CFB51 College Football Fan Community
The Power Five => Big Ten => Topic started by: OrangeAfroMan on December 02, 2018, 01:06:57 AM
-
So I wasted about 20 min trying to set up a system (groan) to rank general game outcomes. Logical stuff like a win is better than a loss, a blowout win is better than a close win, and a win over a good team is better than a win over a bad team.
Logical, right?
So the data involved are:
Win or Loss
Power5 or Group5
MOV - by 3+ possessions, by 2 possessions, or a 1 possession game
SOS -
2*...Was the opponent at .500 or a game above or below .500?
3*...Were they 2+ games over .500?
1*...Were they 2+ games below .500 (in conference games, to even out the scheduling disparity some)
So basically how did you do, against whom, by how much did you win or lose, and how good was your opponent?
The top and bottom of this system are super simple.
The top/best:
Win vs a 3* P5 team by 3 or more possessions
Win vs a 3* P5 team by 2 possessions
Win vs a 3* P5 team by 1 possession
-------------------------
Win vs a 2* P5 team by 3 or more possessions
Win vs a 2* P5 team by 2 possessions
Win vs a 2* P5 team by 1 possession
--------------------------
Okay, and the bottom, worst listed last:
Loss vs 2* G5 team by 1 poss.
Loss vs 2* G5 team by 2 poss.
Loss vs 2* G5 team by 3+ poss.
----------------------
Loss vs 1* G5 team by 1 poss.
Loss vs 1* G5 team by 2 poss.
Loss vs 1* G5 team by 3+ poss.
So that leaves a big middle, right? I intertwined wins vs bad P5 teams and good G5 teams, then wins vs bad G5. And the opposite as well - losses vs bad P5 teams and good G5 teams are better than losses vs bad G5 teams. Logical, still.
I'll list the entire hierarchy, then slide in the OU and OSU schedules and see what we get. Yes, I'm a nerd.
-
So the top and bottom of this will be a copy of what I posted already, but I want the whole thing on one post.
The hierarchy of outcomes:
W vs 3* P5 by 3+ poss.
W vs 3* P5 by 2 poss.
W vs 3* P5 by 1 poss.
W vs 2* P5 by 3+ poss.
W vs 2* P5 by 2 poss.
W vs 2* P5 by 1 poss.
W vs 1* P5 by 3+ poss.
W vs 3* G5 by 3+ poss.
W vs 1* P5 by 2 poss.
W vs 3* G5 by 2 poss.
W vs 1* P5 by 1 poss.
W vs 3* G5 by 1 poss.
W vs 2* G5 by 3+ poss.
W vs 2* G5 by 2 poss.
W vs 2* G5 by 1 poss
W vs 1* G5 by 3+ poss.
W vs 1* G5 by 2 poss.
W vs 1* G5 by 1 poss.
-----------------------------------I'd argue losing to a good P5 is "better" than here, but I digress
L vs 3* P5 by 1 poss.
L vs 3* P5 by 2 poss.
L vs 3* P5 by 3+ poss.
L vs 2* P5 by 1 poss.
L vs 2* P5 by 2 poss.
L vs 2* P5 by 3+ poss.
L vs 1* P5 by 1 poss.
L vs 3* G5 by 1 poss.
L vs 1* P5 by 2 poss.
L vs 3* G5 by 2 poss.
L vs 1* P5 by 3+ poss.
L vs 3* G5 by 3+ poss.
L vs 2* G5 by 1 poss.
L vs 2* G5 by 2 poss.
L vs 2* G5 by 3+ poss.
L vs 1* G5 by 1 poss.
L vs 1* G5 by 2 poss.
L vs 1* G5 by 3+ poss.
Remember that:
3* = a team that is at least 2 games over .500 in their conference
2* = a team that is a game above or below .500 or at-.500 in their conference
1* = a team that is 2 or more games below .500 in their conference
Army's overall record is used, as an independent, and as G5.
-
Okay, so here's where the Sooners' and Buckeyes' outcomes are:
W vs 3* P5 by 3+ poss. -----------------------OSU (UM,NW)
W vs 3* P5 by 2 poss.------OU (ISU,Tex)
W vs 3* P5 by 1 poss.------OU (WV)------------OSU (PSU)
W vs 2* P5 by 3+ poss.------OU (Bay,TCU)-------OSU(MSU)
W vs 2* P5 by 2 poss.------------------------------OSU (TCU)
W vs 2* P5 by 1 poss.
W vs 1* P5 by 3+ poss.----OU (UCLA,KSU)-------OSU (OrSt, RU, IU)
W vs 3* G5 by 3+ poss.------------------------OSU (Tul)
W vs 1* P5 by 2 poss.-----OU (KU)---------------OSU (Minn)
W vs 3* G5 by 2 poss.
W vs 1* P5 by 1 poss.-----OU (TTU, OkSt)----------OSU (UNL, UMd)
W vs 3* G5 by 1 poss.-----OU (Army)
W vs 2* G5 by 3+ poss.
W vs 2* G5 by 2 poss.
W vs 2* G5 by 1 poss
W vs 1* G5 by 3+ poss.------OU (FAU)
W vs 1* G5 by 2 poss.
W vs 1* G5 by 1 poss.
-----------------------------------
L vs 3* P5 by 1 poss.-------OU (Tex)
L vs 3* P5 by 2 poss.
L vs 3* P5 by 3+ poss.
L vs 2* P5 by 1 poss.
L vs 2* P5 by 2 poss.
L vs 2* P5 by 3+ poss.--------------------------OSU (Pur)
L vs 1* P5 by 1 poss.
L vs 3* G5 by 1 poss.
L vs 1* P5 by 2 poss.
L vs 3* G5 by 2 poss.
L vs 1* P5 by 3+ poss.
L vs 3* G5 by 3+ poss.
L vs 2* G5 by 1 poss.
L vs 2* G5 by 2 poss.
L vs 2* G5 by 3+ poss.
L vs 1* G5 by 1 poss.
L vs 1* G5 by 2 poss.
L vs 1* G5 by 3+ poss.
-
If it's hard to understand, Ohio State has the best 2 wins (Michigan, N'Western). They also have the worst loss (Purdue). Oklahoma has the next 3 least-impressive outcomes, though (L vs Texas, W over FAU and Army).
There are 36 levels on the hierarchy, so like the points system in the AP poll, we could see which team's outcomes yields the most points:
Oklahoma = 35+35+34+33+33+30+30+28+26+26+25+21+18
Ohio State = 36+36+34+33+32+30+30+30+29+28+26+26+13
OU = 374
OSU = 383
I would rank OSU over OU. And this is a big, fat support as to why. Yes, it assumes P5 conferences are equal. No, it doesn't count home or away. And no, it doesn't differentiate the elite teams. But it's a quick-n-dirty way to value each outcome.
And OU's defense sucks. :96:
-
Again, for clarity:
3* = good
2* = average
1* = bad
-
36 outcomes:
2x (W or L)
3x (3* or 2* or 1*)
2x (P5 or G5)
3x (3+ possession game, 2 poss, or 1 poss)
2 x 3 x 2 x 3 = 36 possibilities
-
I think, in part, of which team I'd rather face, and it would be Ohio State rather than OU.
I think OU gives Bama problems.
-
I don't understand it but I fully support it
-
If it's hard to understand, Ohio State has the best 2 wins (Michigan, N'Western). They also have the worst loss (Purdue). Oklahoma has the next 3 least-impressive outcomes, though (L vs Texas, W over FAU and Army).
There are 36 levels on the hierarchy, so like the points system in the AP poll, we could see which team's outcomes yields the most points:
Oklahoma = 35+35+34+33+33+30+30+28+26+26+25+21+18
Ohio State = 36+36+34+33+32+30+30+30+29+28+26+26+13
OU = 374
OSU = 383
I would rank OSU over OU. And this is a big, fat support as to why. Yes, it assumes P5 conferences are equal. No, it doesn't count home or away. And no, it doesn't differentiate the elite teams. But it's a quick-n-dirty way to value each outcome.
And OU's defense sucks. :96:
This is not to say things one way or another, but for a full record, OSU’s defense also very much sucks.
I think the OSU case is sneaky a little better. For some reason, I’d rather see OU, though Haskins trying to make a shootout with Bama would be just fine.
-
I think OSU is better, but I also think OU has the better overall resume. The WTF losses are keeping OSU out of the final four.
-
I lean to OU, but it's very close
Not only the bad loss to Purdue, but the Maryland game hurts the Buckeyes in my opinion
Bucks vs Michigan is more impressive than any Sooner victory
I'd like to see these two teams on the same field
-
I'll list the entire hierarchy, then slide in the OU and OSU schedules and see what we get. Yes, I'm a nerd.
thanks for crunching the numbers, nerd
-
and interesting, I haven't heard the talking heads mutter the name: Notre Dame
no sense talking about Notre Dame's resume. They are in with their shiny helmet and undefeated season
-
and interesting, I haven't heard the talking heads mutter the name: Notre Dame
no sense talking about Notre Dame's resume. They are in with their shiny helmet and undefeated season
Best win is a close one in home night game against Michigan.
-
I'd still go OU. As many wins over teams with winning records. More wins over bowl eligible teams. As many wins over ranked teams (maybe more if NW drops out). OU beat a common opponent by 25 and Ohio St beat them by 12.
The losses is the big one though. OU's was a 3 point game to a ranked team and they avenged it. Ohio St's was a 29 point loss to a 6-6 team.
As an aside, someone is going to have to explain ESPN's Strength of Record to me like I'm a 6 year old I guess. I keep seeing shown on graphics all morning how Ohio St has a better SOR than OU (4 to 6). What they aren't showing is ESPN's SOS rankings. Those have OU's SOS ranked #27 and Ohio St's ranked #46. If Ohio St and OU have the exact same record and OU's SOS is deemed superior then how is Ohio St's SOR higher? SOR is supposed to determine the strength of the record based against the strength of the schedule. How is Ohio St's SOR higher?
-
Best win is a close one in home night game against Michigan.
that's a good win, better than Georgia's best win, better than the Sooner's best win, much better than Clemson's best win
-
Rumblings I'm hearing is the committee is going to buffer their OU choice.
OU is 3, ND will be 4 and Georgia 5. With OSU still at 6 there "should be no debate" about their non-inclusion.
-
Vegas is saying Georgia would be favored over OU, OSU and ND.
-
I'd still go OU. As many wins over teams with winning records. More wins over bowl eligible teams. As many wins over ranked teams (maybe more if NW drops out). OU beat a common opponent by 25 and Ohio St beat them by 12.
The losses is the big one though. OU's was a 3 point game to a ranked team and they avenged it. Ohio St's was a 29 point loss to a 6-6 team.
As an aside, someone is going to have to explain ESPN's Strength of Record to me like I'm a 6 year old I guess. I keep seeing shown on graphics all morning how Ohio St has a better SOR than OU (4 to 6). What they aren't showing is ESPN's SOS rankings. Those have OU's SOS ranked #27 and Ohio St's ranked #46. If Ohio St and OU have the exact same record and OU's SOS is deemed superior then how is Ohio St's SOR higher? SOR is supposed to determine the strength of the record based against the strength of the schedule. How is Ohio St's SOR higher?
I'm not certain but my guess is because OU has a higher average per opponent, but OSU has a better opponent (Michigan) than any team OU faced. The calculation is more about the chances of having a particular record against the particular schedule, and so having a very strong team on the schedule throws it towards that schedule. In other words, beating several decent teams and one great team is "stronger" than beating several decent to good teams.
-
Heh. Cowherd had a slight against ND (and cold football teams in general.) He claimed there are 2 types of football, North and South. ND is the best northern team, with the caveat that OSU plays Southern Football. And all 4 southern teams (Alabama, Clemson, Oklahoma, and Ohio State) would crush ND by multiple scores.
-
Vegas is saying Georgia would be favored over OU, OSU and ND.
And I'm the a--hole suggesting the absurd...
-
I'd still go OU. As many wins over teams with winning records. More wins over bowl eligible teams. As many wins over ranked teams (maybe more if NW drops out). OU beat a common opponent by 25 and Ohio St beat them by 12.
The losses is the big one though. OU's was a 3 point game to a ranked team and they avenged it. Ohio St's was a 29 point loss to a 6-6 team.
As an aside, someone is going to have to explain ESPN's Strength of Record to me like I'm a 6 year old I guess. I keep seeing shown on graphics all morning how Ohio St has a better SOR than OU (4 to 6). What they aren't showing is ESPN's SOS rankings. Those have OU's SOS ranked #27 and Ohio St's ranked #46. If Ohio St and OU have the exact same record and OU's SOS is deemed superior then how is Ohio St's SOR higher? SOR is supposed to determine the strength of the record based against the strength of the schedule. How is Ohio St's SOR higher?
Seems like it incorporates much of what I did above.....ranking types of outcomes and plugging in the actual outcomes. Taking into account margin of victory or loss matters. OSU's worse loss is offset by some very low-rated wins by OU.
-
And I'm the a--hole suggesting the absurd...
Welp, you are an a--hole for sure. :)
Does the committee use Vegas odds as criteria?
-
And I'm the a--hole suggesting the absurd...
I got knocked out of two suicide pools because Vegas posted two NFL teams as 12-14 pt favs that lost outright.Happens every week - they just have to be right 53-55% of the time to turn a profit.
-
I'm pretty sure Vegas makes a profit on nearly every game no matter the outcome.
-
Every game no,but they handicap the Handicappers and middle on the betting curves.Just keep action pretty even on both sides and collect the Vig
-
Nearly every game, is what I said.
-
I'm pretty sure Vegas makes a profit on nearly every game no matter the outcome.
if they balance the wagers by adjusting the line
line is set to encourage balanced wagers, not to determine the outcome of the game
-
I'd like to know exactly but I think it's more around the lines of 60-70%.The Sharps(betting syndicates) usually win as much as the house but the betting public usually takes a bath
-
line is set to encourage balanced wagers, not to determine the outcome of the game
Ed Zachery
-
if they balance the wagers by adjusting the line
line is set to encourage balanced wagers, not to determine the outcome of the game
Right, which is the same as getting the consensus of thousands of informed, interested people.
-
We're about to find out. I think the top four will be "boring" and obvious, nothing tricky.
-
The "Georgia is one of the 4 best teams" idea isn't an argument for them to be included in the playoff. It's to show the committee and it's claim to let the 4 best in is BS. Same as the Heisman claiming to go to the most outstanding football player.
The committee should be honest and say it'll include the P5 teams with the best record as long as they thought highly of them from the start of the season. Just as the Heisman should change it to the best skill position player on a top team, but almost always the QB of a helmet team.
But honestly isn't a priority, it seems.
-
I'd like to know exactly but I think it's more around the lines of 60-70%.The Sharps(betting syndicates) usually win as much as the house but the betting public usually takes a bath
I tracked my picks for a year ATS back before I had kids and had more time to pay attention to CFB teams outside of Purdue. I was 62.5% ATS. All you need is >55%.
I had a few rules: no Purdue or ND (bias), don't bet the first two weeks (don't know what teams really are until week 3 or so), betting every B1G game and a few others where I saw something wrong with the line.
I ran the numbers and if I had been betting $100/game, I would have ended up making I think around $1200 over the course of the season. Which wasn't enough to risk ~$800/week over 13 weeks.
-
If it's hard to understand, Ohio State has the best 2 wins (Michigan, N'Western). They also have the worst loss (Purdue). Oklahoma has the next 3 least-impressive outcomes, though (L vs Texas, W over FAU and Army).
There are 36 levels on the hierarchy, so like the points system in the AP poll, we could see which team's outcomes yields the most points:
Oklahoma = 35+35+34+33+33+30+30+28+26+26+25+21+18
Ohio State = 36+36+34+33+32+30+30+30+29+28+26+26+13
OU = 374
OSU = 383
I would rank OSU over OU. And this is a big, fat support as to why. Yes, it assumes P5 conferences are equal. No, it doesn't count home or away. And no, it doesn't differentiate the elite teams. But it's a quick-n-dirty way to value each outcome.
And OU's defense sucks. :96:
I looked at Georgia with this method:
33+29+32+30+30+16+36+36+30+21+36+18 = 347 Well behind both OU and OSU.
-
I'm not certain but my guess is because OU has a higher average per opponent, but OSU has a better opponent (Michigan) than any team OU faced. The calculation is more about the chances of having a particular record against the particular schedule, and so having a very strong team on the schedule throws it towards that schedule. In other words, beating several decent teams and one great team is "stronger" than beating several decent to good teams.
Exactly.
SoS is basically the average strength of your opponents . SoR is chance of attaining a given record against a given slate of opponents.
SoR effectively corrects for something @ELA (https://www.cfb51.com/index.php?action=profile;u=55) complained about for years with the old RPI formula in basketball, namely:
For SoS purposes, playing #99 and #101 is equal to playing #1 and #199 but you obviously have a much better chance of losing a game if you play 1/199 than if you play 99/101.