Calpreps playoff boost defined

[golfaddict1]

7 minutes ago, NYHSFAN33 said:

I don't think a link is necessary when someone with even the most fundamental understanding of statistics would know that manually scaling a starting point for every team is literally the only way to create an algorithm like CalPreps does.  The fact that overall state point totals do not / nominally change over the course of a season unless out of state games are introduced is the only proof that is necessary.  I am stepping out of this one, because I will choose not to live in a real life Groundhog Day re-hashing the same talking points over and over.  It is a fact not really up for debate.  Without getting into too much boring detail, CalPreps has set up a Linear Program where all values are contained.  Read up on it if you are a math fan, it is interesting stuff and a tough thing to create, so kudos to Ned on that.  Just not sure it is executed correctly in order to rate teams across the country.

I'm also done discussing what is pretty obvious... I always get reeled in.  I need to be like Duran and say no mas LOL.  As for tinkering and how folks feel about every playoff round equaling a 210% boost vs regular season games, I'm all ears.  

[steeler01]

1 minute ago, golfaddict1 said:

I'm also done discussing what is pretty obvious... I always get reeled in.  I need to be like Duran and say no mas LOL.  As for tinkering and how folks feel about every playoff round equaling a 210% boost vs regular season games, I'm all ears.  

There should be no playoff boost. 

[Ararar]

9 minutes ago, steeler01 said:

There should be no playoff boost. 

I would agree.They don't get any boost from playing that same playoff team in the regular season 

[golfaddict1]

12 minutes ago, steeler01 said:

There should be no playoff boost. 

That's one way to go... as it keeps the ratings fluid throughout and not a regular season vs playoffs and divide by 2.  I'd be interested if this model would reduce the 2010 Camden County #847 type of ratings from happening, but I don't think I can test it as there was no trend option from CP in 2010.  

[steeler01]

6 minutes ago, golfaddict1 said:

That's one way to go... as it keeps the ratings fluid throughout and not a regular season vs playoffs and divide by 2.  I'd be interested if this model would reduce the 2010 Camden County #847 type of ratings from happening, but I don't think I can test it as there was no trend option from CP in 2010.  

It also keeps teams who play in terrible areas from getting unwarranted points.

[noonereal]

2 hours ago, golfaddict1 said:

 

  Intellects on this forum haven't contested my findings and Stanscript usually avoids my posts like the plague.  

I don't know you well enough to know how to read this.... I know your ego is.... more needy than most  but I also think you are a good guy at heart. 

In any event, an apology puts us square. 

just an FYI.

I told Cajun the same thing and he chose not to. 

Think about it. 

 

 

 

[noonereal]

3 hours ago, NYHSFAN33 said:

I don't think a link is necessary when someone with even the most fundamental understanding of statistics would know that manually scaling a starting point for every team is literally the only way to create an algorithm like CalPreps does. 

ya see to me, eliminating starting values is easy to do 

But a link is necessary because they keep speaking to it as if it's a given.

Question, why must each STATE be given a "scale"? Why not each section or region or area of the country encompassing many states? See? 

Did you understand something in my argument? Be honest. 

[954gator]

22 minutes ago, noonereal said:

ya see to me, eliminating starting values is easy to do 

But a link is necessary because they keep speaking to it as if it's a given.

Question, why must each STATE be given a "scale"? Why not each section or region or area of the country encompassing many states? See? 

Did you understand something in my argument? Be honest. 

I think if this wasn't done we'd get rankings more like we did in the early 2004-2007 versions.   You'll get a lot of random high rated WI, NM, NY teams that throughout the season, might not play an OOS.   How could a computer compare them to a team from say South Carolina?

Scaling has made the rankings much better, only very recently were Central, MNW, BTW, Norland (IE toughest Dade county schedules) given due credit in the ratings.   

[noonereal]

2 minutes ago, 954gator said:

I think if this wasn't done we'd get rankings more like we did in the early 2004-2007 versions.   You'll get a lot of random high rated WI, NM, NY teams that throughout the season, might not play an OOS.   How could a computer compare them to a team from say South Carolina?

Scaling has made the rankings much better, only very recently were Central, MNW, BTW, Norland (IE toughest Dade county schedules) given due credit in the ratings.   

please share how you know why scaling was done and why it was done by state rather than geographical area larger or smaller than a state.

thanks

[954gator]

Just now, noonereal said:

please share how you know why scaling was done and why it was done by state rather than geographical area larger or smaller than a state.

thanks

Just look at the trends man.  

[954gator]

Nor, no one is going to be able to give you Ned's exact formula.   

[noonereal]

2 minutes ago, 954gator said:

Just look at the trends man.  

differentiate as I asked

thanks

[noonereal]

5 minutes ago, 954gator said:

Nor, no one is going to be able to give you Ned's exact formula.   

I don't want it!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

The product is obviously awesome but some that reason from want other than than objectivity carry on like children., 

 

[954gator]

3 minutes ago, noonereal said:

differentiate as I asked

thanks

dy+7x dx=0

dy=−7x dx

∫dy=−∫7x dx

y=−2/7x^2+K

 

 

[954gator]

5 minutes ago, noonereal said:

I don't want it!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

The product is obviously awesome but some that reason from want other than than objectivity carry on like children., 

 

Oh you want it!  I kind of do too though. 

[NYHSFAN33]

30 minutes ago, noonereal said:

ya see to me, eliminating starting values is easy to do 

But a link is necessary because they keep speaking to it as if it's a given.

Question, why must each STATE be given a "scale"? Why not each section or region or area of the country encompassing many states? See? 

Did you understand something in my argument? Be honest. 

Your argument is actually correct IMO, starting points are eliminated, which is why CalPreps does a good job distributing the points amongst in state, same class competition.  It is actually an excellent algorithm in that sense and pretty impressive how close it comes sight unseen in rating these teams.  However, with any Linear Program, regardless of the desired outcome, there is a finite amount of points available in the data set to be distributed.  This finite number is set and spread among every team in the country.  Some states (I.e. Cali) start with a higher number of points to spread around.  I am not sure if I am explaining it with as much clarity as I would like to.  Basically my point is that if no team played any out of state competition, the total number of points for each state would either change very nominally or not at all throughout the course of a season, the points would just be spread out among those states teams. I might be correct, I might not be, but I have yet to see much if any evidence to disprove my hypothesis.

[noonereal]

Just now, 954gator said:

Oh you want it!  I kind of do too though. 

No I don't.

I know I don't agree with particulars but I am fine at taking it at what it is. 

It's the nuts that want to toss the whole thing because it's objective and they cant influence it that are unbearable. (Golf and Echo)

[954gator]

1 minute ago, noonereal said:

No I don't.

I know I don't agree with particulars but I am fine at taking it at what it is. 

It's the nuts that want to toss the whole thing because it's objective and they cant influence it that are unbearable. (Golf and Echo)

I don't think anyone wants it tossed.   We just want to argue about it. 

[golfaddict1]

54 minutes ago, NYHSFAN33 said:

Your argument is actually correct IMO, starting points are eliminated, which is why CalPreps does a good job distributing the points amongst in state, same class competition.  It is actually an excellent algorithm in that sense and pretty impressive how close it comes sight unseen in rating these teams.  However, with any Linear Program, regardless of the desired outcome, there is a finite amount of points available in the data set to be distributed.  This finite number is set and spread among every team in the country.  Some states (I.e. Cali) start with a higher number of points to spread around.  I am not sure if I am explaining it with as much clarity as I would like to.  Basically my point is that if no team played any out of state competition, the total number of points for each state would either change very nominally or not at all throughout the course of a season, the points would just be spread out among those states teams. I might be correct, I might not be, but I have yet to see much if any evidence to disprove my hypothesis.

You are correct and one look at Freeman's tinkering after the 2007 season confirms it.  States annual top ratings numbers mirror one another... it's not a coincidence.   It's not a coincidence how WI and NY can have more top 500 teams than GA if your scale starts below them!  Magically 2008, GA came alive thanks to an improved decision by Freeman.   

Please stop feeding the troll.  He's the only one who's lost and needs a pacifier.  

[Pops]

6 hours ago, golfaddict1 said:

That's one way to go... as it keeps the ratings fluid throughout and not a regular season vs playoffs and divide by 2.  I'd be interested if this model would reduce the 2010 Camden County #847 type of ratings from happening, but I don't think I can test it as there was no trend option from CP in 2010.  

2010 Camden's best win, besides MC, was a team outside the top 2,000 that they beat by 10

their rating was correct

not saying there were 846 better teams, but there were that many better resumes

dont know what right factor to use but why wouldn't there be a playoff boost if those are considered the most important games to be measured by in human polls 

[steeler01]

21 minutes ago, Pops said:

2010 Camden's best win, besides MC, was a team outside the top 2,000 that they beat by 10

their rating was correct

not saying there were 846 better teams, but there were that many better resumes

dont know what right factor to use but why wouldn't there be a playoff boost if those are considered the most important games to be measured by in human polls 

Most important games instate, not when measuring teams nationally. 

 

You don't need to add bonus points to teams to find out whose who instate. 

 

That's why you get 3A teams in Texas finishing in the top 10 in the state in calpreps. No interplay between 6A and 3A and bonus points during the playoffs. 

[CCBlackhatter]

16 minutes ago, Pops said:

2010 Camden's best win, besides MC, was a team outside the top 2,000 that they beat by 10

their rating was correct

not saying there were 846 better teams, but there were that many better resumes

dont know what right factor to use but why wouldn't there be a playoff boost if those are considered the most important games to be measured by in human polls 

We damn sure know that a game versus Clayton Valley Charter is nowhere near 210% more important, challenging, etc. than a regular season opponent like AHP, Roswell, Mill Creek, Valdosta, Lowndes, or Tucker.

Ned should deduct 10 points from every team in Norcal until one of them can play with and/or beat DLS.

At most, Ned should do this boost for a State Championship game.

 

[Pops]

14 minutes ago, steeler01 said:

Most important games instate, not when measuring teams nationally. 

 

You don't need to add bonus points to teams to find out whose who instate. 

 

That's why you get 3A teams in Texas finishing in the top 10 in the state in calpreps. No interplay between 6A and 3A and bonus points during the playoffs. 

All fair points, especially the TX 3As

i still believe, whether instate or not, those are most important games and most meaningful time to be playing your best

the factor may be overcooked a bit in algorithm but i like having "some" boost.... or maybe only include for later rounds i.e. 

[Pops]

7 minutes ago, CCBlackhatter said:

We damn sure know that a game versus Clayton Valley Charter is nowhere near 210% more important, challenging, etc. than a regular season opponent like AHP, Roswell, Mill Creek, Valdosta, Lowndes, or Tucker.

Ned should deduct 10 points from every team in Norcal until one of them can play with and/or beat DLS.

At most, Ned should do this boost for a State Championship game.

 

Is AHP in the GA playoffs?

You just compared the 5th best team in one of CA's 10 sections to a top FL team plus 5 of best teams in GA so I'll give you this one, although CVC and Lowndes would be close to a coin toss 

not sure if you'd feel same way about deductions had DLS been in GA last 25 years with 18-20 championships in a drawer in the coachs' office (they all share one)

[golfaddict1]

59 minutes ago, Pops said:

2010 Camden's best win, besides MC, was a team outside the top 2,000 that they beat by 10

their rating was correct

not saying there were 846 better teams, but there were that many better resumes

dont know what right factor to use but why wouldn't there be a playoff boost if those are considered the most important games to be measured by in human polls 

Or maybe less resumes would be better if the playoff boost wasn't in play?   

I like that we are in the nuts and bolts portion and welcome your thoughts. 

Im not ready to choose no boost, 210 pct or a tiered boost as a best scenario without hearing from others and gathering info and testing it .   

Seems to me if the ratings are adjusting weekly that playoff boost is infiltrating the regular season ratings as well no?  That may be a bit too much... 

A plus b divided by 2 is your avg. but A has B playoff boost factored in.  I've had some drinks tonight but does this make any sense?