Mps opinions, anyone?

I see a real dichotomy. People with no understanding say something like “oh, you’re working on a credential” like you’re studying for the written part of a driver’s test. People with the least bit of understanding (either directly or indirectly) are like “oh, good on you. Good luck mate. Those things are tough”

Yes, even more so because the CFA Ch. is not well known outside the finance world… I just tell unknowing people I’m doing “continuing education”, but the wife doesn’t quite buy it.

I think the thing to keep in mind too is that the MPS estimates computed from 40/60/80, etc are based on ADJUSTED data so assuming your score can only be adjusted upwards (not sure if that’s a reasonable assumption) it could be possible to pass with a 63% unadjusted. After doing some calculations I’m right around 66% with a high of 69% and a low of 63% so I hope it’s enough…

I don’t want to be that guy, but the CFAI website has a new announcement saying the MPS is increased to 71% for this year for levels 1 and 2.

link/pic or didn’t happen :stuck_out_tongue:

Shouldn’t be hard for a novice researcher to go to the CFAI’s page

CFA doesn’t publish MPS scores lol.

If you clicked my link, you’d see there’s no MPS…I was hoping to egg on this ridiculous wild goose chase for a bit. Your post is the reason why any discussion of the MPS is irrelevant-- something I’ve said for a while on here…unless it comes from the CFAI. The “study” by 300 hours is absolute garbage, and other claims about it aren’t well supported. The only credible source, in my opinion, is the CFAI themselves, or someone who helped set the MPS and who has nothing to lose by disclosing that information.

There are tons of problems with trying to figure out the MPS from these bracketed scores, especially with any potential “if-then” clauses that may not be expressly stated in CFAI published documents.

Sure it’s a fun game, but you’ll never have a winner, and people talk with much more conviction than is reasonably allowed based on the available information.

Ok so I’ll throw my hat into this little game. I’m thinking it’ll be somewhere between 63-65. The lower the better of course! This is my 2nd time around as I took it last year and failed, got Band 8. Though to be honest, not sure how I pulled that off as looking at my ratings for each topic, I would’ve thought I scored in a lower band.

The only useful way to ascertain is to have a bunch of band 10 failures post their topic results with an estimation of their overall.

I was band 3 fail last year estimated approx 47-50% (I think there is a couple of percent between bands 5-10, approx 55% - 65% so I reckon 65% is approximate pass, with anyone very close being pushed over the line by a good score in Ethics, which I don’t think is possible this year…), I was not very well prepared 2016, although the exam was very fair, similar to this year in terms of difficulty, results below:

AI, Deriv, Econ, Ethics, FRA, FI: <50%

CF: 51-70%

Equity, PM, QM: >70%

Repost of Momoe for comparison (I felt very similar to how you described your exam btw, difficult Ethics, Dervis and FI, everything else very fair)

Band 9 in 2016.

Less than 50. AI. Corp F.

Between 51-70- Deri, Equity, Ethics, FI, PM

Over 71- Econ, FRA, QM.

I worked out a worst case scenario based on the above and it came out to 63%

Thanks Goldman. I hope we get over the line this year.

The problem with this is you don’t know how close band 10 is to passing. You’re riding the assumption that band 10 is right below the pass line and possibly the assumption that if “adjustments” to borderline candidates exist, that they don’t move the candidates score further from the pass/fail line.

There are plenty of scenarios for how grading is handled that make it tricky to do as you’ve suggested. The other issue stems from the qualitative data (score categories) that people are trying to turn into quantitative (numerical score to pass). Just because the category is 50-70% doesn’t make that a quantitative variable.

It’s pretty reasonable to say that the CFAI determines the MPS and a “grey area” range. Anyone at or above the MPS without adjustment passes. For the sake of context, say the MPS is determined to be 68%. People in the grey zone, say from 65-68%, excluding 68% on the upper side (so, less than 68%), are re-graded. Once people are re-graded, then any possible adjustments are looked at-- if the ethics adjustment is real, then say anyone who “does well” in ethics is bumped into passing. Anyone remaining after adjustments is classified as failing. The scores for those moved can be adjusted to reflect the classification, or they could be left as is-- that is a 66% with a positive adjustment may just be classified as passing with a 66% or the 66% may be given the points to meet the MPS. This clearly would also affect what we see. The idea behind a process like this is to have a clear dividing line where there is more of a difference between those who pass and fail, and it would show that a band 10 might not be “right there” by the pass line.

Of course, I have no idea if any of that is correct, but there isn’t any documentation on the exact process. I just think this scenario provides an example of how the process might go which further complicates these theories.

Band 10 is right below the pass line - next step after band 10 is a pass. A score band of 10 indicates performance in the top 10% of people who did not pass the exam. The categories are 50% or less, 67%, and 83% or better.

What I would find concerning is how well Momoe did, 67% in 5 topics, 83% or better in three topics, and less than or equal to 50% in only two areas (one of them being AI which is typically a smaller area of examination) and he was still a Band 9 fail.

It’s all conjecture, just interesting to see and discuss nonetheless.

Where do you get these numbers?

Less than 50 on two topics can mean as low as 0 and as high as 49 on those two topics.

Between 51-70 on five topics can mean as low as 51 on each of them to as high as 70 on each of them or anything in between. It doesn’t mean he had 67% on each.

Over 71 on three topics can mean as low as 71 on each topic to as high as 100 on each topic, or again anything in between. It doesn’t mean 83% or better. Two topics with 0 points, 5 topics with 51 points and 3 topics with 71 points as a worst case can easily end up in band 9 or worse depending on the weights of the topics…

FYI, I found this topic by chance

https://www.analystforum.com/forums/cfa-forums/cfa-level-iii-forum/91353345

and

I don’t know much about the 40/60/80 scores, But according to firstandonlypost , the MPS is 60 (%)!!!

Honestly, I don’t believe it. But maybe what he said is true.

I would look at that more as he was at the upper end of the 50-70 range for topics in which he scored in that range rather than the MPS is 60.

I thought 60 (%) was his score estimated by 40/60/80 method? So, the MPS according to his result and the 40/60/80 must be at most 60%, right?

ps: yes, if we take into account the error of the 40/60/80 method, MPS will be higher.

isn’t it said that if you pass you pass with flying colors, and if you fail, you fail with flying colors.

in other words the probability of someone passing/failing by exactly one wrong/right answer is pretty much zero.

No it just gives you a base line estimate based on the ranges you input from your score matrix. For example, assuming a passing score and getting 50-70 on every single topic it would estimate my most likely score to be a 60. That doesn’t mean that I passed with a 60, it more likely means that my actual score within the 50-70 range was probably closer to 70 than it was 50 given that the general consensus is that a passing score is mid 60’s.

I haven’t seen it in anything provided by the CFA Institute, but I might’ve missed it. Can you provide a source for the bold statement? Otherwise, I think it’s possible that band 10 is just below the pass line, but I also think there are several scenarios where there is an appropriate buffer between passing and band 10, such as the scenario I detailed in my last post. I think it’s entirely possible that an initial MPS is set at x, and people between y and x are retabulated and evaluated for potential adjustments that push them over x to pass or below y to fail. This would mean the band 10 doesn’t necessarily indicate “just below the pass line.”