sida1049
Well-Known Member
- Joined
- Jun 18, 2013
- Messages
- 926
- Gender
- Male
- HSC
- 2015
Hello fellow BoS'ers!
As someone who loves working with statistics and numbers, I'm interested in running an experiment. I would like to analyse the accuracy of ATAR estimates based on both yearly ranks and trials performances, as this respected forum is among the most popular methods of individual ATAR estimation.
After receiving enough (as many as possible!) ATAR estimates from this forum and my final ATAR result, I'll write up an analysis of the the overall accuracy and reliability of BoS ATAR estimates. I will also be ranking the accuracy of the estimates from users, and if I can think of some kind of a reward for the closest estimate, then that'll also be included, which would make this a competition! (Not definite at this point in time, but join in the fun anyway!)
So I invite the current HSC cohort, previous HSC students, admins, and the general users alike to join in on this experiment to hopefully gain some insight into the accuracy of estimates in general, and perhaps to the benefit of future students. The larger the sample set we work with, the more reliable the analysis will be, so challenge your friends! If anyone would like to contribute to the idea of some kind of a reward, we'd all like to hear it!
Alright, so let's do this.
School ranking: 90s/587
Subject rankings:
- English Advanced: 11/85
- Mathematics Extension 1: 2/57
- Mathematics Extension 2: 4/24
- Chemistry: 3/51
- Physics: 3/53
- Economics: 3/45
Trials performance:
- English Advanced: 89%
- Mathematics Extension 1: 93%
- Mathematics Extension 2: 81%
- Chemistry: 87%
- Physics: 86%
- Economics: 90%
Things to note:
- If possible, give a little detail about yourself, e.g. the year of your HSC, and other information which may be relevant.
- Will be taking estimates to the second decimal place so rankings will be differentiated better (so if you estimate 50, I'll register that as 50.00).
- If a range is provided, I'll take the midpoint.
- You may also edit your estimation up to the 17th of December.
- Avoid being influenced by the estimations of others.
Good luck!
UPDATE #1 (2/10/15): I propose that if we can get at least 40 ATAR estimates from different users, then I'm willing to offer a $20 iTunes or Steam gift card to the most accurate estimate...
Update #2 (13/10/15): We did it! Over 40 ATAR estimations, which makes up for a healthy statistical data set! As promised, I will be offering the choice of an iTunes or a Steam gift card to the person who gets the closest to my actual ATAR. But keep guessing guys! The more the merrier, and the more reliable the statistics will be. If there are significantly more estimates, I'll offer runner up prizes.
The competition is now closed.
I would like to thank everyone who participated in this ATAR estimation competition!
So here it is...
My ATAR: 99.30
Rankings:
View attachment 32735
View attachment 32736
Congratulations to atargainz and danpowell! I will be contacting both of you - check your private messages!
I will post up my (amateur) analysis of the data once I complete it. I'm a little busy these days, so it may take a while before I find the time to do so. If anyone else also would like to analyse these results, go right ahead and post them here or in an independent thread.
As someone who loves working with statistics and numbers, I'm interested in running an experiment. I would like to analyse the accuracy of ATAR estimates based on both yearly ranks and trials performances, as this respected forum is among the most popular methods of individual ATAR estimation.
After receiving enough (as many as possible!) ATAR estimates from this forum and my final ATAR result, I'll write up an analysis of the the overall accuracy and reliability of BoS ATAR estimates. I will also be ranking the accuracy of the estimates from users, and if I can think of some kind of a reward for the closest estimate, then that'll also be included, which would make this a competition! (Not definite at this point in time, but join in the fun anyway!)
So I invite the current HSC cohort, previous HSC students, admins, and the general users alike to join in on this experiment to hopefully gain some insight into the accuracy of estimates in general, and perhaps to the benefit of future students. The larger the sample set we work with, the more reliable the analysis will be, so challenge your friends! If anyone would like to contribute to the idea of some kind of a reward, we'd all like to hear it!
Alright, so let's do this.
School ranking: 90s/587
Subject rankings:
- English Advanced: 11/85
- Mathematics Extension 1: 2/57
- Mathematics Extension 2: 4/24
- Chemistry: 3/51
- Physics: 3/53
- Economics: 3/45
Trials performance:
- English Advanced: 89%
- Mathematics Extension 1: 93%
- Mathematics Extension 2: 81%
- Chemistry: 87%
- Physics: 86%
- Economics: 90%
Things to note:
- If possible, give a little detail about yourself, e.g. the year of your HSC, and other information which may be relevant.
- Will be taking estimates to the second decimal place so rankings will be differentiated better (so if you estimate 50, I'll register that as 50.00).
- If a range is provided, I'll take the midpoint.
- You may also edit your estimation up to the 17th of December.
- Avoid being influenced by the estimations of others.
Good luck!
UPDATE #1 (2/10/15): I propose that if we can get at least 40 ATAR estimates from different users, then I'm willing to offer a $20 iTunes or Steam gift card to the most accurate estimate...
Update #2 (13/10/15): We did it! Over 40 ATAR estimations, which makes up for a healthy statistical data set! As promised, I will be offering the choice of an iTunes or a Steam gift card to the person who gets the closest to my actual ATAR. But keep guessing guys! The more the merrier, and the more reliable the statistics will be. If there are significantly more estimates, I'll offer runner up prizes.
The competition is now closed.
I would like to thank everyone who participated in this ATAR estimation competition!
So here it is...
My ATAR: 99.30
Rankings:
View attachment 32735
View attachment 32736
Congratulations to atargainz and danpowell! I will be contacting both of you - check your private messages!
I will post up my (amateur) analysis of the data once I complete it. I'm a little busy these days, so it may take a while before I find the time to do so. If anyone else also would like to analyse these results, go right ahead and post them here or in an independent thread.
Last edited: