Adjusted RPI (4/14)

[GoneGolfin]

_ Could you give us some insight and differences in Boyd Nation's ISR and the current RPI that the NCAA currently uses? I would first have to factor in home and away into the formula and maybe give more weight to OOWP into the formula.

And what other changes you would make if you could? _

The principal difference between the ISR and the RPI is that the ISR is not subject to the same connectivity issues that plagues the RPI. This is because the ISR employs a real algorithm that is recursive in nature and extends far beyond two levels deep into SOS (OWP and OOWP).

From Boyd's page ...
"The basic idea is an iterative one. Begin with all teams set to an even rating -- 100 in this case. Then, for each game played, give each team the value of their opponent's rating plus or minus a factor for winning or losing the game -- 25 in this case. Total all of a team's results, divide by the number of games played, and that's the end of a cycle. Then use those numbers as the start of the next cycle until you get the same results for each team for two consecutive cycles."

This is the best I can explain it without coding the algorithm for you. But simply imagine an RPI with no limits (until the recursive stop condition is hit) as to the depths that SOS is calculated. Meanwhile, RPI has an arbitrary depth of two (OWP and OOWP). Imagine OWP, OOWP, OOOWP, OOOOWP, ... until the stop condition is hit.

The only knock I have against the ISR is that it does not take into account home vs. road games. But that was not its intent. It was intended to provide a better version of what the RPI was attempting to accomplish.

As far as "upgrading" the RPI if I was stuck using it ... I would equally weight win/loss percentage and strength of schedule. I would also apply a factor to home wins and road wins. I have not decided what that should be. But obviously a home win would be worth less than 1.0 and a road win would be worth more than 1.0.

But again, the above is simply patching a flawed system. You cannot accurately assess SOS by simply going two levels deep. It is an arbitrary system.

Brian

[J-Town Cajun]

_ The principal difference between the ISR and the RPI is that the ISR is not subject to the same connectivity issues that plagues the RPI. This is because the ISR employs a real algorithm that is recursive in nature and extends far beyond two levels deep into SOS (OWP and OOWP).

From Boyd's page ...
"The basic idea is an iterative one. Begin with all teams set to an even rating -- 100 in this case. Then, for each game played, give each team the value of their opponent's rating plus or minus a factor for winning or losing the game -- 25 in this case. Total all of a team's results, divide by the number of games played, and that's the end of a cycle. Then use those numbers as the start of the next cycle until you get the same results for each team for two consecutive cycles."

This is the best I can explain it without coding the algorithm for you. But simply imagine an RPI with no limits (until the recursive stop condition is hit) as to the depths that SOS is calculated. Meanwhile, RPI has an arbitrary depth of two (OWP and OOWP). Imagine OWP, OOWP, OOOWP, OOOOWP, ... until the stop condition is hit.

The only knock I have against the ISR is that it does not take into account home vs. road games. But that was not its intent. It was intended to provide a better version of what the RPI was attempting to accomplish.

As far as "upgrading" the RPI if I was stuck using it ... I would equally weight win/loss percentage and strength of schedule. I would also apply a factor to home wins and road wins. I have not decided what that should be. But obviously a home win would be worth less than 1.0 and a road win would be worth more than 1.0.

But again, the above is simply patching a flawed system. You cannot accurately assess SOS by simply going two levels deep. It is an arbitrary system.

Brian _

Got it. Thanks! If Boyd's ISR system makes more sense I wonder why the NCAA would not use it. I know why home and road games are not factored in, it is because it would penalized the big conference teams. If the NCAA would implement this formula with road and home games factored in, then we would see more of the big boys travel which in the long run create more parity.

[CajunRed]

Correct. Right now they have ZERO incentive RPI wise to ever leave home.

Got it. Thanks! If Boyd's ISR system makes more sense I wonder why the NCAA would not use it. I know why home and road games are not factored in, it is because it would penalized the big conference teams. If the NCAA would implement this formula with road and home games factored in, then we would see more of the big boys travel which in the long run create more parity.

igeaux.mobi

[CajunRed]

Golfin...just want you to know that when i finally read your 4th to last paragraph...the one that mentions algorithum...i had to stop to hurry and go take two tylenol.

The principal difference between the ISR and the RPI is that the ISR is not subject to the same connectivity issues that plagues the RPI. This is because the ISR employs a real algorithm that is recursive in nature and extends far beyond two levels deep into SOS (OWP and OOWP).

From Boyd's page ...
"The basic idea is an iterative one. Begin with all teams set to an even rating -- 100 in this case. Then, for each game played, give each team the value of their opponent's rating plus or minus a factor for winning or losing the game -- 25 in this case. Total all of a team's results, divide by the number of games played, and that's the end of a cycle. Then use those numbers as the start of the next cycle until you get the same results for each team for two consecutive cycles."

This is the best I can explain it without coding the algorithm for you. But simply imagine an RPI with no limits (until the recursive stop condition is hit) as to the depths that SOS is calculated. Meanwhile, RPI has an arbitrary depth of two (OWP and OOWP). Imagine OWP, OOWP, OOOWP, OOOOWP, ... until the stop condition is hit.

The only knock I have against the ISR is that it does not take into account home vs. road games. But that was not its intent. It was intended to provide a better version of what the RPI was attempting to accomplish.

As far as "upgrading" the RPI if I was stuck using it ... I would equally weight win/loss percentage and strength of schedule. I would also apply a factor to home wins and road wins. I have not decided what that should be. But obviously a home win would be worth less than 1.0 and a road win would be worth more than 1.0.

But again, the above is simply patching a flawed system. You cannot accurately assess SOS by simply going two levels deep. It is an arbitrary system.

Brian

igeaux.mobi

[GoneGolfin]

_ Got it. Thanks! If Boyd's ISR system makes more sense I wonder why the NCAA would not use it. I know why home and road games are not factored in, it is because it would penalized the big conference teams. If the NCAA would implement this formula with road and home games factored in, then we would see more of the big boys travel which in the long run create more parity. _

Yes, there is certainly a powerful lobby here that plays a significant role in the decision to retain the current system.

I do not know why the NCAA has not considered the use of the ISR or something else superior to the RPI. But if they ditched the RPI in baseball, they would need to be prepared to ditch it in all sports. Anything else would come under harsh criticism.

To my knowledge, the NCAA has never even consulted Boyd to obtain more information. It make take the current generation of NCAA statisticians to retire before change is made.

Brian

[GoneGolfin]

_ Golfin...just want you to know that when i finally read your 4th to last paragraph...the one that mentions algorithum...i had to stop to hurry and go take two tylenol.

LOL. I think you are sandbagging me now. But the humor is appreciated.

Brian

[J-Town Cajun]

_ Correct. Right now they have ZERO incentive RPI wise to ever leave home.






igeaux.mobi _

There is a small incentive. If you get a top 25 road victory you will be awarded bonus points, and if you get a home lose against a weak RPI team (something like 200 or worse) you will get points taken away. So there is incentive but not enough to encourage the big teams to play away from home.

[CajunRed]

Still time to make some movement upward...but a 1 and 2 weekend means we're still at "63" as of Monday morning. We needed to finish out the season winning 2 of every 3 remaining games. Losing yet another series means we will need a sweep to cancel this one out.

Our hole got just a little deeper this weekend.

igeaux.mobi

[GoneGolfin]

_ Still time to make some movement upward...but a 1 and 2 weekend means we're still at "63" as of Monday morning. We needed to finish out the season winning 2 of every 3 remaining games. Losing yet another series means we will need a sweep to cancel this one out.

Our hole got just a little deeper this weekend.

Honestly, this team is doing a little better than I thought they would (overall). I figured the most likely scenario was that the Cajuns were a year out from another NCAA regional appearance. That still may be the case, but the Cajuns are not out of it.

Brian

[Dad04]

They are doing well for a team that lost nearly automatic winner Zach Osborne and now has two freshman closers. Next year should be better, assuming we don't get hurt in the draft too badly.