Morgan Stanley Traders Lost $390 Million in One Day in August

[Jaxon]

Quote from BJL:

Or if you are like the OP you just stick with your original choice

you know, I really am sympathetic(not actually the OP) because I had a hard time accepting the fact that you should always change your choice with a 3 cup (or 3 curtains with Lets Make a Deal) problem. It wasn't until I expanded the problem to 100 choices that I saw the logic in switching. Took me a while too.


ok, I finally read through most of this tortuous thread and I see that my example using 100 or 1000 cups was presented several pages ago. so.... nevermind.

 

[Capablanca]

Quote from Batman28:

at least someone here knows how to calculate probability..

u make 2 decisions.. first select one of the three, and then to switch or not..

I wanted to show to some of the others that how you came up with the 1/6 probability was understandable. Still after reviewing the Monte Hall wikipedia link I think it makes a convincing case that switching improves the odds.

Perhaps it will help to think of the switch not in terms of switching cups but in terms of switching odds.

The "switch" offered is actually between your original odds of

1/3*1/2 = 1/6

and the newly proffered odds of

2/3*1/2 = 1/3

1/3 odds is greater than 1/6 so you should switch.

It might be useful to note that, presuming knowledge on the part of the facilitator, revealing that one of the cups does not contain the ball doesn't change the odds. One of those cups was always going to be empty. It's just a distraction.

How exactly this relates to trading I'm not sure of though

 

[GTS]

Quote from Capablanca:

I wanted to show to some of the others that how you came up with the 1/6 probability was understandable. Still after reviewing the Monte Hall wikipedia link I think it makes a convincing case that switching improves the odds.

Perhaps it will help to think of the switch not in terms of switching cups but in terms of switching odds.

The "switch" offered is actually between your original odds of

1/3*1/2 = 1/6

and the newly proferred odds of

2/3*1/2 = 1/3

1/3 odds is greater than 1/6 so you should switch.

The original odds of picking the correct cup are 1/3. The odds of getting the correct cup if you switch your choice after an empty cup is eliminated is 2/3. The odds are never 1/6.

 

[youngtrader]

Ok somebody earlier mentioned taking the other side of this trade (its way more fun to go against the grain). I think this all is total BS. You guys can't try to quantify every piece of the market. It can't be done because it has no rules and anything goes. If that guy is telling the truth about that interview then I would say the guy looking to higher him is the dumbass for calling the interview over because of such a gay question like that.

The market only has 2 cups. You can go long or short nothing else. There is no option C. Whether I go long or short I still have a 50-50% chance no matter what. So the cup theory does absolutely no good here. It basically comes down to gut instinct, crowd phycology, and fundamentals.

Suppose market is rallying hard but you think a tops got to be coming up here pretty quick. Your thinking is we are coming up to some resistance on the daily chart and have had kind of a run away buying spree over the last few days, with a report out at the end of the week maybe guys will take some money and even up for the report? Finally market starts to show topping action and all of a sudden bids are being hit with good size but bids appear to be holding their ground. Do I go long and take advantage of the setback as a tremendious buying opportunity or do I sell this thing short?

You have no time to sit there with a pen and paper and try to figure out the probability of what I should do next. You have to act and act fast in order to capatilize on the markets mistake. Thinking on your feet and at times shooting from the hip with good money management is the key to success not crunching numbers and plugging them into an automated system based on physics and statistical probabilities.

This of course is my view and I am probably in the minority but try to understand and think about what im saying here. Seems like the people that use the quant techniques are the people who have never encountered anything like the markets before. They are used to encountering things thats actions can be explained by rules and formulas. These people in my opinion do not make good traders (probably not bad anaylists though). Don't try to make total since out of the market because you never will. Like they say "You don't have to know why the market is doing what its doing you just have to be able to profit from it."

 

[Equalizer]

Dood, this was simply an interview question to test Batman28's analytical thinking. Pure and simple. Do not conjecture upon market connections and modeling and sheah. IBs ask probability/combinatoric/brain-teaser type questions for quantitative related roles, as well as the general-finance/stoch-calculus/esoteric-C++ stuff. They want to see how you think - outside the phreakin' box.

Sheah, they can ask you how many grains of sand are on the beach. Of course you can't give them an exact answer, they do however want to observe your thought process, how you'd go about estimating, what problems/difficulties you can flag, etc, etc. A simple "you can't be serious" response won't cut the mustard.

 

[Capablanca]

Quote from GTS:

The original odds of picking the correct cup are 1/3. The odds of getting the correct cup if you switch your choice after an empty cup is eliminated is 2/3. The odds are never 1/6.

I get the feeling you are unfamiliar with the use of decision trees in statistics.

The chances of you picking the correct cup is 1/3 true but the chances of you picking the correct cup and choosing not to switch as opposed to switching is 1/6.

The odds of getting the correct cup if you switch your choice after an empty cup is eliminated is 2/3.

2/3? How did you come up with this figure? If we are presuming an independent trial between 2 cups then the chances are 1/2. But we have a dependent case. How can the chances be greater than 50%? Are you saying undergoing a previous choosing actually increases the chances? Logically the extra step should decrease the chances. The correct probability in this case is 1/3.

 

[GTS]

Quote from Capablanca:
The chances of you picking the correct cup is 1/3 true but the chances of you picking the correct cup and choosing not to switch as opposed to switching is 1/6.

The decision to switch or not switch is not subject to "chance". Saying that the odds of picking the correct cup and not switching implies that you dont have control over the decision to switch which is false.

You can enumerate all the possible choices (picking the correct cup to begin with, not picking the correct cup, switching not switching) but that has nothing to do with the odds. The odds as I stated in my previous post are correct.

Quote from Capablanca:
2/3? How did you come up with this figure? If we are presuming an independent trial between 2 cups then the chances are 1/2. But we have a dependent case. How can the chances be greater than 50%? Are you saying undergoing a previous choosing actually increases the chances? Logically the extra step should decrease the chances. The correct probability in this case is 1/3.

Seriously man, if you can't comprehend why switching after an an empty cup has been eliminated gives you a 2/3 chance of picking the correct cup then do yourself a favor and go read the Wiki link on the first page of this thread. There is no reason to keep re-explaining this a hundred times.

If after reading the entire Wiki article you still don't get it then please come back here and say so. Please also revisit the case (listed a few times already in this thread) where there are 100 cups instead of three and 98 of them are eliminated (all empty) and you decide if you want to keep your original guess or the other remaining cup. If thinking that through that doesn't create an ah-ha moment for you then I don't know what will.

 

[Capablanca]

I've read the aforementioned article GTS. Have you?

 

[GTS]

Quote from Capablanca:

I've read the aforementioned article GTS. Have you?

You read the article where is says in big bold print "Why the probability is 2/3" and yet you still believe that the correct answer is 1/3?
http://en.wikipedia.org/wiki/Monty_Hall_problem#Why_the_probability_is_2.2F3

What the article says has a 1/6 chance of happening is that you originally picked the correct door (cup) AND the host shows you "Goat A". It also says that there is 1/6 chance that you happen to pick the correct door (cup) and the host shows you "Goat B". The sum of these two is 1/3. Note that it is irrelevent to the odds of picking the door correct originally whether the host shows you Goat A or Goat B - that is just an enumeration of all the possible events.

When you and batman said that the original odds were 1/6 you both were wrong. 1/6 refers to a very specific circumstance where you pick the right cup/door and the host reveals a specific (wrong) door.

Let me help if you are forgetting what you said:

Quote from Capablanca:

The "switch" offered is actually between your original odds of

1/3*1/2 = 1/6

and the newly proffered odds of

2/3*1/2 = 1/3

1/3 odds is greater than 1/6 so you should switch.

Everything you wrote there is wrong. Plain and simple.

 

[Capablanca]

Went back and read the article again. I see an early paragraph says the chances are 2/3. The phrasing and posing of the problem and the explanation in that portion are unfortunate. It should be clarified.

Actually everything I wrote is correct.

The perspective is what is misaligned. I'm looking at it in finer detail.

1/6+1/6 = 1/3

1/3+1/3 = 2/3
______ ____

1 = 1