# The "three door puzzle" approach to guessing a question

You remember the “three door puzzle” - the one where there is a car behind one door and a wooden spoons behind the other 2 doors. You pick a door and the quiz master tells you another door which has the wooden spoon behind it and asks if you want to change? Since you are unlikely to have picked the right one first time, and the fact that the quiz master has told you which one it’s not behind, most of the time, the final door will be the correct answer. The conclusion being you should always change your guess. Could we use this strategy to improve your chance at the new 3 answer setup when you need to do a pure guess? This is what I’m thinking: 1. Guess an answer randomly. 2. Read the question and rule out a remaining ridiculous answer 3. Change your answer to the remaining one I assume this won’t work because you cannot know with certainty the one to rule out.

just like the movie 21… But you should be able to rule out one of the wrong answers right away, so go for it.

The reason why its 50% is because the quiz show host rules out a wrong answer. If you can rule out a wrong answer on your own, without guessing, you have a 50% shot anyway. Same principle, no convoluted reasoning.

My understanding was you’re more likely to get it right if you change from your original answer since 2/3 you were wrong with the first guess.

So after you switch, what is your chance of getting it right? 2/3? Also, what if you eliminate your guess? d11j0d Wrote: ------------------------------------------------------- > My understanding was you’re more likely to get it > right if you change from your original answer > since 2/3 you were wrong with the first guess.

i always end up with the wooden spoon !!! Used to do “ini-meenie-minie-mo” - but now Mo’s gone!!

this is the famous monty hall problem, in which marilyn vos savant correctly solved. the original riddle has a goat (instead of spoons) in two of the 3 doors. :o

Correct - 2/3 I think you’d have to eliminate an alternative after you’ve guessed for it to work for the CFA institute 3 question setup. But the problem is successfully eliminating a wrong one (and that not being the one you randomly picked first) raoul duke Wrote: ------------------------------------------------------- > So after you switch, what is your chance of > getting it right? 2/3? > > Also, what if you eliminate your guess? >

Yeah its some crazy statistics probability theory. And mathematically it works out. If you change your answer your probability goes up to 2/3 not 1/2 http://en.wikipedia.org/wiki/Monty_Hall_problem

ok well, one to remember for the exam. 1. Guess an answer (You’re probably wrong) 2. Eliminate an alternative unlikely one. 3. Chose the other one.

CF_AHHHHHHHHH Wrote: ------------------------------------------------------- > Yeah its some crazy statistics probability theory. > And mathematically it works out. If you change > your answer your probability goes up to 2/3 not > 1/2 > > http://en.wikipedia.org/wiki/Monty_Hall_problem No, it isn’t really probability theory as much as it is a brain teaser. Mathematics is more elegant than the MHP. Think of the MHP in two ways. 1) Monty offers you TWO doors for your one door. Obviously the odds are 2/3. Monty throws in the condition: Since one of the two doors is empty, I will open AN EMPTY door after you switch but before you open the door(s). This doesn’t change the odds. 2) Restate the problem with 100 doors. Monty offers you 99 doors for one, etc.

>I assume this won’t work because you cannot know with certainty the one to rule out. Wrong. This doesn’t apply to test taking whatsover. This won’t work because the rule relies on the fact that Monty knows the correct answer. Keep on reading CF_AHHHH; under the scenario where Monty (i.e. you on test day) has no idea which of the two is the right answer, the probability is still 50/50. Now, let’s get back to work on something that might actually be useful on exam day and stop trying to play silly games.

Well said on all accounts Aimee

4 me at this exact point in time there are 2 doors: - this forum - studying more

I will just study for next 10 days and get 100% on most of the answers. Those probability definitely helps you with a higher score if you know nothing but would hurt when you want to get higher score.

Aimee Wrote: ------------------------------------------------------- > >I assume this won’t work because you cannot know > with certainty the one to rule out. > > Wrong. This doesn’t apply to test taking > whatsover. This won’t work because the rule relies > on the fact that Monty knows the correct answer. > Keep on reading CF_AHHHH; under the scenario where > Monty (i.e. you on test day) has no idea which of > the two is the right answer, the probability is > still 50/50. Now, let’s get back to work on > something that might actually be useful on exam > day and stop trying to play silly games. Just pointing everyone to the Wiki-page. Not necessarily endorsing this as a strategy. I really don’t know much about it, other than what i picked up in an econ class that covered it in college (which i mostly slept through)

There is another option. If answer is like this: A) 0.5 and 0.9 B) 0.5 and 0.4 C) 0.7 and 0.4 then choose B if you dont know. Cos it has to be 0.5 because there’s 2 of them so if you were able to correctly work it out you’d get 0.5 and then be forced to work out the other one too. Similar for 0.4! i bet that works a good % of the time

Bobsters, i bet a lot of people take that approach, including myself, so they will probably make .5 the GAAP answer and .9 the IFRS answer, just to make sure we are paying close attention

Hehe, perhaps the option B should be the one you avoid then. Gives you a 50-50 chance ?