2019 FEI and PPD Data

[ibeattetris]

Now that FEI is officially 2019 only data I am starting this thread for weekly updates regarding our FEI and PPD. If you want to see my previous thread for PPD data, it can be found here: https://gtswarm.com/threads/bcf-toys-2019-point-per-drive-data-now-live.19614/#post-635386
For information on what the stats are

Spoiler: What each stat means

The Fremeau Efficiency Index (FEI) is a college football rating system based on opponent-adjusted possession efficiency, representing the per possession scoring advantage a team would be expected to have on a neutral field against an average opponent. Unadjusted possession efficiency (PE) is calculated as a function of offensive, defensive, and special teams game splits. Schedule strength is represented by each team's average per possession opponent adjustment (OA). Opponent-adjusted offense ratings (OFEI), opponent-adjusted defense ratings (DFEI), and opponent-adjusted special teams ratings (SFEI) are calculated in a similar manner as overall FEI ratings.

GT

Interesting thing to note here is we have a top 20 strength of schedule so far according to FEI.

For the offense

Spoiler: What each stat means

OFEI Offense Ratings (OFEI) represent the per possession scoring advantage a team's offense would be expected to have on a neutral field against an average defense. Unadjusted offensive possession efficiency (OPE) is calculated as a function of offensive game splits. Schedule strength is represented by each team's average per possession offensive opponent adjustment (OOA). Touchdown rate (OTD) is the percentage of offensive drives that result in a touchdown. First Down rate (OFD) is the percentage of offensive drives that result in a touchdown or at least one first down. Available Yards percentage (OAY) is the total yards earned by the offense divided by the total yards available to be earned based on starting field position. Explosive Drive rate (OED) is the percentage of offensive drives that average at least ten yards per play. Busted Drive rate (OBD) is the percentage of offensive drives that earn zero or negative yards. Turnover rate (OTO) is the percentage of offensive drives that result in a fumble or interception.

Things I take away from this. We have faced top 5 strength of schedule for defenses. The other stat I found interesting was OED which is a measure on explosive drives. It is our best stat and seems to indicate that when our drives are successful, they are *really* successful.

For the defense

I don't have much to comment here. Hope to see improvement.

On to PPD data. From last week:

Another week in the books. Things still not looking good statistically for the good guys. Slid down two more ranks to 109 overall. We did increase our offensive points per drive from 0.68 to 0.98, the problem though is we still fell in OPD rankings (from 126 to 128) due to other teams improving by more. If we erase the first half it looks better, but we sadly can't do that. I hope to see continued improvement (there has certainly been improvement). We continue to have zero points on drives starting inside our own 20, hopefully this can be something we fix this week. Defense slid again, but I am not too surprised as they are the best offense we've seen since Clemson. Duke is currently scoring 2.42 (46) PPD while giving up 2.00 (51).

109 Georgia Tech
Net point score per drive: -1.41 (109)

Offense
PPD: 0.98 (128)
PPD (long field position): 0.00 (113)
PPD (medium): 1.33 (101)
PPD (short): 1.17 (126)

Defense
PPD: 2.39 71
PPD (long): 1.58 67
PPD (medium): 2.32 83
PPD (short) 4.00 85

This week showed the first offensive improvement across the board. We scored our first points on a drive >80 yard, and also improved PPD in all other starting field positions. Considering Duke is currently ranked 26 in DFEI, I think we can be proud of that. We still need to keep getting better (obviously), but week over week the offense seems to be coming together.

Defense slips for the second week in a row. I've posted elsewhere that I am not concerned about this because we are playing better offenses now than we were previously (excluding Clemson). Miami is currently the 48th FEI ranked team with 61st ranked OFEI and 36st DFEI. The one place I'd like to see improvement this week on defense is in short drives. We need to start holding teams to more FG's on drives starting inside our 40.

115 Georgia Tech
Net point score per drive: -1.40 (115)

Offense
PPD: 1.18 (121)
PPD (long field position): 0.23 (119)
PPD (medium): 1.39 (98)
PPD (short): 1.86 (121)

Defense
PPD: 2.57 (89)
PPD (long): 2.00 (86)
PPD (medium): 2.13 (74)
PPD (short) 4.45 (103)

That's it for this week. If anyone has any suggestions or stats they want to see let me know.

 

[SteamWhistle]

It is true that most of our best drives have ended with a long TD pass or were sparked by a long run. Very similar to most Spread Option teams. Everyone loved to talk about Death March but the Triple was at its best when we could get to the 2nd level and make the right read. Right now Tech isn’t comfortable just trying to Nickel and Dime teams like the UGA’s of the world. Even with the offense change we are still trying to hit the home runs. I think the more reps the more consistent it will get and we will be able to sustain some good drives without big plays. They sure had me fooled after the opening drive last week.

 

[ibeattetris]

It is true that most of our best drives have ended with a long TD pass or were sparked by a long run. Very similar to most Spread Option teams. Everyone loved to talk about Death March but the Triple was at its best when we could get to the 2nd level and make the right read. Right now Tech isn’t comfortable just trying to Nickel and Dime teams like the UGA’s of the world. Even with the offense change we are still trying to hit the home runs. I think the more reps the more consistent it will get and we will be able to sustain some good drives without big plays. They sure had me fooled after the opening drive last week.

I think right now teams are playing us pretty similar to how they did in the past. Stack the box until we prove we can beat them going over the top. We just haven't had consistent success doing that.

 

[alagold]

I think right now teams are playing us pretty similar to how they did in the past. Stack the box until we prove we can beat them going over the top. We just haven't had consistent success doing that.

yep, the passing stats look a LOT like the Sprd Option stats of past..The Run stats are nowhere close except maybe the QB run stats.

 

[gtrower]

Thanks for putting this all together. Depressing, but informative.

 

[deeeznutz]

Sad as it is to say, but we probably are a big factor in the defenses we’ve played being a top 5 SOS. We’ve made a few very average defenses look like world beaters.

 

[ibeattetris]

Sad as it is to say, but we probably are a big factor in the defenses we’ve played being a top 5 SOS. We’ve made a few very average defenses look like world beaters.

If I'm ever bored enough and have the time, I want to see if I can replicate something close to FEI numbers, specifically, so I can try to calculate what the values would be excluding certain games. I do think we are helping boost some of our opponents D, but we are also lowering their SOS, which should also hurt them (but I don't know where that balance is).

 

[ibeattetris]

Now that FEI is officially 2019 only data I am starting this thread for weekly updates regarding our FEI and PPD. If you want to see my previous thread for PPD data, it can be found here: https://gtswarm.com/threads/bcf-toys-2019-point-per-drive-data-now-live.19614/#post-635386
For information on what the stats are

Spoiler: What each stat means

The Fremeau Efficiency Index (FEI) is a college football rating system based on opponent-adjusted possession efficiency, representing the per possession scoring advantage a team would be expected to have on a neutral field against an average opponent. Unadjusted possession efficiency (PE) is calculated as a function of offensive, defensive, and special teams game splits. Schedule strength is represented by each team's average per possession opponent adjustment (OA). Opponent-adjusted offense ratings (OFEI), opponent-adjusted defense ratings (DFEI), and opponent-adjusted special teams ratings (SFEI) are calculated in a similar manner as overall FEI ratings.

GT
View attachment 6985

Interesting thing to note here is we have a top 20 strength of schedule so far according to FEI.

For the offense

Spoiler: What each stat means

OFEI Offense Ratings (OFEI) represent the per possession scoring advantage a team's offense would be expected to have on a neutral field against an average defense. Unadjusted offensive possession efficiency (OPE) is calculated as a function of offensive game splits. Schedule strength is represented by each team's average per possession offensive opponent adjustment (OOA). Touchdown rate (OTD) is the percentage of offensive drives that result in a touchdown. First Down rate (OFD) is the percentage of offensive drives that result in a touchdown or at least one first down. Available Yards percentage (OAY) is the total yards earned by the offense divided by the total yards available to be earned based on starting field position. Explosive Drive rate (OED) is the percentage of offensive drives that average at least ten yards per play. Busted Drive rate (OBD) is the percentage of offensive drives that earn zero or negative yards. Turnover rate (OTO) is the percentage of offensive drives that result in a fumble or interception.

View attachment 6986

Things I take away from this. We have faced top 5 strength of schedule for defenses. The other stat I found interesting was OED which is a measure on explosive drives. It is our best stat and seems to indicate that when our drives are successful, they are *really* successful.

For the defense
View attachment 6987

I don't have much to comment here. Hope to see improvement.

On to PPD data. From last week:

This week showed the first offensive improvement across the board. We scored our first points on a drive >80 yard, and also improved PPD in all other starting field positions. Considering Duke is currently ranked 26 in DFEI, I think we can be proud of that. We still need to keep getting better (obviously), but week over week the offense seems to be coming together.

Defense slips for the second week in a row. I've posted elsewhere that I am not concerned about this because we are playing better offenses now than we were previously (excluding Clemson). Miami is currently the 48th FEI ranked team with 61st ranked OFEI and 36st DFEI. The one place I'd like to see improvement this week on defense is in short drives. We need to start holding teams to more FG's on drives starting inside our 40.

115 Georgia Tech
Net point score per drive: -1.40 (115)

Offense
PPD: 1.18 (121)
PPD (long field position): 0.23 (119)
PPD (medium): 1.39 (98)
PPD (short): 1.86 (121)

Defense
PPD: 2.57 (89)
PPD (long): 2.00 (86)
PPD (medium): 2.13 (74)
PPD (short) 4.45 (103)

That's it for this week. If anyone has any suggestions or stats they want to see let me know.

Next week in FEI data came out on Wednesday. GT sees a six spot jump after its win in Miami.

Spoiler: What each stat means

The Fremeau Efficiency Index (FEI) is a college football rating system based on opponent-adjusted possession efficiency, representing the per possession scoring advantage a team would be expected to have on a neutral field against an average opponent. Unadjusted possession efficiency (PE) is calculated as a function of offensive, defensive, and special teams game splits. Schedule strength is represented by each team's average per possession opponent adjustment (OA). Opponent-adjusted offense ratings (OFEI), opponent-adjusted defense ratings (DFEI), and opponent-adjusted special teams ratings (SFEI) are calculated in a similar manner as overall FEI ratings.

Overall offense and special teams is the main contributor for the jump as we see each improving by 7 spots this week.

Offense:

Spoiler: What each stat means

OFEI Offense Ratings (OFEI) represent the per possession scoring advantage a team's offense would be expected to have on a neutral field against an average defense. Unadjusted offensive possession efficiency (OPE) is calculated as a function of offensive game splits. Schedule strength is represented by each team's average per possession offensive opponent adjustment (OOA). Touchdown rate (OTD) is the percentage of offensive drives that result in a touchdown. First Down rate (OFD) is the percentage of offensive drives that result in a touchdown or at least one first down. Available Yards percentage (OAY) is the total yards earned by the offense divided by the total yards available to be earned based on starting field position. Explosive Drive rate (OED) is the percentage of offensive drives that average at least ten yards per play. Busted Drive rate (OBD) is the percentage of offensive drives that earn zero or negative yards. Turnover rate (OTO) is the percentage of offensive drives that result in a fumble or interception.

Nothing improved by much. It has our OTD rate increasing by a decent amount, so I assume the data source includes our special teams touch down for the offense (otherwise we had two touch down on ten drives, which seems to low for the jump we had). Going forward I would really like our OFD to increase. We can't afford to have so many three and outs. 40% of our drives last week were three and outs. The two long drives in the second half certainly helped us keep the drives down, I'd like to see more of those (and hopefully some points from them as well).

Defense:

Spoiler: What each stat means

DFEI Defense Ratings (DFEI) represent the per possession scoring advantage a team's defense would be expected to have on a neutral field against an average offense. Unadjusted defensive possession efficiency (DPE) is calculated as a function of defensive game splits. Schedule strength is represented by each team's average per possession defensive opponent adjustment (DOA). Touchdown rate (DTD) is the percentage of opponent offensive drives that result in a touchdown. First Down rate (DFD) is the percentage of opponent offensive drives that result in a touchdown or at least one first down. Available Yards percentage (DAY) is the total yards surrendered by the defense divided by the total yards available to be earned based on starting field position. Explosive Drive rate (DED) is the percentage of opponent offensive drives that average at least ten yards per play. Busted Drive rate (DBD) is the percentage of opponent offensive drives that earn zero or negative yards. Turnover rate (DTO) is the percentage of opponent offensive drives that result in a fumble or interception.

General improvement overall, but nothing really noteworthy.

PPD Data. From last week

We see a general bump in most of the stats. Hopefully the week over week improvement continues as it is great to see. My happiest offensive stat this week is the jump in PPD from short distance (less than 60 yards) as that was the game winning OT drive . Defense improved in each category. The PPD in short drives is still killing us. We gave up 3.5 ppd in short fields, which while much better than we've been, is still not great. Defense was helped by missed kicks, but that's part of the game.

Spoiler: PPD data explained

Teams are ranked in the table below by net points scored per drive (NPD), the difference between points scored per offensive drive (OPD) and points allowed per opponent offensive drive (DPD). Points per drive from long starting field position for the offense (OLD) and opponent offenses (DLD) are calculated on possessions that begin inside the offense's own 20-yard line. Points per drive from middle starting field position for the offense (OMD) and opponent offenses (DMD) are calculated on possessions that begin from the offense's own 20-yard line to its own 40-yard line. Points per drive from short starting field position for the offense (OSD) and opponent offenses (DSD) are calculated on possessions that begin less than 60 yards from the end zone.

107 Georgia Tech
Net point score per drive: -1.14 (107)

Offense
PPD: 1.32 (119)
PPD (long field position): 0.19 (124)
PPD (medium): 1.52 (94)
PPD (short): 2.50 (107)

Defense
PPD: 2.46 (85)
PPD (long): 1.86 (82)
PPD (medium): 1.97 (65)
PPD (short) 4.20 (100)

 

[Sidewalking]

Tech is statistically pretty awful. Better hope the turnover margin is in the positive for the remainder of the year.

 

[ibeattetris]

Tech is statistically pretty awful. Better hope the turnover margin is in the positive for the remainder of the year.

Yeah, that will help for sure. Our current offensive turnover rate is 16.7% of drives which really needs to improve if we want a chance at squeaking out some more wins.

 

[AE 87]

Rather than starting a new thread, I thought I'd add my PPDvPwr5 data.
As of this week, our OffPPDvPwr5 is 1.60, and our DefPPDvPwr5 is 2.96. For the 71 teams who have played at least 3 Pwr5 opponents, the average OffPPDvPwr5 is 2.12, and we are ranked 54 of 71. The average DefPPDvPwr5 for these 71 teams is 2.26, and we are ranked 60.

An estimated Team ranking can be calculated by either subtraction or division of offense over defense scoring efficiency. We are #63 by subtraction and #62 by division. For those interested, here are the team rankings based on OffPPDvPwr5/DefPPDvPwr5 compared to AP and Coaches:

 

[Deleted member 2897]

Rather than starting a new thread, I thought I'd add my PPDvPwr5 data.
As of this week, our OffPPDvPwr5 is 1.60, and our DefPPDvPwr5 is 2.96. For the 71 teams who have played at least 3 Pwr5 opponents, the average OffPPDvPwr5 is 2.12, and we are ranked 54 of 71. The average DefPPDvPwr5 for these 71 teams is 2.26, and we are ranked 60.

An estimated Team ranking can be calculated by either subtraction or division of offense over defense scoring efficiency. We are #63 by subtraction and #62 by division. For those interested, here are the team rankings based on OffPPDvPwr5/DefPPDvPwr5 compared to AP and Coaches:
View attachment 7067

That data is clearly flawed as it has 5 ACC teams in the top 25. Minnesota having a heck of a year. Can we take an assist for giving them just the confidence they needed?

 

[bobongo]

That data is clearly flawed as it has 5 ACC teams in the top 25. Minnesota having a heck of a year. Can we take an assist for giving them just the confidence they needed?

Speaking of Minnesota, they are having a good year but to date their P-5 schedule has consisted of Nebraska, Illinois, Purdue, Maryland, and Rutgers, whose combined Big 10 record is 6-19. Without the context of SOS other than just "P-5", those stats can be quite misleading. While imperfect, at least the pollsters take SOS more into consideration. We'll find out what Minnesota is made of when they run up against Penn State in a couple of weeks.

 

[Deleted member 2897]

Speaking of Minnesota, they are having a good year but to date their P-5 schedule has consisted of Nebraska, Illinois, Purdue, Maryland, and Rutgers, whose combined Big 10 record is 6-19. Without the context of SOS other than just "P-5", those stats can be quite misleading. While imperfect, at least the pollsters take SOS more into consideration. We'll find out what Minnesota is made of when they run up against Penn State in a couple of weeks.

When those two teams play, who ya gonna gopher?

 

[AE 87]

Speaking of Minnesota, they are having a good year but to date their P-5 schedule has consisted of Nebraska, Illinois, Purdue, Maryland, and Rutgers, whose combined Big 10 record is 6-19. Without the context of SOS other than just "P-5", those stats can be quite misleading. While imperfect, at least the pollsters take SOS more into consideration. We'll find out what Minnesota is made of when they run up against Penn State in a couple of weeks.

Pollster's taking SOS into account is clear from the inclusion of App St, Boise St, and Memphis.

Seriously though, no ranking system is perfect because sports' teams are not static entities. Two teams will not have the exact same result even if they play each other successively. Even with the polling, coaches don't have the same opinion as writers, and they're skewed by pre-season expectations.

So, take it for what it's worth. My ranking is based on a raw stat. It is not an advanced stat like footballoutsiders which builds-in a strength of schedule adjustment (whose accuracy I dispute). My approach tries to account for strength of schedule by assuming that Pwr5 schedules largely balance out over the course of the season. I only rank teams who've played at least 3 Pwr5 teams because a team that plays only 1 or 2 pwr5 teams are more likely to have skewed results because of the strength of the opponents.

 

[bobongo]

When those two teams play, who ya gonna gopher?

I'd be lion if I said Minnesota.

 

[ibeattetris]

Rather than starting a new thread, I thought I'd add my PPDvPwr5 data.
As of this week, our OffPPDvPwr5 is 1.60, and our DefPPDvPwr5 is 2.96. For the 71 teams who have played at least 3 Pwr5 opponents, the average OffPPDvPwr5 is 2.12, and we are ranked 54 of 71. The average DefPPDvPwr5 for these 71 teams is 2.26, and we are ranked 60.

An estimated Team ranking can be calculated by either subtraction or division of offense over defense scoring efficiency. We are #63 by subtraction and #62 by division. For those interested, here are the team rankings based on OffPPDvPwr5/DefPPDvPwr5 compared to AP and Coaches:
View attachment 7067

Thanks for the content. I always enjoy reading these.

 

[bobongo]

Pollster's taking SOS into account is clear from the inclusion of App St, Boise St, and Memphis.

Seriously though, no ranking system is perfect because sports' teams are not static entities. Two teams will not have the exact same result even if they play each other successively. Even with the polling, coaches don't have the same opinion as writers, and they're skewed by pre-season expectations.

So, take it for what it's worth. My ranking is based on a raw stat. It is not an advanced stat like footballoutsiders which builds-in a strength of schedule adjustment (whose accuracy I dispute). My approach tries to account for strength of schedule by assuming that Pwr5 schedules largely balance out over the course of the season. I only rank teams who've played at least 3 Pwr5 teams because a team that plays only 1 or 2 pwr5 teams are more likely to have skewed results because of the strength of the opponents.

Well it will even out for Minnesota somewhat (and probably become more even generally for most teams) by the end of the year, after they play Penn State, Iowa, and Wisconsin.
But a legitimate ranking system has to take into account wins and losses combined with SOS in order to be complete, because the ultimate goal of the game is to win, not rack up PPD or any other stat.
Now, if you're talking about team strength instead of where they deserve to be ranked based on their record, I can see it. Those are two different things. But stats would have to incorporate special teams as well.

 

[takethepoints]

Thanks to both Tetris and AE for these posts. I know it must be a lot of work to put them together, but it sure is enlightening to those of us who are too lazy (sheepish look) to ferret this out for ourselves. Keep it up, if you can. We'll keep applauding.

And, yes, Minnesota is and was since the last half of last season a pretty good country football team.

 

[ibeattetris]

It is not an advanced stat like footballoutsiders which builds-in a strength of schedule adjustment (whose accuracy I dispute).

I’ve begun thinking that most ranking algorithms are very similar to Google’s search rank algorithm. They are all generally good at picking the top 5-6, but everything after that is a crap shoot.