i really don’t get probabilities, in fact, they piss me off.
I particularly hate when people say things like ‘if you just turned up to the CFA level 1 exam and randomly guessed you’d get 33%’ or ‘if you absolutely nail 2/3’s of the syllabus and guess the remaining 1/3 on exam day you’ll get 33% on the guessed 1/3 you’ll get over 70% so you should pass’
Yes, If I alone were a large sample set of data i would get 33% on average when guessing from 3 options. But alas, I am but one man and could therefore get any fecking score from 0-100%.
Very good points. Cheaters tend to cheat on the harder questions. Those are the ones that most people get wrong and maybe will even be thrown out after the exam (I think the Modified Angoff method allows for this).
As for the overall discussion, the math argument seems more like a grammar argument to me - everybody seems to agree on the math and on what a perfect amoral undetectable robot-cheater should do. Ex-ante, the robot-cheater should cheat. Ex-post, maybe the robot-cheater would get screwed.
If I were them, I would write a program to test the outliers: probability of cheating by comparing patterns of mistakes and right answers. This would not guarantee that you could accuse them but you could tell the proctors a list of potential cheaters for the next exam.
Cheaters are more likely to have non-random patterns of errors and stumble on questions which may be easy while answering correctly questions which are not trivial.
To answer OP’s plea, you should write a letter and really apologise if it is sincere, but I welcome their decision as the exam would have no value if many people could cheat and pass. Also I think you should not post that you cheated on a public forum under your real name (Michelle …), I hope that your login is not your real name, because googling ‘michelle k… cfa’ points directly to this thread.
That would not be surprising as it would very hard to track the real poster behind this. I also made a typo in my post: I should have said “… you should NOT post … under your real name”.
it comes down to how you use always in the sentence. you are always better off cheating based on the information you have. as you will never know whether cheating was the correct move or not, the ex-ante probability is all that matters. if you can determine that cheating provides a better probability of getting the correct answer than 33% based on ex-ante probabilities, which we can based on CFA exam data collected via this site, then you are always better off cheating.
as for your scenario where a question may be so difficult that almost nobody gets it right based on their knowledge, so long as one person gets it right based on their knowledge and the point isn’t the product of chance, then the probabilty is above 33.333333%. think about it, 99,999 people guess or answer a question thinking they know the answer providing a probabily of 33.33333% for those people and 1 person gets the question right, 100% confident in his answer and reasoning, boosting the probability of cheating beyond 33.33333%. the odds of one person out of 100,000 knowing how to solve a question is 100%. this is a matter of confidence. in sample of 5 people, the act of cheating may not be better as your confidence that at least 1 person knows the answer diminishes and is no longer 100%. the bottom line is that we’re not writing the hardest exam ever written here, i’d fathom to guess that even the hardest CFA exam questions still have success rates of 40% or more as some people will know the answer for sure and many others will guess. no question will ever have a success rate below 33.3333% (ex-ante at least). you can’t fully discount the 30 million in hours of studying performed by the 100,000 test takers. the only way you’d get a question success probability low 33% is if you put a bunch of monkeys in there. finally, are we really going to depend on a cheater’s estimation of question difficulty as a factor in making a decision whether to cheat or not? since the potential cheater doesn’t know whether a question is considered “difficult” or not, the potential cheater must assume it is of average difficulty. under this pretense, you are undisputably better cheating than not. the fact that all we have to work with are ex-ante probabilities and we’ll never know whether we were right to cheat means that you’re always better off cheating.
in the context of the CFA exam which was the initial discussion, you’re always better off cheating. on a risk-adjusted basis, maybe not. in a different context, maybe not.
Matt Likes Analysis: I think there may be questions that fewer than 33% of candidates get right. Those would be situations in which the majority of the candidates may get tricked into the wrong answer.
Kartelite: I just googled “define slander”. It seems that slander as a verb is not restricted to spoken defamation. In law, as nouns, slander and libel mean exactly what you just said.
I didn’t do this as an attempt to disprove you or argue about it. I was just trying to improve my vocabulary a little bit. And I don’t even know how trustworthy Google’s dictionary is, but I thought it was worth sharing.
"noun
LAW
1. the action or crime of making a false spoken statement damaging to a person’s reputation. “he is suing the TV network for slander”
verb
1. make false and damaging statements about (someone). “they were accused of slandering the head of state” synonyms: defame (someone’s character), blacken someone’s name, tell lies about, speak ill/evil of, sully someone’s reputation, libel, smear, cast aspersions on, spread scandal about, besmirch, tarnish, taint; More "
I’m in exactly the same situation but my case has been going on for 5 months. I’m awaiting the verdict, but am sure my exam would be voided. Is it standard protocol to ban if your guilty under similarity analysis? I responded to the PCP- providing my tutors details and mentioned the hard work I put in etc. Any advice would be appreciated. Thanks