A reasonably reliable method to predict points- Medicine (Other popular courses soon)

partyatmygaff

To begin with, we need to understand how the CAO works. The "Points" of a course published in newspapers are the points of the lowest-scoring candidate that received an offer for the course in question. That is to say that if there are 300 offers made then the offer made to candidate #300 will serve as a cut-off point with regard to the points required for a course.

Bearing this in mind, it is possible to obtain a somewhat accurate estimation of a course's points providing that a large proportion of applicants for the course in question take part in the "survey".

Essentially, what this spreadsheet does is take people's points and their first preference (All anonymously) and with this generates highest, lowest, mean and median points for the individual courses. If X is the number of places available on a course, it takes the X highest scores, finds the lowest of them and uses that to generate an estimation. The more people that take part, the more accurate the estimation gets.

Now, first of all there's one thing people need to understand. This method is not perfect. It may be off. It may be spot on. I don't know but it sure beats random speculation by newspapers and their journalists.

Needless to say, this survey (Unlike the one I made for the Mock) needs massive participation. It needs to spread to as many candidates as possible so if people could spread the word (Preferably before Monday ) that would be a great help to the effort. The mass uptake of this survey is absolutely vital to get a good estimation so do your best to spread the word to other applicants further afield than Boards.

I've made this spreadsheet system for Medicine but it's easily adaptable to any other set of courses such as Engineering, Business, Law e.t.c.

If there's enough demand for other courses, I can create similar spreadsheet systems for them.


Survey: https://docs.google.com/spreadsheet/viewform?hl=en_US&formkey=dGVHUHR3Tm5oRUt5SGdYZUpJX1YweGc6MQ#gid=1

Statistics and estimation: https://docs.google.com/spreadsheet/ccc?key=0AjdswhIupxjadGVHUHR3Tm5oRUt5SGdYZUpJX1YweGc&hl=en_US#gid=1


I've got the ball rolling with my own points.

Won't be that accurate as not enough people will see/fill it out

imo

partyatmygaff

AdamD wrote: »

Won't be that accurate as not enough people will see/fill it out

imo

That's why I'm telling people to spread the word further away than just boards.

There's a lot of people here and every one of us has our own network of people that we know. This is something that i'm sure a lot of people would be interested in. Everyone wants to know how many points they'll need for their course so hopefully that will entice people to participate and do their best to spread the word (Facebook, Google+ e.t.c.). If a significant number takes part then it may well be reasonably accurate.

eVeNtInE

This post has been deleted.

partyatmygaff

Don't be so pessimistic.

Just spread the word. I'll deal with any BS results.

At the very least, any estimations generated will serve as the very minimum points required.


I have the actual estimation system working now. All it needs is more people.


Edit: Removed one BS entry.

Edit2: Stupidly realised that I forgot to make the actual spreadsheet public...

Exothermic

Would you like if I enter two of my friends results? Neither of them use boards.
I know their points, HPAT and first preference, it's not something I've made up, promise! :P

partyatmygaff

Go ahead.

Edit: Working again after the brief interruption caused by a bad array formula.

QueenOfLeon

Galway added 19 places this year btw.

partyatmygaff

Updated with the correct number of places.

I can't find the number of places available in UCC however. If anyone knows, could they please let me know?

cairdre

Sorry to point out the flaws in your plan but you have not allowed for non-traditional entry routes. Different colleges allocate various percentages of places to matures, disability and HEAR applicants. This varies, as not all courses have many mature applicants.These applicants are exempt for the rating of points on courses. Therefore even though say the course has 300 places you will not know how many places went to these applicants. Therefore you cannot calculate the average?

LilMissCiara

By the time you have enough people to do it the offers will be out! :P

joshrogan

I have a foolproof method for predicting all points, wait till Monday.

partyatmygaff

cairdre wrote: »

Sorry to point out the flaws in your plan but you have not allowed for non-traditional entry routes. Different colleges allocate various percentages of places to matures, disability and HEAR applicants. This varies, as not all courses have many mature applicants.These applicants are exempt for the rating of points on courses. Therefore even though say the course has 300 places you will not know how many places went to these applicants. Therefore you cannot calculate the average?

That's a good point. There's a simple enough solution however. Taking in to account international students, DARE/HEAR students and mature students I can say that at my own estimate they form 25% of the entire undergraduate medicine cohort with varying levels of each group in each university/college. It's not a perfect method but I will reduce the number of available places in each university by that average percentage of non-standard applicants.

joshrogan wrote: »

I have a foolproof method for predicting all points, wait till Monday.

I have a foolproof method of making that comment sound less stupid. Open a dictionary and look at the meaning of the word "Prediction".

QueenOfLeon

partyatmygaff wrote: »

That's a good point. There's a simple enough solution however. Taking in to account international students, DARE/HEAR students and mature students I can say that at my own estimate they form 25% of the entire undergraduate medicine cohort with varying levels of each group in each university/college. It's not a perfect method but I will reduce the number of available places in each university by that average percentage of non-standard applicants.

From what I know, international students are not included in the figure on the prospectus. Obviously, I can only speak for one college, but of the figure given, some go into premed and some into first med, and then there are extras (maybe 30-40 or so) not included in that figure at all.

Not that this will make any difference anyway, operation beat-the-CAO won't get the 3000 stats it needs :P

LilMissCiara

GSI wrote: »

are the offers out monday?? i thought it was tuesday

Nope, Monday at 6am!

QueenOfLeon

GSI wrote: »

are the offers out monday?? i thought it was tuesday

Yup, Monday morning.

GSI

^ haha thanks, I deleted my post because i had just found out before you posted

rainbowtrout

partyatmygaff wrote: »

That's a good point. There's a simple enough solution however. Taking in to account international students, DARE/HEAR students and mature students I can say that at my own estimate they form 25% of the entire undergraduate medicine cohort with varying levels of each group in each university/college. It's not a perfect method but I will reduce the number of available places in each university by that average percentage of non-standard applicants.


I have a foolproof method of making that comment sound less stupid. Open a dictionary and look at the meaning of the word "Prediction".

88% of statistics are made up on the spot.

A former student of mine got Pharmacy on her second appeal a few years back. She was offered a deferral for the following year or start (first week of December) straight away and work like mental to catch up. She started straight away, but was told by admissions in the college that half of all places in Pharmacy for the following year were already gone from deferrals/upgrades etc.

You can't account for the number of deferrals or people who were offered it as a result of an upgrade and had to wait until this year to start.

You won't get a representative sample. Just wait until Monday. Anyway how do you determine that a result is made up in your spreadsheet?

jumpguy

Sample size too small.

Human bias.

C'mon, gaffster, remember your scientific method! This will be about as accurate as me picking a number between 726 and 745 out of a hat.

randylonghorn

jumpguy wrote: »

This will be about as accurate as me picking a number between 726 and 745 out of a hat.

How many times have you done that now? >_>

partyatmygaff

rainbowtrout wrote: »

A former student of mine got Pharmacy on her second appeal a few years back. She was offered a deferral for the following year or start (first week of December) straight away and work like mental to catch up. She started straight away, but was told by admissions in the college that half of all places in Pharmacy for the following year were already gone from deferrals/upgrades etc.

You can't account for the number of deferrals or people who were offered it as a result of an upgrade and had to wait until this year to start.

jumpguy wrote: »

Sample size too small.

Human bias.

C'mon, gaffster, remember your scientific method! This will be about as accurate as me picking a number between 726 and 745 out of a hat.

I know it's not a perfect method. I never said it was. I only created it to offer an alternative to random off-the-wall points predictions by students and journalists alike.

You won't get a representative sample. Just wait until Monday. Anyway how do you determine that a result is made up in your spreadsheet?

Results that fall outside the range of this year's HPAT results (Max and Min) are removed automatically. Entries by the same person in quick succession are also removed. Any other suspicious looking entries are investigated and removed if needs be.

Carsinian Thau

rainbowtrout wrote: »

You can't account for the number of deferrals or people who were offered it as a result of an upgrade and had to wait until this year to start.

It is very very difficult to defer a place in medicine though. They don't like to hold too many places off for too long.

In my own class, those who got in very late were told to go to pre-med. There was no mention of deferral for them. Other medschools without pre-med programs obviously can't insist on this so in those cases they're probably told to wait the year.

And even then, very few seem to get in following rechecks etc. We had 4-5 and talking to others in different years, that seems to be about average.

nothing

partyatmygaff wrote: »

I know it's not a perfect method. I never said it was. I only created it to offer an alternative to random off-the-wall points predictions by students and journalists alike.

Tbh, the best prediction method there is available is to base it on the previous years' points. (That is to say, a model built from the data of several years results, based on leaving cert results in all subjects, particularly subjects of those who commonly enter into Medicine, percentages of non-traditional students on the courses, etc etc)

Essentially what you seem to be trying to do is exactly what the CAO do, except you're hoping to do it with voluntary participation... Even if you managed to get a representative sample, you'll never be able to get the lower bound to be accurate enough to make a prediction that's any better than what you call random off-the-wall predictions. Any journalist worth their salt would actually have done some research before spouting a guess at the figure anyways.

drrkpd

Mine is 726

Konata

Honest to God, I know it's a nerve wrecking wait but it's what, less than 48 hours? The only person who can predict the CAO points is the one looking at them on the CAO computer system right now. There is absolutely 0 point in predicting anything because eh... no one knows! You may be right, you may be wrong but at 6am or whatever on Monday morning, it's not going to matter a damn.

Go distract yourself with activities for the next day - doing this sorta thing only torments people further.

mmmuir

That 890 as the highest score for NUIG looks pretty spectacular... And sadly impossible

partyatmygaff

mmmuir wrote: »

That 890 as the highest score for NUIG looks pretty spectacular... And sadly impossible

I'm shocked at the amount of people who made the same mistake despite the clear instructions in the form. It's about the tenth entry i've had to edit manually. Doctors of the future eh?

Fixed that entry anyway... Not that it matters this close to 6AM.

mmmuir

partyatmygaff wrote: »

I'm shocked at the amount of people who made the same mistake despite the clear instructions in the form. It's about the tenth entry i've had to edit manually. Doctors of the future eh?

Fixed that entry anyway... Not that it matters this close to 6AM.

Ah yeah, it's up to the Gods now. Nice idea though, although the predictions are somewhat depressing, could take off next year Papers'll be in shops at 02.30, badum-badum-badum-...

partyatmygaff

No need. http://www.irishtimes.com/newspaper/frontpage/2011/0822/1224302808028.html?utm_medium=twitter&utm_source=twitterfeed

My system was spot on for TCD. Near for RCSI. One off for UCD and close for NUIG and UCC.

A mix of a good algorithm and pure luck made it work

I always knew you guys would be eating your words

mmmuir

partyatmygaff wrote: »

No need. http://www.irishtimes.com/newspaper/frontpage/2011/0822/1224302808028.html?utm_medium=twitter&utm_source=twitterfeed

My system was spot on for TCD. Near for RCSI. One off for UCD and close for NUIG and UCC.

A mix of a good algorithm and pure luck made it work

I always knew you guys would be eating your words

After three cups of tea, I have come to terms with repeating. I'll just get 1000 points next year to redeem myself...

Allie79

partyatmygaff wrote: »

No need. http://www.irishtimes.com/newspaper/frontpage/2011/0822/1224302808028.html?utm_medium=twitter&utm_source=twitterfeed

My system was spot on for TCD. Near for RCSI. One off for UCD and close for NUIG and UCC.

A mix of a good algorithm and pure luck made it work

I always knew you guys would be eating your words

Good predicting - though I'm sure you wish you were wrong! Hope it all works out for you with rechecks and subsequent rounds.

kellogscoffey

Another doozy of a survey! Ignore those naysayers witht their facts and all that... Nice job!