve toyed with these metrics over the past few years and found they are too generic, although I love the concept.

What I’m looking for is a better way to answer how the team is performing. I don’t think these metrics reflect game strategy.

For example, in golf, the PGA Tour publishes all sorts of stats on total driving, shots under 150, shots 150-200, 200+, putting, etc.

And yet, one statistician said to forget all that. One stat, Greens in Regulation, is the single best predictor of score.

The search continues ...

Can I take a stab at trying explain why you might find FEI generic? Feel free to let me know if I’m off the mark.

In your example, greens in regulation is a predictive stat for overall performance. One could assume that if they increase Greens in Regulation, their performance will increase overall.

Offense and defense FEI on the other hand are descriptive stats. They try to determine how efficient an offense or defense has been. The stat itself uses the actual score as part of its equation. This would be the equivalent of saying a golfer will be better if the shoot fewer strokes per hole (while true, it doesn’t examine why a player may shoot less strokes per hole).

I agree with your premise a little, but I think football is a bit different than golf in this regard since there is an offense and defense component. If you are looking for non scoring stats that affect win rate, Yards per play diff (yards per play offense - yards per play defense) is one of the stronger stats that will predict outcome.

 

[Augusta_Jacket]

Can I take a stab at trying explain why you might find FEI generic? Feel free to let me know if I’m off the mark.

In your example, greens in regulation is a predictive stat for overall performance. One could assume that if they increase Greens in Regulation, their performance will increase overall.

Offense and defense FEI on the other hand are descriptive stats. They try to determine how efficient an offense or defense has been. The stat itself uses the actual score as part of its equation. This would be the equivalent of saying a golfer will be better if the shoot fewer strokes per hole (while true, it doesn’t examine why a player may shoot less strokes per hole).

I agree with your premise a little, but I think football is a bit different than golf in this regard since there is an offense and defense component. If you are looking for non scoring stats that affect win rate, Yards per play diff (yards per play offense - yards per play defense) is one of the stronger stats that will predict outcome.

Also, to tackle the greens in regulation point, data sets often are impacted by other data sets. Take GIR for instance, I would imagine that golfers with higher GIR % also have a higher % of fairways hit. SO while GIR is the stat most indicative of golfing success, it is largely driven by hitting fairways. FEI tends to look at ALL the stats that can be predictive of success and measure them as an efficiency.

 

[Vespidae]

Can I take a stab at trying explain why you might find FEI generic? Feel free to let me know if I’m off the mark.

In your example, greens in regulation is a predictive stat for overall performance. One could assume that if they increase Greens in Regulation, their performance will increase overall.

Offense and defense FEI on the other hand are descriptive stats. They try to determine how efficient an offense or defense has been. The stat itself uses the actual score as part of its equation. This would be the equivalent of saying a golfer will be better if the shoot fewer strokes per hole (while true, it doesn’t examine why a player may shoot less strokes per hole).

I agree with your premise a little, but I think football is a bit different than golf in this regard since there is an offense and defense component. If you are looking for non scoring stats that affect win rate, Yards per play diff (yards per play offense - yards per play defense) is one of the stronger stats that will predict outcome.

I agree with your characterizing metrics as predictive or descriptive. Here's the rub on those efficiency metrics though ...

If you think about football as Expected Value per Possession, the the more possessions a team has, the more likely they would converge towards their long-term EV/PP. (That's just statistics.) So if Clemson has an EV/PP of 4.25 and Tech has an EV/PP of 2.5, Tech would almost always lose.

So, if you are a weaker team, you are best served by limiting the number of possessions to increase the variance in the result. In other words, play for the upset. (This is the same concept in poker and gambling in general.)

I looked at CGC's results last year at Temple and his team's were all over ... 9 possessions per game, 16 possessions per game ... so it's not possible for me to answer ... did the strategy work?

That's what I'm struggling to evaluate ... the effectiveness of the game plan itself. I don't think the FEI ratings do that, they describe, as you said ... but they don't help evaluate. Tech could beat Clemson 7-6 but show horrible FEI numbers.

 

[Vespidae]

Also, to tackle the greens in regulation point, data sets often are impacted by other data sets. Take GIR for instance, I would imagine that golfers with higher GIR % also have a higher % of fairways hit. SO while GIR is the stat most indicative of golfing success, it is largely driven by hitting fairways. FEI tends to look at ALL the stats that can be predictive of success and measure them as an efficiency.

I thought so too. But the latest data shows that it's not Fairways in Regulation that count ... it's remaining distance to the hole. Golf has changed dramatically in the last ten years and why everyone is bombing it off the tee and not worrying about accuracy.

 

[tmhunter52]

Now this feels more like the gtswarm I'm used to.

Geeks

 

[Augusta_Jacket]

I thought so too. But the latest data shows that it's not Fairways in Regulation that count ... it's remaining distance to the hole. Golf has changed dramatically in the last ten years and why everyone is bombing it off the tee and not worrying about accuracy.

There is still a huge correlation between fairway accuracy in driving and greens in regulation, which is correlated to scoring. While driving longer reduces distance to hole, closer shots from bad lies reduces GIR. Golf is also harder to be consistent with statistically as each course is unique and rewards accuray/length in different ways.

 

[ibeattetris]

I agree with your characterizing metrics as predictive or descriptive. Here's the rub on those efficiency metrics though ...

If you think about football as Expected Value per Possession, the the more possessions a team has, the more likely they would converge towards their long-term EV/PP. (That's just statistics.) So if Clemson has an EV/PP of 4.25 and Tech has an EV/PP of 2.5, Tech would almost always lose.

So, if you are a weaker team, you are best served by limiting the number of possessions to increase the variance in the result. In other words, play for the upset. (This is the same concept in poker and gambling in general.)

I looked at CGC's results last year at Temple and his team's were all over ... 9 possessions per game, 16 possessions per game ... so it's not possible for me to answer ... did the strategy work?

That's what I'm struggling to evaluate ... the effectiveness of the game plan itself. I don't think the FEI ratings do that, they describe, as you said ... but they don't help evaluate. Tech could beat Clemson 7-6 but show horrible FEI numbers.

Every point you raise is spot on. One thing I’ve wanted to dive into more is the S&P+ metric. Since it is based on per play data, it should have less overall variance than drive data. The general rule bring up still holds: underdog teams should limit possessions to increase variance as a game plan (taken in a vacuum when there are no specific personnel mismatches that can be exploited). From what I read about the Temple offense, they wanted to play fast all the time. I think this is a mistake, just like I think CPJ not implementing and maintaining a hurry up was a mistake.

I would be interested to know which Temple game only had 9 possessions. I would guess the Navy game, so wouldn’t take as much stock in it.