a to b is 2x, a to c is 3x, and a to d is 4x, use the hs gpa formula for grade

effect = (GPA team 1 - Gpa team 2)+1

the difference between A+ and B+ in D1 is at least 2x. 
IMO, no further discussion on that topic is necessary.

I would guess D1 is somewhere along recruiting effort per prestige:   A+ = x10;  B+  x5;  C+  x3,  D+   x2

Posted by oldresorter on 2/3/2011 2:26:00 PM (view original):

Posted by jetwildcat on 2/3/2011 1:08:00 PM (view original):
 OR's 2x model just doesnt make sense to me, whether its linear or multiplicative (i'm not sure what he means)

for example

if D is our baseline (1x), D+ is 1.333x, C- is 1.667x, C is 2x

say you want to compare D+ to C-, you're talking about a ratio of C- = 1.25xD+. then C=1.2xC-. it just wouldnt make sense anymore.

if it truly goes up in a linear fashion, and A=4x while A-=3.667x, you're dealing with a difference of 1.09x. I REALLY find that difficult to believe, that the "average" A- is that close to the "average" A

if it's 2x and it's multiplicative, then it would be 1.26x each third of a grade (if D=1, D+=1.26, C-=1.59, C=2). this is also possible but it seems a little severe. I can think of at least a few instances where i've tried to bump off a D prestige sim with a B d1 team...if i had a 4x prestige advantage, there's no way i'd have had as much trouble.

despite popular opinion of a lot of successful coaches i'm not sippin the kool-aid here.

 I can't make out the math you are trying to express, and I am pretty lucid when it comes to math.

why would anyone lie to you or anyone about this - sippin the kool-aid? - I don't get what you are implying?

figuring out the real ratio in d1 is fairly easy due to how many sims recruit, I had even forgot how I did it the last time until I read this thread, I will verify the 2x, last time I checked, it was 2x for a full grade, 3x for two grades, 4x for three grades, but if someone told me it was 1.7, 2.4, and 3.1, it would not shock me either, too bad, I just got done recruiting in a world where I probably could have gotten an exact answer as I took on several sims, but another chance will come.  My gut says the 2x / 3x ratio was accurate, as I seem to recall I was considered and went ahead at almost the exact points I would have thought, but I really was not paying that much attention at the time to nail it to the dollar, or to the % in this case.

For anyone with the will to do so, just take out a sim competitor, do the math, after a few tries, the ratio should jump out at you too. 

I don't play much d2/d3, but I would have to prove to myself the ratio's are the reason for the difference and not the equality effect of promises, as there is just very little reason to program prestige different in d2/d3.   nor is it logical to program the effect of promises different, hence, since money is way different in d1 vs d3, prestige gets trumped by the fixed value of promises, but I am guessing at this, I will try to verify it for myself, although at one time, few if any played d3 better than I did, so I am not a complete dolt at d3 either.

OR-
can you elaborate a bit on the part where you say
"For anyone with the will to do so, just take out a sim competitor, do the math, after a few tries, the ratio should jump out at you too. "

i don't see how you translate competing with sims to prestige values? it seems like you would have to be fighting sims with the same # of recruits, and be confident who their #1 is to have a chance... is that what you are looking at?

NO - I don't think I will, I probably am wrong - LOL

two of my conf mates were in a battle for a player, that just finished. they posted their effort on the CC so i feel its ok to repost it here -
a- school, 37 campus, 10 home, 10 evals, start, 20 mins
b school,  43 campus, 21 home, 1 eval, start, 15 mins, some letters (im guessing about 1-1.5k worth, he said he maxed it out)

the b school also appears to have had some considering credit advantage.

in this case, the b school won, at the 11am cycle the day after signings (read: kind of close, but not very close)

the prestige advantage of an a- vs b school could theoretically be as small as .34, and as high as 1. looking at their history, both schools appear to be on the low end - if they were equally low, that would be a .66, which is the expected different of an a- and b school.

with a .66 advantage, the effective credit by the a- school in b school terms would be:
61.6 campus, 16.7 home, 16.7 scouting, 1.67* start, 1.67*20 mins    vs
43 campus, 21 home, 1 eval, start, 15 mins, some letters, some considering credit

seems to me like a big win for the first school, no doubt about it, in the average case.

going with a pretty conservative value of .5, the a- school in terms of the b school would be:
55.5 campus, 15 home, 15 scouting, 1.5 * start, 1.5* 20 mins   vs
43 campus, 21 home, 1 eval, start, 15 mins, some letters, some considering credit

i would pretty strongly expect the a- school to win in that case as well, and if the b school did win, i very much doubt it would have been by enough to sign the player at 11am. and based on the resume of the B school, i am very reluctant to believe this is a minimum case of a .333 difference, i think .5 is a pretty reasonable figure, significantly lower than the average case.

using a prestige value of .5, in the average case of a .666 prestige difference, using a geometric progression, we get an advantage of 31%. so that translates to:
48.5 campus, 13.1 home, 13.1 scouting, 1.3 * start, 1.3 * 20 mins    vs
43 campus, 21 home, 1 eval, start, 15 mins, some letters, some considering credit

in this case, i feel like its pretty close to a wash. at 2:1 hv:cv, the a- school has roughly 3 more home visits, 12 evals, .3 starts, a bit on minutes, vs considering credit and some letters. considering credit could be as high as 10%, which would be like 10 home visits for the b school, which would probably put them ahead slightly. but, i feel like that is probably liberal - in favor of the b school. probably, the a- school would win. so i feel like even the 1.5x prestige multiplier puts the a- school ahead here, in the average case.

now if you go with that same, lower end .5 prestige grades figure, thats a 25% advantage. which is,
46 campus visits, 12.5 home, 12.5 eval, 1.25* start, 1.25 * 20 mins    vs
43 campus, 21 home, 1 eval, start, 15 mins, some letters, some considering credit

now, finally, this seems to line up. the schools are roughly tied on campus/home/evals. considering credit at 10% would probably be enough to overcome the slight mins/start advantages and could justify the 11am day after signings signing. still, it seems like the maybe 10 home visits ahead this leaves the B school is only questionably enough to cause a signing at 11am, but i suppose its pretty close. but anyway, what i am getting at is, even using the conservative .5 prestige grades figure, it *still* seems like 1.5x prestige over states the prestige advantage slightly. so maybe it was only .45 prestige grades or something. but, 2x really doesn't seem to fit to me. anybody see anything i missed/messed up?

Posted by oldresorter on 2/3/2011 3:01:00 PM (view original):
a to b is 2x, a to c is 3x, and a to d is 4x, use the hs gpa formula for grade

effect = (GPA team 1 - Gpa team 2)+1

that doesn't make sense cuz if you have 3 teams in a battle the ratio doesn't work out. lets say an A competes with a B and a C. by your system, the A is 2x the B and 3x the C. the B is also 2x the C. This should make the A now 4x the C (2x the B and 2x the C) but by your system he's somehow only 3x the C?

to elaborate again

if team A prestige puts in 100 "recruiting effort" points and team B puts in 180 "recruiting effort" points, A should beat B.
if team B prestige puts in 180 and C puts in 350, B should beat C
but if team A puts in 100 and team C puts in 350, C should beat A? i thought C just lost to B who lost to A!

jets great point, I have not dabbled in this in quite some time & really am not too concerned about what C or D schools are doing, if that makes any sense, but what you point out is 100% true, the real formula is actually this:

effect = (prestige team 1 - presstige team 2)*2

note if billy's 1.5 is accurate, then the formula would be effect = (pr team 1 - pr team 2)*1.5 or lm2's 1.7

I will do a little testing and verify the base rate, i.e. what one grade is worth, as well as what 2 grades or maybe even 3 grades are worth, there are so many sim schools, it should not take me very long, I might even start a thread on my experiment

billy, in your example, the base rate appears to be 1.25, for 2/3's a grade difference, which would translate to a full grade at 1.37, what conf is that from, I have a few ?'s about what your boys posted, as well as the timing of the offers, etc - I'd just like to take a quick look as long as you guys got me started on this.

Posted by oldresorter on 2/4/2011 10:08:00 AM (view original):
jets great point, I have not dabbled in this in quite some time & really am not too concerned about what C or D schools are doing, if that makes any sense, but what you point out is 100% true, the real formula is actually this:

effect = (prestige team 1 - presstige team 2)*2

note if billy's 1.5 is accurate, then the formula would be effect = (pr team 1 - pr team 2)*1.5 or lm2's 1.7

I will do a little testing and verify the base rate, i.e. what one grade is worth, as well as what 2 grades or maybe even 3 grades are worth, there are so many sim schools, it should not take me very long, I might even start a thread on my experiment

billy, in your example, the base rate appears to be 1.25, for 2/3's a grade difference, which would translate to a full grade at 1.37, what conf is that from, I have a few ?'s about what your boys posted, as well as the timing of the offers, etc - I'd just like to take a quick look as long as you guys got me started on this.

OR - in my example where i used a 1.25, that wasn't the base rate - that was a rate of 1.5x assuming a prestige difference of .5. im not sure how i got 1.25, that is wrong, really the formula i use is -
prestige grade = 1.5x
1/3rd grade = 1.5^(1/3) = 1.1447
.5 prestige difference = 1 + 1/2 partial grades = 1.1447^(1.5) = 1.22

i was just sloppy there, i was thinking half of 1.5 because i was using half grade prestige difference. i also showed it for the avg case of a .667 prestige grade difference. the first 2 examples i went through were at 2x and the second 2 at 1.5x, hopefully that wasn't too confusing or whatever.

the posts are in the sec in tark, georgia/auburn

no the 1.25 came from me, not you - or lets say the 1.37, it was the estimated effort ratio of team A- vs team B - which implies the 2x that I use, is not only wrong, it is wrong by a huge amount.  We'll see, I'll do a little analysis and see what I come up with.

The coach of team A- contacted me directly, he is going to give me further info.

oh, i see OR. i think 1.37 is about right. honestly in the average case 1.5x just doesn't justify it either, i was expecting it to - looking at the two teams, i felt georgia's B (its a b not b- not sure if you typoed there or ?) was a lower  B probably than auburn's A-.

so OR -

effect = (prestige team 1 - presstige team 2)*2

just to make sure we are on the same page, are we agreeing the growth is multiplicative, not linear? i.e. if it is 2x, a->b = 2x, a->c = 4x, a->d = 8x?

it has to be, if you follow jet's very clear example above:

A prestige at 100 effort credits
B prestige at 180 effort credits
C prestige at 350 effort credits

a linear scheme would have a beating b - 100*2 vs 180
while a would lose to c - 100*3 vs 350
yet b would beat c - 180*2 vs 350

in a multiplicative sense, A has 400 points, B has 360, and C has 350 and the world seems right

a great case for the formula is multiplicative and not linear, I honestly have not paid that close attention to this issue for a long, long, long time

now we just need to come up with the right multiplier

OR, a linear scale could work consistently.  E.g., suppose the scale is D = 1, C = 2, B = 3, A = 4.  Then in jet example the adjusted credits are

A prestige 100 raw => 400 effective
B prestige 180 raw => 540 effective
C prestige 350 raw => 700 effective

In this case the C prestige team wins over both, and B beats A.

Note that I am not saying it is linear, just that that would work.

What I've seen on the forums points to a "multiplicative" scale (which is really exponential, but let's not quibble) as most people seem to agree that the ratio between A and B is at least the same as the ratio between B and C.

hannibal, in that case, the prestige difference of a grade would change between every two grades. that is linear in the way that each grade goes up by 1, but the people who talk about it being linear mean like, if a to b is 2x, a to c is 3x, a to d is 4x. i don't think you could get a model like that, based on what jet is showing. i do think a model like yours is possible. but like you, i don't think it works that way (from experience).

 

Posted by _hannibal_ on 2/4/2011 1:04:00 PM (view original):
OR, a linear scale could work consistently.  E.g., suppose the scale is D = 1, C = 2, B = 3, A = 4.  Then in jet example the adjusted credits are

A prestige 100 raw => 400 effective
B prestige 180 raw => 540 effective
C prestige 350 raw => 700 effective

In this case the C prestige team wins over both, and B beats A.

Note that I am not saying it is linear, just that that would work.

What I've seen on the forums points to a "multiplicative" scale (which is really exponential, but let's not quibble) as most people seem to agree that the ratio between A and B is at least the same as the ratio between B and C.

that it is,

interesting part of your analysis, a mythical tie exists if A spends 100, B spends 133, and c spends 200, so although the equation may be linear, the level of effort 'steps' in this case may not be.

given that billy's conf mate's example yielded a full letter grade ratio in the 1.37 to 1.40 range, for a A- school vs a B school, makes you wonder?

I have ample info to go try to figure out the equation with sims

If anyone has data from human vs human battles they want to send me, feel free, I will probably start a seperate thread in some conf board documenting what I find.