New forums are now live. Check it out for yourself.

2018 Offseason Stuff

Forum: College Football Page 1 of 8  1 2 3 4 5   »
Author: [College Football] Topic: 2018 Offseason Stuff
TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 2/2/2018 7:50:17 PM
I hope everyone is enjoying the college football offseason thus far!  

I'm nearly done with a final #s review of every team's cumulative 2017 performance (PR #s with various breakdowns).  Will share more of that later when I'm finally ready... guessing a week or two depending upon how busy I am at work.

The idea is to have an accurate full season PR# going into the offseason, then simply adjust up or down accordingly going into Week 1 2018.  Easier said than done.  And I'm sure all of us go to different levels of detail for their offseason analysis and adjustments.  This exercise also forces me to see how well a team did against different groupings of opponents (vs. top 10 teams, vs. #11-25, etc. whatever breakdown makes sense as dictated by their schedule), or how much worse a team performed on average when a key player was out with an injury, etc, etc.

Anyway, I'm not quite ready yet, but Mr. Connelly is.... see the article posted below for a good summary of returning production for 2018.  Mr. DoubleUp's Mississippi State team is mentioned among others.  Really like Clemson as well.  The entire D-line turning down the NFL tells us something IMO.  They have their eyes on the prize.

https://www.sbnation.com/college-football/2018/1/31/16950222/2018-ncaa-football-returning-starters-experience

Looking forward to comparing offseason notes for the next several months.  Take care guys.  And enjoy the Super Bowl this weekend.  


WISEGUY36 PM WISEGUY36
Joined: Jul 2012
Posts: 23778
 
WISEGUY36
Participation Meter
Hall of Fame
Posted: 2/2/2018 8:51:43 PM
TD , Thanks for posting . It's good to get some early info

I did some similar work right after the Jan.15th deadline for declaring for the NFL draft.

I've got as my top 5 with best  returning production :

1.  Michigan state
2.  Michigan
3.  Clemson
4.  USC ***  ( QB )
5.  Washington

Managed to buy some early futures on MSU and Wash.

Season can't get here soon enough.

TD  

 
WISEGUY36 PM WISEGUY36
Joined: Jul 2012
Posts: 23778
 
WISEGUY36
Participation Meter
Hall of Fame
Posted: 2/2/2018 10:08:54 PM
QUOTE Originally Posted by WISEGUY36:

TD , Thanks for posting . It's good to get some early info

I did some similar work right after the Jan.15th deadline for declaring for the NFL draft.

I'm only listing teams that most think might compete
for a National Championship.

I've got as my top 5 with best  returning production :

1.  Michigan state
2.  Michigan
3.  Clemson
4.  USC ***  ( QB )
5.  Washington

Managed to buy some early futures on MSU and Wash.

Season can't get here soon enough.

TD  

 




My Bottom 5 in returning production are as follows:

1.  Penn State , LSU
2.  TCU
3.  Ok.St.
4.  Okla.
5.  Wisc.

TD ,  
WISEGUY36 PM WISEGUY36
Joined: Jul 2012
Posts: 23778
 
WISEGUY36
Participation Meter
Hall of Fame
Posted: 2/3/2018 10:17:30 AM
I think that Bill Connely should start a new .......

" Preseason College Football Magazine " 

I'd buy it .   


MikeMed PM MikeMed
Joined: Aug 2010
Posts: 4793
 
MikeMed
Participation Meter
Veteran
Posted: 2/3/2018 12:08:57 PM

Great stuff TD

THNX for sharing.

 

For the record I believe Michigan may be the most underrated team going into 2018 season.........I think that offense will be much better after only having just a couple returning starters on offense coming in to 2017.......There defense brings back a ton of production in an already elite defense..................I actuall have them ranked 5th overall.

 

Obviously gotta go with Bama Clemson GA and OHST to fill out the top 5

Pecador PM Pecador
Joined: Jan 2010
Posts: 3074
 
Pecador
Participation Meter
Veteran
Posted: 2/5/2018 8:44:03 PM
TD,  So good your doing some Hot Stove football.  Hope the family is well.  

I couldn't care less about the top 5.  Bore me.   I'll make hundreds of bets this year and that top 5 will seldom be mentioned.  I'm looking forward to info on all but the so called top 5.  

I'm scuffling around trying to figure some Tex. area teams.  Nothing positive to say now but maybe will have a couple I think will disappoint.

My mind is numb from playing NCAAB.  I'm not going to study and play 250 teams playing twice a week any more. It's arduous and not particularly profitable.  I suck betting NCAAB

Dilly, Dilly
THEMUGG PM THEMUGG
Joined: Oct 2003
Posts: 32506
 
THEMUGG
Participation Meter
Legend
Posted: 2/7/2018 7:59:20 AM
 Interesting link, TD. I like Connelly's take on returning starters based on production. But it's also interesting that #1 is MSU.......followed by Kansas........1-11 Kansas. As has been said many times over the years here, having a lot of returning starters from a really bad team isn't always a positive. But I get the gist of it.

 Good info.
TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 3/8/2018 10:16:26 PM

Alright team, I’m a little later than estimated. 

Ever look at those Against the Spread (ATS) statistics at the end of the year and notice that some teams were great ATS (2017:  Iowa State 11-1-1 and Fresno State 11-2-1) and some teams were terrible ATS (2017:  Hawaii 1-10-1, UTEP 2-9-1, and WKU 3-10)?  What does this mean?  …. Some teams were undervalued most of the season from their true value and other teams were overvalued most of the season from their true value.  So if a team’s true CUMULATIVE value (numerically – their PR#) was “100% accurate” all season long, then theoretically their ATS record is 6-6 and their full season Against the Spread margin is net zero (*ATS margin being the number of points they did or didn’t cover by).  I can tell you that “100% accurate” doesn’t exist for determining a “one size fits all” PR#, but going through the process of trying to get close to a 6-6 ATS record and net 0.0 margin for ATS margin does make you notice things about each team.  Some teams are quite consistent week to week and other teams are more volatile in their performance.  I also notice tiers for a lot of teams.  As just one example, a  team might have what looks like a pretty accurate PR# against most of the teams on their schedule in the “same weight class” but as soon as they play a top 10 team they can’t keep up even with the point spread.  And/or the opposite also occurs; they may not only beat but also cover vs. terrible teams.  So, slight adjustments up or down to a team’s base PR# for a certain situations may be necessary.  Some good to great teams choose not to run up the score on lesser opponents.  Some ball control teams do well at keeping it close and covering vs. tough opponents.  The list of different situations  goes on and on.  A lot of this is intuitive and not surprising, but going through the numbers review process I find to be both fascinating and rewarding in the long run.

I was done with this final 2017 review effort, then recently decided to take one more pass at it and adjust individual team ratings based on conference ratings (i.e., the entire conference’s combined non-conference ATS record and ATS margins—because their in-conference ATS record and ATS margin are 0.500 and net zero by definition).  Basically, slide the entire conference up or down together.

So, reasons to look back at the previous season, rather than starting over from scratch every year:  1. Establish an accurate starting point for offseason adjustments.  2.  Noticing how well a team improved (or regressed) in the last 4-8 games of the season (especially with first or second year head coaches).  3.  Noticing how bad a team performed with a lame duck coach that ends up being fired – i.e., the new head coach inherits a team that is not as bad as the raw #s indicate.  4.  Noticing what types of tier teams exist and being on the lookout for similar teams in the upcoming season.  5. Determining what types of teams are consistent or volatile.

Anyway, here are my final 2017 post season ratings.  I’ll try to get into some more detailed breakdowns later this weekend for each team.  And there do end up being a few teams that I intentionally keep at a PR# that results in an ATS W/L record a couple notches above or below a 6-6 record for various reasons.  Specific examples later as I find time to post (Pretty damn busy lately).  And before I forget to mention, I choose to count any ATS margin result that is between +0.99 and -0.99 (less than a full point win or loss) as a push in their ATS record.  This just helps me notice when a team is very consistent (some teams have 2, 3, or 4 pushes in a season with this approach).



TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 3/8/2018 10:18:32 PM
7800 characters.... really!?

Alabama  98.2

Georgia  96.2
Clemson  95.4
Oklahoma  95.4
Penn State  95.1
Ohio State  94.9
Auburn  91.8
Wisconsin  91.2
Notre Dame  89.6
Washington  89.1
Oklahoma State 87.7
TCU  86.4
LSU  85.0
Central Florida  84.7
Mississippi State  83.3
Miami  83.2
USC  82.3
Texas  82.3
Virginia Tech  81.9
Stanford  81.5
NC State  81.4
Northwestern  80.6
Michigan  80.5
Iowa  80.4
Louisville  80.2
Florida State  79.9
Memphis  79.9
Wake Forest  79.6
Iowa State  79.4
Michigan State  79.2
Oregon  78.4
Georgia Tech  77.9
Boise State  77.7
Kansas State  77.1
Florida Atlantic  77.0
Washington State  76.9
Purdue  76.3
Boston College  76.2
Syracuse  75.7  *based on games with QB Dungey healthy.

TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 3/8/2018 10:20:22 PM
West Virginia  75.2
Texas A&M  74.9
Utah  74.8
South Carolina  74.7
Texas Tech  74.5
Houston  74.5
Missouri  74.2
Arizona  74.1
San Diego State  73.8
Duke  73.7
Fresno State  73.6
Pittsburgh  73.6
Arizona State  73.4
Toledo  73.3
UCLA  72.4
Mississippi  72.0
Troy  71.7
California  71.5
Indiana  71.4
Navy  70.1
Ohio U.  69.7
Army  69.6
Nebraska  69.5
Appalachian State  69.3
Florida  68.7
Virginia  68.4
Kentucky  68.0
Minnesota  67.4
SMU  67.3
Arkansas  67.2
Northern Illinois  67.0
Arkansas State  66.9
Colorado  66.4
Colorado State  66.1
North Carolina  66.1
Western Michigan  65.7
Wyoming  65.7
Temple  65.6

TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 3/8/2018 10:22:40 PM
Baylor  65.1
Central Michigan  64.8
Eastern Michigan  64.7
Maryland  64.1
Utah State  63.7
Middle Tennessee 63.5
Tennessee  63.2
Tulane  63.0
Air Force  63.0
Marshall 63.0
Rutgers  63.0
Vanderbilt  62.8
Buffalo  61.9
Southern Miss  71.8
North Texas  61.0
Louisiana Tech  60.0
Texas San Antonio  59.9
Miami (Ohio)  59.8
Massachusetts  59.2
New Mexico State  58.9
Nevada  58.9
Illinois  58.7
BYU  58.2
Idaho  58.7
UAB  57.3
Tulsa  57.2
UNLV  57.2
Louisiana Monroe  56.5
Akron  56.5
Florida International  55.5
Oregon State  55.5
Cincinnati  55.4
South Alabama  54.6
Connecticut  54.4
Georgia State  53.6
Western Kentucky  53.2
New Mexico  52.8
Bowling Green  52.7
East Carolina  51.9
Kansas  51.2
Coastal Carolina  50.9
Hawaii  50.0
Georgia Southern  49.8
Louisiana Lafayette  48.9
Old Dominion  47.5
Charlotte  44.8
Kent State  44.0
Rice  43.3
Texas State  41.7
San Jose State  40.0
UTEP  39.0
Ball State  36.4
TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 3/8/2018 10:26:26 PM

Just a sample of what I'll try to show next (running out of time tonight):

ACC:

Boston College: 

Power Rating = 76.2

ATS record:  5-7-1  *suggests they should be rated lower to improve their ATS record.

ATS margin:  +3.1 ppg  *suggests they should be rated higher to lower their ATS margin toward 0.0.

**This ATS margin figure decreases to +0.8 if the Syracuse game is excluded (QB Dungey out). 

ATS margin first 6 regular season games:  -9.0 ppg  (ATS record: 1-5)

ATS margin last 6 regular season games:  +16.1 ppg  (ATS record 4-1-1)

ATS margin average/mean vs. #3 -#21 rated opponents:  -8.4 ppg (ATS record: 0-3-1)

ATS margin average/mean vs. #24 -#28 rated opponents:  +4.1 ppg (ATS record: 2-2)

ATS margin average/mean vs. #41 -#85 rated opponents:  +14.6 ppg (ATS record: 3-1)

*Notice the very first team is an example of you can’t have it both ways (Record suggests they are overvalued, Margins suggest they are undervalued).  This is obviously because when they won, they won big.  Inconsistent teams for ATS margin lead me to do a standard deviation on this statistic so that I could then rank all teams from most consistent ATS margins to most volatile ATS margins.  Will post that info later.

Clemson: 

Power Rating = 95.4

ATS record:  7-7 

ATS margin:  +0.2 ppg 

ATS margin first 10 regular season games:  -1.4 ppg  (ATS record: 4-6) *Also, for games 6-10, they were 0-5 ATS.  Clemson seems to intentionally put it on cruise control until their best effort is needed.

ATS margin last 3 regular season games:  +10.7 ppg  (ATS record 3-0)  *I do include conference championship games as regular season (both teams are motivated, unlike some bowl games).

ATS margin average/mean vs. #1 -#25 rated opponents:  +3.4 ppg (ATS record: 4-2)  *This figure was +7.1 ppg going into the Alabama game.  Clemson plays well (generally) against better opponents.

ATS margin average/mean vs. #26 -#45 rated opponents:  -3.5 ppg (ATS record: 2-4)  *Anybody have Wake Forest +21 against Clemson last year?  Never in doubt, right!!??  Down 28-0 with 12 minutes left to play.  Dabo tends to go on cruise control vs. opponents in this tier.  Same thing happened in the 2016 championship season, they turned it on when they needed to. 
WISEGUY36 PM WISEGUY36
Joined: Jul 2012
Posts: 23778
 
WISEGUY36
Participation Meter
Hall of Fame
Posted: 3/9/2018 10:21:53 AM
Nice work , TD .  Digging into the " Science " of
establishing " Power ratings " can lead to some
great breakthroughs into how to get maximum
accuracy and benefit from them.
Digging into the analytics is not for everyone , but
I've found that in doing so it leads to more questions
that demand answers . It's actually a fun journey and
leads to some great  rewards. Well done , Bud.
You might consider taking your research just a bit
farther and use your ratings to establish a " win
probability " for each matchup as well.

TD , Great work , Buddy 

TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 3/28/2018 10:51:28 PM
Where are all the offseason college football fans!?  - I'm just too d@mn busy with work and coaching baseball and soccer to post very many of my thoughts for going into the 2018 season.  Hopefully I'll find a little more time in the coming weeks to post.

And whoever is moving a couple week 1 lines in my favor... Thanks!

2018 Week 1 plays locked in:
Texas -7 (hook)  (0.2u)  *posted previously in my bowl thread actually -- grabbed it that early.

Added this morning:
Washington +5.5 (0.5u)
Florida Atlantic +24 (0.3u)

LonghornHoosier PM LonghornHoosier
Joined: Sep 2007
Posts: 7098
 
LonghornHoosier
Participation Meter
Captain
Posted: 3/31/2018 2:05:38 AM
Nice work TD! 
TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 3/31/2018 3:35:31 PM
How much tougher does wagering on college football get from beginning to end of a season?

1 year sample size only, but here is what I've found for 2017.

Using the cumulative full season PR#s I listed above to determine a point spread vs. the actual Las Vegas point spread (closing lines each week per this site), the following are the average difference in point spread per game from a * "true line":

Weeks 0 / 1:  All games = 10.0
  • FBS vs. FBS = 9.0
  • FBS vs. FCS = 11.0
Week 2:  All games = 6.7
  • FBS vs. FBS = 6.5
  • FBS vs. FCS = 7.1
Week 3:  All games = 6.8
  • FBS vs. FBS = 6.3
  • FBS vs. FCS = 8.7
Week 4:  All games = 6.5
Week 5:  All games = 5.3
Week 6:  All games = 6.8
Week 7:  All games = 4.8
Week 8:  All games = 4.6
Week 9:  All games = 4.7
Week 10:  All games = 4.3
Week 11:  All games = 2.8
Week 12:  All games = 3.4
Week 13:  All games = 2.8
Week 14:  All games = 3.6
Week 15:  1 game = 2.0
Bowls:  All games = 2.9

* I put an asterisk by "true line" because my PR# per team is a full season cumulative # (because my offseason effort is an overview to find a starting point for 2018 offseason adjustments.  My week to week in-season approach would further correct the late season ebbs & flows indicated above).  In reality I think LV gets sharper every week late in the year.  Late season injuries, lame duck coaches, etc. are the explanation for the #s above bouncing around late in the year.  And week 14 is a smaller sample size as well.  Think of the full season cumulative PR# as if Team A plays Team B all 12 weeks, what is the one single baseline point spread for all 12 games between these two teams.  In-season, the PR#s (and the resulting point spread) adjust slightly week to week.
TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 3/31/2018 4:23:21 PM
I should add that some of the early season opening lines are off more (some significantly more) than the data above (delta vs. closing lines).

One extreme example from 2017 week 1 would be the UAB vs. Alabama A&M game.

After a full season, the larger amount of data available arrives at a fair line point spread of right around 38 to 39 points.  The actual week 1 LV closing line is listed as UAB -27.5.  So, 10.5 to 11.5 points off of what the line would have been if these two teams met again at the end of the season.  I was able to lock in UAB -9.5 last summer well before the season started.  UAB is obviously an extreme example because books did not know how to value them early in the year.

Several other examples out there I am sure, just wanted to point out that these figures are averages.  So, if week 1 lines are 'off' by 10 points on average, some games are 'off' more than 10 points, and some games are 'off' less.  Makes sense.  Staying ahead to find softer lines is a great approach early in the year.  But the #s above clearly show how much tighter the lines get week to week as a season goes on.

LonghornHoosier PM LonghornHoosier
Joined: Sep 2007
Posts: 7098
 
LonghornHoosier
Participation Meter
Captain
Posted: 3/31/2018 8:33:39 PM
So based on your summary, one's best chances ATS are in the first 6 weeks...After that, you have to get even sharper...
lvmike32 PM lvmike32
Joined: Oct 2010
Posts: 56
 
lvmike32
Participation Meter
Prospect
Posted: 4/5/2018 11:26:02 PM
thanks TD.  Great info and truly appreciate your willingness to share.  Are your Week 1 lines so far from 5D?  Or is there another book offering lines now?  thanks again
TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 5/19/2018 12:54:25 PM

Speaking of hammering early lines....

Betonline posted a few week 0/1 lines.

Locked in on:

UCF -16.5

Wyoming +1 (vs. NMSU)

Syracuse -1.5

TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 5/19/2018 12:59:42 PM

Slightly less on UCF with a new staff, but still the value is there on all three.

Also, smaller on RSW totals:

LSU over 6.5

Tennessee Under 6.5

Waiting on one Big Ten RSW total in particular to be offered.  Will be a larger play than the other two.

Sox PM Sox
Joined: Mar 2010
Posts: 72
 
Sox
Participation Meter
Prospect
Posted: 5/19/2018 4:43:16 PM
TD, thanks for the hard work and info you share with this site. 

I like the UCF and Wyoming play. I too see value at those numbers. As far as the Syracuse game, what do you have as a fair line? I capped them at -5.5 and with my book already at -3.5 I have lost the edge there. Based on your number do you still see value at -3.5? 

TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 5/19/2018 6:14:42 PM
QUOTE Originally Posted by Sox:

TD, thanks for the hard work and info you share with this site. 

I like the UCF and Wyoming play. I too see value at those numbers. As far as the Syracuse game, what do you have as a fair line? I capped them at -5.5 and with my book already at -3.5 I have lost the edge there. Based on your number do you still see value at -3.5? 


About -7 looks right to me with a healthy Dungey and Wassink at QB for both teams.  Maybe even -7.5, but that just me.  IMO still some value at -3.5.  Line should get bet into the -6 to -7 range is my guess.

Sox PM Sox
Joined: Mar 2010
Posts: 72
 
Sox
Participation Meter
Prospect
Posted: 5/19/2018 7:31:34 PM
Thanks TD. Appreciate the feedback. 
TD21 PM TD21
Joined: Jul 2013
Posts: 3725
 
TD21
Participation Meter
Veteran
Posted: 5/20/2018 9:24:47 PM
One more small play, think it changes soon....  updated card below:

Week 0/1 plays:
Syracuse -1.5  (2u)
Wyoming +1 (1.5u)
Central Florida -16.5  (0.75u)
Washington +5.5  (0.5u)
Florida Atlantic +24 hook (0.3u)
Texas -7 hook (0.2u)
Added:  Louisville +31.5  hook (0.2u)

RSW Win Totals:
Tennessee Under 6.5 -165  (risk 0.5 to win 0.303)
LSU Over 6.5 -110 (0.5u)

Forum: College Football Page 1 of 8  1 2 3 4 5   »
You have entered the forum as a GUEST. 
You must login/register to post or reply.
Desktop View: Switch to Mobile View