Page 1 of 2
Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 1:51 pm
by SuperHornet
I've just been going over Big Sky point differential, wondering how it would affect the rankings. Things would be WAY different. BTW, I'm factoring in the newbies. I'm also eliminating the skewing of FBS and D-II games from the PD; that is, I'm only counting FCS games. The results are astonishing in the middle, but the Top One and Bottom One are unchanged. Perhaps that shouldn't be surprising.
By Record
1. Montana State 4-1 (2-0)
2. Northern Arizona 3-1 (1-0)
3. Cal Poly 3-2 (1-0)
4. Portland State 2-2 (1-0)
5. Eastern Washington 3-2 (2-1)
5t Montana 3-2 (2-1)
7. Sacramento State 2-3 (1-2)
7t Weber State 2-3 (1-2)
7t Northern Colorado 2-3 (1-2)
10. UC Davis 2-3 (0-2)
11. Idaho State 1-4 (0-3)
By Point Differential
1. Montana State +53
2. Portland State +43
3. Northern Arizona +28
4. Montana +15
5. Sacramento State +1
5t Cal Poly +1
7. Eastern Washington -1
8. Weber State -4
9. UC Davis -5
10. Northern Colorado -50
11. Idaho State -74
I know it's a bit early because not everyone's played two conference games yet. But if this is telling me something, it's the following:
1. I've been wrong on MSU all year. They look to be for real.
2. I was right about Mr. Tiger. It may have taken a bit longer for his stuff to work than I'd thought, but PSU is looking tough for the rest of the way.
3. While it's way too early to say "it's over," the Griz MAY be in trouble.
4. There's a big bunching in the middle of the pack. It's going to take time to sort it all out.
5. There's a HUGE dropoff with the last two. And even there, UNC is a whole lot better than the Bungles.
6. The newbies are going to be middle-of-the-pack or better to start. Cal Poly will challenge for titles immediately, while UCD will be better than most Hornet fans expected (though it may take a year or two for them to stabilize).
Is any of this valid? Discuss, please.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 1:58 pm
by Silenoz
1. Montana State +53
2. Portland State +43
3. Northern Arizona +28
4. Montana +15
5. Sacramento State +1
5t Cal Poly +1
7. Eastern Washington -1
8. Weber State -4
9. UC Davis -5
10. Northern Colorado -50
11. Idaho State -74
I cannot even being to figure out what those numbers are
Okay I see, all non-D2 games. Well if we had to drop a D2, MSU can't count Drake.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 2:02 pm
by SuperHornet
Silenoz wrote:I cannot even being to figure out what those numbers are
This is point differential in FCS games. So, for Sac-MSU (for example), 64-61 = 3. MSU gets credit for +3, while Sac gets -3. Add up each team's FCS games (to avoid the skewing inherent in most games vs FBS or D-II), and look at the list. It's a comparison of the general worth of a team's offense and defense. At the very least, it validates MSU's #1 position and ISU's cellar position. In the middle seems open to debate. That's the debate I'm asking for.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 2:04 pm
by clenz
Silenoz wrote:
Okay I see, all non-D2 games. Well if we had to drop a D2, MSU can't count Drake.
Why? Drake is FCS
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 2:04 pm
by Silenoz
Well I don't want to get flamed by Pioneer fans, if there any here, but they are DIII teams trapped in FCS. Non-counter imo
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 2:10 pm
by Silenoz
SuperHornet wrote:Silenoz wrote:I cannot even being to figure out what those numbers are
This is point differential in FCS games. So, for Sac-MSU (for example), 64-61 = 3. MSU gets credit for +3, while Sac gets -3. Add up each team's FCS games (to avoid the skewing inherent in most games vs FBS or D-II), and look at the list. It's a comparison of the general worth of a team's offense and defense. At the very least, it validates MSU's #1 position and ISU's cellar position. In the middle seems open to debate. That's the debate I'm asking for.
Right, I missed your "non FBS/D2" part. The problem here is you now have teams with 2 qualifying games, including blowout wins over ISU. At this point they are very dependant on SOS
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 2:33 pm
by Ursus A. Horribilis
Silenoz wrote:Well I don't want to get flamed by Pioneer fans, if there any here, but they are DIII teams trapped in FCS. Non-counter imo
You may not have watched that game Sil but Drake is a pretty big physical team and they would compete with the bottom 3 Big Sky teams. They were give MSU a hell of a game in the first 1/2 and then things fell apart for them and MSU was able to take advantage.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 2:36 pm
by Silenoz
Ursus A. Horribilis wrote:Silenoz wrote:Well I don't want to get flamed by Pioneer fans, if there any here, but they are DIII teams trapped in FCS. Non-counter imo
You may not have watched that game Sil but Drake is a pretty big physical team and
they would compete with the bottom 3 Big Sky teams. They were give MSU a hell of a game in the first 1/2 and then things fell apart for them and MSU was able to take advantage.
Is that really saying anything?
Man, I could be making a lot of enemies with this thread
And I did not watch the game, but word in Bozeman-land is that the Cats played utterly, utterly terrible for a half
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:05 pm
by SuperHornet
This is getting good.
Sure, it's a bit early. But could this have an effect on deciding who might win a game between Team A and Team B later in the season? Or who might be more "playoff-worthy" among tied or near-tied teams?
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:08 pm
by blueballs
Given a large enough sample size, point differential is almost always a good predictor of relative strength and success, especially in pro basketball.
IMO 4 or 5 games is not a large enough sample size to draw solid conclusions from though.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:11 pm
by Ursus A. Horribilis
Silenoz wrote:Ursus A. Horribilis wrote:
You may not have watched that game Sil but Drake is a pretty big physical team and they would compete with the bottom 3 Big Sky teams. They were give MSU a hell of a game in the first 1/2 and then things fell apart for them and MSU was able to take advantage.
Is that really saying anything?
Man, I could be making a lot of enemies with this thread
And I did not watch the game, but word in Bozeman-land is that the Cats played utterly, utterly terrible for a half
Eh, some truth (very little) to that. You just couldn't compare most D2's to them. They were about 300 across the O line and had some pretty darn good LB's and a D End that was good enough to play on about any BSC team.
They were not a shit team. I expected them to be a shit team.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:15 pm
by Ursus A. Horribilis
SuperHornet wrote:This is getting good.
Sure, it's a bit early. But could this have an effect on deciding who might win a game between Team A and Team B later in the season? Or who might be more "playoff-worthy" among tied or near-tied teams?
Well since I believe it is one of the tie breakers in most conferences then I think you are on to a winner here....if it was 1970. I believe it's about the 3rd thing looked at in most tie breaker scenarios so....
Good work though, you've stumbled onto something that's about a 1/2 century old.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:22 pm
by Silenoz
Ursus A. Horribilis wrote:Silenoz wrote:
Is that really saying anything?
Man, I could be making a lot of enemies with this thread
And I did not watch the game, but word in Bozeman-land is that the Cats played utterly, utterly terrible for a half
Eh, some truth (very little) to that. You just couldn't compare most D2's to them. They were about 300 across the O line and had some pretty darn good LB's and a D End that was good enough to play on about any BSC team.
They were not a **** team. I expected them to be a **** team.
Okay, this is going to look retarded, but I'm going to try anyways (and I can't upload anything to a photobucket-type site from here):
|--------------D3----------------|
....................|--------------D2--------------|
...............................|------------------------------FCS--------------------------|
..............................................................................|--------------------------------FBS----------------------|
There is my rough theory of where any given team is going to perform for any given day. The worst FCS teams (Pioneer league) are going to perform (for the most part) on the bottom of our spectrum, akin to a typical D2, and hovering in the D3 area
edit: Okay, ignore the periods (that's what she said)
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:41 pm
by clenz
Silenoz wrote:Ursus A. Horribilis wrote:
You may not have watched that game Sil but Drake is a pretty big physical team and they would compete with the bottom 3 Big Sky teams. They were give MSU a hell of a game in the first 1/2 and then things fell apart for them and MSU was able to take advantage.
Is that really saying anything?
Man, I could be making a lot of enemies with this thread
And I did not watch the game, but word in Bozeman-land is that the Cats played utterly, utterly terrible for a half
So, using your logic you can't count point differential for your games with Idaho State Portland State, and whoever else is in the bottom three or four.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:55 pm
by Ursus A. Horribilis
Silenoz wrote:Ursus A. Horribilis wrote:
Eh, some truth (very little) to that. You just couldn't compare most D2's to them. They were about 300 across the O line and had some pretty darn good LB's and a D End that was good enough to play on about any BSC team.
They were not a **** team. I expected them to be a **** team.
Okay, this is going to look retarded, but I'm going to try anyways (and I can't upload anything to a photobucket-type site from here):
|--------------D3----------------|
....................|--------------D2--------------|
...............................|------------------------------FCS--------------------------|
..............................................................................|--------------------------------FBS----------------------|
There is my rough theory of where any given team is going to perform for any given day. The worst FCS teams (Pioneer league) are going to perform (for the most part) on the bottom of our spectrum, akin to a typical D2, and hovering in the D3 area
edit: Okay, ignore the periods (that's what she said)
No I get it. All I'm saying is that Drake was a much better team than I expected to see out there and I bet you would have been a little surprised by them as well.
Clenz made the initial point that I was making anyway.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:56 pm
by Silenoz
No, I'm saying we shouldn't discount any of the teams, D2, FBS, anyone. EWU had a tougher opponent in CWU then MSU had in Drake, labels be damned
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 3:57 pm
by Silenoz
And Ursus, Drake may very well be quite good, better than 1/3 of the teams in FCS, like I said I didn't get to catch the game (in fact, wasn't it happening while we were playing Sac?). But dropping D2s because of their label is the same as dropping Pioneer teams because of their label
Am I making any sense or do I just sound crazy?
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 4:44 pm
by Ursus A. Horribilis
Silenoz wrote:And Ursus, Drake may very well be quite good, better than 1/3 of the teams in FCS, like I said I didn't get to catch the game (in fact, wasn't it happening while we were playing Sac?). But dropping D2s because of their label is the same as dropping Pioneer teams because of their label
Am I making any sense or do I just sound crazy?
I'm pretty sure it was the weekend we were at Cal Poly or EWU but anyway I was taking you as saying it should be dropped instead of all should be included. I get what your saying. I'm sure it's crazy for that reason alone.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 7:08 pm
by JayJ79
whether you win by 1 point or 100, it's just just 1 in the W column.
Re: Is Point Differential a Valid Stat?
Posted: Wed Oct 06, 2010 7:24 pm
by CatMom
So the theory would hold that SFA gets more credit for their run-up score against Lamar than we do beating Cal Poly by 9?
Re: Is Point Differential a Valid Stat?
Posted: Thu Oct 07, 2010 7:50 am
by JBB
Its a valid stat when considered within the context of a teams statistical profile when used in comparison to another for purposes of predicting an outcome.
The stat can also be used as a measure of strength of schedule. There is a casual inverse relationship between point differential and schedule strength.
Re: Is Point Differential a Valid Stat?
Posted: Thu Oct 07, 2010 8:28 am
by UAalum72
CatMom wrote:So the theory would hold that SFA gets more credit for their run-up score against Lamar than we do beating Cal Poly by 9?
Depending on how much work you want to put into it, you could calculate diminishing returns for larger margins, or just cut off additional credit for margins above 40 points (or whatever number you choose), on the theory that if you can beat them by 35 you can win by any amount you choose.
Sagarin used to use diminishing returns until the BCS committee told him he couldn't.
Re: Is Point Differential a Valid Stat?
Posted: Thu Oct 07, 2010 6:48 pm
by FWJack
CatMom wrote:So the theory would hold that SFA gets more credit for their run-up score against Lamar than we do beating Cal Poly by 9?
CM, I don't think point differential can be a factor given the situation you listed. I believe it can be a valid stat when used to compare teams with common opponents ie. if TXST had played Lamar you might draw comparisons/conclusions about the differences between SFA & TXST.
Re: Is Point Differential a Valid Stat?
Posted: Thu Oct 07, 2010 6:51 pm
by FWJack
UAalum72 wrote:CatMom wrote:So the theory would hold that SFA gets more credit for their run-up score against Lamar than we do beating Cal Poly by 9?
Depending on how much work you want to put into it, you could calculate diminishing returns for larger margins, or just cut off additional credit for margins above 40 points (or whatever number you choose), on the theory that if you can beat them by 35 you can win by any amount you choose.
Sagarin used to use diminishing returns until the BCS committee told him he couldn't.
Interesting (the concept and the fact that the BCS committe told Sagarin not to use the diminishing returns factor). I like the idea of a cut off. Your right, if you can beat a team by 40 then you could probably beat them by any margin you choose.
Re: Is Point Differential a Valid Stat?
Posted: Thu Oct 07, 2010 7:14 pm
by CatMom
FWJack wrote:CatMom wrote:So the theory would hold that SFA gets more credit for their run-up score against Lamar than we do beating Cal Poly by 9?
CM, I don't think point differential can be a factor given the situation you listed. I believe it can be a valid stat when used to compare teams with common opponents ie. if TXST had played Lamar you might draw comparisons/conclusions about the differences between SFA & TXST.
I tend to agree. I imagine SFA will have larger margins in conference, so comparing point differentials there would be ok with me.
I wonder when, if ever, the SLC had to go to this, because we all play each other and, therefore, someone wins the head-to-head, they have never had to.