@avo & happyhappyland
Have a look here: http://vcenotes.com/forum/index.php/topic,6688.0.html
You might be interested in the maths going on there.
I never seemed to have used that excel document properly, but for me Ive worked out somethign with probability Ive learnt from methods...
Say for an easier subject like Chemistry
Get a total out of 50 so
33% of 50 for each thus each GA is worth ~16.5 marks.
You enter ur predicted score into your calc using normalCDF with say 0 to [given score] with the given SD and Mean. X that number by 16.5 and ull get something over 50. But since 40/50 isnt top 10% but top 8% I used basic chemistry-like ratios... Heres an example.
Chemistry 2009 Data
Exam 1: 114/146 GA1
Thus
Lower: 0
Upper: 114
SD: 30.1
Mean: 73.8
NormalCDF= 0.902044856
GA2
190/200
Lower: 0
Higher: 190
SD:31.1
Mean: 151.9
NormalCDF: 0.8897270136
GA2:
140/152
Lower: 0
Upper: 140
SD:33.3
Mean: 88.8
NormalCDF = 0.9340887544
add up all the percentages and average it 0.908620208
then if 40ss=0.92 (top 8%)
"x"ss = 0.908620208
39.5 ss. But you see its not exactly THAT accurate. . Only if the Mean and SD is provided for the actual STUDY SCORE itself. In which in MAO's post you stated
Mean=30 SD = 7
so if i did invNormCDF "L" (left) and the prob was .9086 with SD=7 Mean=30
I get 39.3 But remember the stats I used was
GA1: Lowest A+
GA2: Mid A+ (but mid A+ for sacs is around top 12%)
GA3 : Low-mid A+
So even with 3A+ you get around a 39-40 if one of your A+ is the lowest.