O O O It's Magic (More Wombat Love)
Neri
nkafkafi at yahoo.com
Tue Oct 10 19:09:55 UTC 2006
No: HPFGUIDX 159363
> Talisman:
> <huge snip>
> Once again we are left to prod the bloody entrails of exam results,
> hoping some patterns congeal, so that we can conjure our clues from
> the stinking carnage.
>
> But, as we will see, the mysteries of Ms Rowling's "quite
> complicated" marking system, run deep.
>
Neri:
I think there's a simple explanation to the mystery. Since the
questions in many cases don't have definite answers in canon, JKR
takes it into account. She doesn't necessarily have answers that are
wrong and answers that are right. Instead she has answers that are
more right than others.
So for example, taking question 3: what's the most dangerous
dementors, inferi, hags, vampires or werewolves. JKR doesn't give
(say) "inferi" 10 points and 0 for any other answer. Instead, inferi
get 10, dementors get 9, werewolves get 6, hags get 3 and vampires
get 1.
If most of the questions are graded this way then it is possible to
get "Outstanding" with many different variations, as long as they're
good enough. In addition we don't know what was JKR's benchmark for
getting an O.
Now, I haven't actually tried to formulate this problem in algebraic
terms <shudder> but a simple algebraic law says that in order to
solve for n unknowns you need at least n equations. This means that
if we have something like 18 questions times 5 answers per question =
90 unknowns (the grade for each answer) then you'll need *at least*
90 different O-grade sets of answers to discover these unknowns
(throw in the benchmark for O as one additional unknown). That's
assuming you know how to solve a system of 90 equations with 90
unknowns. Don't look at *me*
But in general, I'd say the way you've been going about it isn't
unreasonable. If you gather as many different O-grade sets of answers
as you can find, and grade each answer by its popularity, the results
should probably be pretty close to JKR's original grading. Not
identical necessarily, but generally close. For example, your
popularity pole might find that inferi are slightly more dangerous
than demenetors while JKR originally graded inferi slightly less
dangerous, but it would still mean that she probably graded inferi
and dementors very close.
Hope this helps,
Neri
More information about the HPforGrownups
archive