ENTER / ATAR calculator (estimator)

[Daniel15]

Sweet, do you intend on making this open-source?

Maybe one day... To be honest, the code is rather messy at the moment, and I don't release code that's messy

I'm interested in how you derived scaled study scores for those that aren't listed on the Scaling Report (e.g. 33, 27, etc.)

I guess it linearly. Not the best, but works alright for an approximation
Here's that bit of the code (at this point, $subject_ids is an array of all the subject IDs chosen, and $scores is an associative array with all the raw study scores):

Code: [Select]

// Get the data on the subjects.
$result = mysql_query('
SELECT ID_SUBJECT, name, raw_20, raw_25, raw_30, raw_35, raw_40, raw_45, raw_50
FROM subjects
WHERE ID_SUBJECT IN (' . implode(', ', $subject_ids) . ')') or die('Could not perform query: ' . mysql_error());

// Loop through all subjects.
while ($row = mysql_fetch_assoc($result))
{
$id = $row['ID_SUBJECT'];
$score = $scores[$id];
// Huh?
// !!! Needs more validation here.
if ($score > 50)
$score = 50;

// Scores lower than 20 are treated somewhat differently.
if ($score < 20)
{
$lower = 0;
$upper = 25;
$step = 25;
// Hack? :P
$row['raw_0'] = 0;
}
else
{
// Score rounded down to the nearest 5 (eg. 32 --> 30).
$lower = $score - ($score % 5);
// And rounded up to the nearest 5 (eg. 32 --> 35). But don't exceed 50.
$upper = ($lower + 5) > 50 ? 50 : $lower + 5;
// "Step" (probably better called "run", or x2 - x1)
$step = 5;
}

// Gradient of the line connecting points (lower, lower_scaled) and (upper, upper_scaled).
$gradient = (($row['raw_' . $upper] - $row['raw_' . $lower]) / $step);
// Difference between "lower" score and the actual score.
$inc = ($score - $lower);
// Estimate the scaled score by getting the "lower" scaled score and adding
// the gradient times the difference to it. Imagine this as a linear graph
// with $gradient as the gradient, and the "lower" score as the Y-intercept.
$scaled_score = $row['raw_' . $lower] + ($gradient * $inc);


Edit: I use tabs for indentation. Why'd phpBB have to be screwy and replace them with spaces? I hate you too, phpBB!  :evil:

[Daniel15]

I started implementing extension study increments, as well as more/better error checking (it will tell you if you haven't put an English study, etc.). Take a look, and please report any bugs to me

[kingmar]

Very good.

I note that if you only select three subjects, it will still count the Extension Study increment (you need four I believe).

Also, you can still input negative study and point study scores. (although point scores don't do anything).

I have lobbied VTAC for an exact scaling report before to have it rejected - if we petition them, do you think they'll relent?

[reg]

I note that if you only select three subjects,

English + 3.

Also, I'm fairly sure that VCAA doesn't want people to be able to reproduce their calculations exactly, be it to prevent abuse or hide design flaws :p.

[cara.mel]

Aggregate: 209.7
ENTER: 99.90

Aggregate: 209.8
ENTER: Less than 30

Aggregate: 209.9
ENTER: 99.90

Aggregate: 210
ENTER: 99.90

Aggregate: 210.1
ENTER: Less than 30

Aggregate: 210.2
ENTER: Less than 30

Aggregate: 210.3
ENTER: Less than 30

Aggregate: 210.4
ENTER: 99.90

Aggregate: 210.5
ENTER: Less than 30

I tried putting my subjects in a different order, and then using completely different subjects, and then trying in IE and it still happens
(and no I don't think I'll get that high, I was just playing with english and spesh being near 50 xD)

[Collin Li]

I note that if you only select three subjects,

English + 3.

Also, I'm fairly sure that VCAA doesn't want people to be able to reproduce their calculations exactly, be it to prevent abuse or hide design flaws :p.

Maybe, but I think they should be completely transparent about how the system works. It will catalyse the process for change to a fairer system.

If they're trying to hide some "abuse," they should correct it. Ultimately the best scores should be received by merit, not by knowing how the system works.

[Daniel15]

Aggregate: 209.7
ENTER: 99.90

Aggregate: 209.8
ENTER: Less than 30

Aggregate: 209.9
ENTER: 99.90

Aggregate: 210
ENTER: 99.90

Aggregate: 210.1
ENTER: Less than 30

Aggregate: 210.2
ENTER: Less than 30

Aggregate: 210.3
ENTER: Less than 30

Aggregate: 210.4
ENTER: 99.90

Aggregate: 210.5
ENTER: Less than 30

I tried putting my subjects in a different order, and then using completely different subjects, and then trying in IE and it still happens
(and no I don't think I'll get that high, I was just playing with english and spesh being near 50 xD)

Oops
I was wondering what would cause a problem like that... Seems like those values were missing from VTAC's Aggregate to ENTER table. Since the aggregate was not on the table, when the calculator goes to look up the value, it ends up being set to 0.  :lol:  :lol:  :lol:.
I'll fix that soon, I'll just insert those three values into my database manually. This problem will only occur for those exact three aggregate values (and, any other values that VTAC left off their chart )

[kingmar]

No no.

I actually mean if you choose three subjects the Extension study is counted.

I.e. choose like methods, Bosnian and Chinese.

The Extension Study increment is still counted even though you didn't choose an English requirement.

[Daniel15]

No no.

I actually mean if you choose three subjects the Extension study is counted.

I.e. choose like methods, Bosnian and Chinese.

The Extension Study increment is still counted even though you didn't choose an English requirement.

My reply above was in response to cara.mel's post
And yeah, that's a bug. If you choose three subjects but don't choose an English subject (English, ESL, Literature, Language) it will still count the increment even though it shouldn't. I'll work on fixing that. Although, without an English, the ENTER won't be accurate anyways

[kingmar]

I was responding to reg's post.

But you're right. The use of this calculator requires someone to not screw around and try to rig it so they get invalid or stupid values.

Otherwise, for our purposes, it works a treat. Good job.

[Daniel15]

Thanks

But you're right. The use of this calculator requires someone to not screw around and try to rig it so they get invalid or stupid values.

Yeah, I need to add more validation (ie. not allow negative values).

And I just noticed now...

I'm interested in how you derived scaled study scores for those that aren't listed on the Scaling Report (e.g. 33, 27, etc.)

I guess it linearly. Not the best, but works alright for an approximation

I was coming up with this as I went along, and just realised that this is that linear approximation thing from Methods [f(x+h) = f(x) + hf'(x)]
Lolz, I'm slow  :lol:

Code: [Select]

$scaled_score = $row['raw_' . $lower] + ($gradient * $inc);

[elaine]

the enter calculator is an awesome idea- nice work  
gives me a rough idea of what i need in each subject

[costargh]

^^^^
There are many calculators around
A strength of this one I believe is that it is accessable directly through a site. I have one that is a program which i had to download and unzip to get it to work. Its very good thou. I think its from 2 years ago thou or last year cant remember

[Daniel15]

^^^^
There are many calculators around
A strength of this one I believe is that it is accessable directly through a site. I have one that is a program which i had to download and unzip to get it to work. Its very good thou. I think its from 2 years ago thou or last year cant remember

Can you post the URL of it here? I'd like to try and get some ideas from it

[costargh]

I can't find the Url now but I have it on my computer. I think the site may have gone down or something.