sublimity Wrote: ------------------------------------------------------- > It’s an interesting “theoretical” and > “mathematical” problem that I’ll have to think > more about, considering all factors: > - scoring of the L2 test takers before/after > adding people > - scoring of the top 1% before/adding people > - impact of 3 or 4 answer chioces > - your probability of passing > - relevance to practical problem > - etc. > > “Practically,” I don’t even care about the MPS and > I’m shooting to get >70% in all topics. I’ve set > my margin of error so high above what these > changes entail that it is irrelevant to me what > these fluctuations are and the average score is > effectively irrelevant. > > > philip.platt Wrote: > -------------------------------------------------- > ----- > > sublimity - but you have to think about the > more > > people that show up for the exam makes the 1% > > bigger, so the total number of candidates is > > actually a factor. > > > > If everyone in the whole world took the CFA > exam, > > the top 1% score would be pretty crappy > (obviously > > everyone can’t take L2/L3, but lets say > everyone > > that could take L2 did take L2, which is not > the > > case). > > > > > > > > ie if 40,000 people took L2 (roughly the number > > this year), top 1% of scores will be the top > 400 > > candidates. The top 1% candidates scores might > be > > something like a range of 88 - 92, having a > mean > > of 90. > > > > Lets say 5,000 more people think because they > > didn’t study and L2 now has 3 choices, they > will > > take it and give it a shot, making the total > > 45,000 for L2. Now the top 450 candidates > consist > > of the top 1% and their scores will probably be > > slightly lower as the number of people in the > 1% > > interval is larger. The range might be 87 - 92, > > with a mean of 89.5. > > > > These extra 5,000 people who signed up might > save > > a handful of people who would have been Band > 10s > > with 40,000 people, but now squeaked by with a > > pass with 45,000 people as the pass score is > > slightly slightly lower. ha - yeah I know you are aiming high above this and are sure to pass . . . I am just analyzing for the sake of analysis
Guys Chilll No way, Im planning to quit… I started this thread to kind of motivate people (with not so good prep ) to turn up for the exam…and thats why wrote it as A & B I am not A but B in the conversation As for my scores are concerned …I am terrified like anything…as agian I got a 73 % in Book 7 ( 1 PM) Wosrt…I hardly see any scope of improvement…whatever I can larn …I did…rest is beyond my capabilities…I dont understand anything about SWAP, Translation, Pension and Quant…so all the time keep on guessing for similar sounding words…as in I know they always try to capitalize leases…why ? I dont know…
In any case…considering Ive started in feb…Im kind of ok to accept a re-take…
I don’t want to get into a very detailed analysis (time reasons), but let’s consider a simple tweaking, concerning how the extra test takers affect things. Case 1: All extra test takers score perfect, 120/120 questions correct, then this brings the MPS up and closer to the requirements that are don’t have a curve, approaching 70% in the limit that the number of test takers is large (relative to the already existing pool). Case 2: All extra test takers score zero, 0/120 questions correct, then this brings the MPS down and helps the curve. Of course there is a continuum of possibilities between these two extreme cases. There are a lot of other complicating factors here too, for examples: - what is relationship between the scores of these new test takers and the top 1% before the extra test takers? - how large is the number of extra test takers vs. the original pool? - etc. I am not saying what you wrote is wrong nor that my original statement is correct, it is just that the effect is unclear (at least to me) and may be dependent on how well the extra test takers do - adding more test takers may either help or hurt us. It’s an interesting problem and contrived examples like I made up may stand as mathematical conclusions, that may be untrue in real world situations. Interesting! philip.platt Wrote: ------------------------------------------------------- > sublimity Wrote: > -------------------------------------------------- > ----- > > It’s an interesting “theoretical” and > > “mathematical” problem that I’ll have to think > > more about, considering all factors: > > - scoring of the L2 test takers before/after > > adding people > > - scoring of the top 1% before/adding people > > - impact of 3 or 4 answer chioces > > - your probability of passing > > - relevance to practical problem > > - etc. > > > > “Practically,” I don’t even care about the MPS > and > > I’m shooting to get >70% in all topics. I’ve > set > > my margin of error so high above what these > > changes entail that it is irrelevant to me what > > these fluctuations are and the average score is > > effectively irrelevant. > > > > > > philip.platt Wrote: > > > -------------------------------------------------- > > > ----- > > > sublimity - but you have to think about the > > more > > > people that show up for the exam makes the 1% > > > bigger, so the total number of candidates is > > > actually a factor. > > > > > > If everyone in the whole world took the CFA > > exam, > > > the top 1% score would be pretty crappy > > (obviously > > > everyone can’t take L2/L3, but lets say > > everyone > > > that could take L2 did take L2, which is not > > the > > > case). > > > > > > > > > > > > ie if 40,000 people took L2 (roughly the > number > > > this year), top 1% of scores will be the top > > 400 > > > candidates. The top 1% candidates scores > might > > be > > > something like a range of 88 - 92, having a > > mean > > > of 90. > > > > > > Lets say 5,000 more people think because they > > > didn’t study and L2 now has 3 choices, they > > will > > > take it and give it a shot, making the total > > > 45,000 for L2. Now the top 450 candidates > > consist > > > of the top 1% and their scores will probably > be > > > slightly lower as the number of people in the > > 1% > > > interval is larger. The range might be 87 - > 92, > > > with a mean of 89.5. > > > > > > These extra 5,000 people who signed up might > > save > > > a handful of people who would have been Band > > 10s > > > with 40,000 people, but now squeaked by with > a > > > pass with 45,000 people as the pass score is > > > slightly slightly lower. > > > ha - yeah I know you are aiming high above this > and are sure to pass . . . I am just analyzing for > the sake of analysis