Say we have an asset that's 100 dollars on day 1. Now, every day, the price can either go up with 30%, or down with 25%, and both price movements have the same probability. What's the return after 4000 days?
Jeremiah Rogers
0?
Luis Nguyen
It depends, anything between 0 and 59348429003252308012834107202943920995257834690948195302156515900662637463168195 96632372919311696898354989794024626401712625896567585218346637688404219389802400 12479766368170756278731995105816861733049592429788853386702024392783945262681472 63983630206546955261697123386215747025011852513255712970874721929085187613702834 04917002657872134725197214970767027400738832483718437245392551891447407827118917 28493694787664017746930975541845856473663472265145982596563147315023225340971807 32398642429911996383922112507100967950111696400348805222163343358133290412008075 75051735993536409699595787715273699720910894780883669041179871059308419290131006 58277997669987336521977748550859446277085027190956590583065047520779562582989400 42162755506157655068441844747181750318059183134394041564271476901426898626717928 46384860863593957636990756956222057130203171121125770501864193850532554284062888 73542161870749600404582923541829782593606950052371949624569932190690832372061896 80638338772542621312720765663875198696621881828315452342636768954945522390174162 12702159217304615341780471236906484630781964556529589632130604414573568399642526 05053352327717366620856745554553599277771034848622053942630506257040448455542415 66397488130924046976419473497706740683087315839907708420949872205420811755834222 13532571368272732688193733956236809804796710968997887734962740685069671612229759 09390149137712779802174841911778096932750209417090204782679734013976177623582185 (cont)
>You can't calculate the return because in some scenarios your asset will go to negative value, which is counter-factual. U are a realy stupid idiot.
Dominic Bell
(OP) >What's the return after 4000 days? You can't calculate the return because in some scenarios your asset will go to zero and stay there, which is counter-factual. The best that can be done is to calculate an expected rate of rate based on a series of test cases, something akin to a Monte Carlo analysis.
But, I'm not doing your summer school homework for you, so do the math yourself.
Michael Clark
>which is counter-factual. But it isn't. Pay good attention to the price of Trumpcoin in the following months if you want an example.
Eli Perry
Shut up coinfag.
Jonathan Rodriguez
Anywhere between 0 and 100(1.3)^4000
Also, if for some reason the up down movement happened same amount of times 100((1.3)(0.75))^{2000)
Which basically ends up at 0 anyway
Jaxon Howard
I just did this simulation of 50 prices, and they all go to zero in the long run.
Lincoln Allen
As I said, that's expected for many cases, since adding 30% to zero in any case when you get to zero leaves you at zero. However, if it's happening for every case, you either have a too small sample size or your model is setup incorrectly.
Blake Ward
Which softrware did you used?
Wyatt Harris
Just thinking about it logically, you're going to most likely end up at 0 well before 4000 days.
Nathaniel Davis
MATLAB.
Thinking logically, it always goes to zero after thousands of days.
Michael Gutierrez
>Thinking logically, it always goes to zero after thousands of days. Thinking logically, there are scenarios where the positive days predominate and the result must necessarily be greater than zero.
Grayson Davis
You mean, the >expected Return. Also anyone that wants to learn stochastic calculus for financial shit here's a textbook.
Brownian Motion Calculus - Ubbo Wiersema
Lincoln Long
>hurr durr anything can happen
Here's the deal if I ask for the "expected" return, and exactly what's bugging the shit out of me. We'll have the following:
X0 is our start price on day 0, and Xn is our price on day n. The price movement on day n is a random variable N_n.
For day i we have: Xi = X0 * N_1 * N_2 * N_3 *.... * N_{i-1}.
So the expectation for Xi would be:
E(Xi) = E( X0 * N_1 * ... N_{i-1} )
With independence follows:
E(Xi) = E(X0) E(N_1)^{i-1} = X0 * 1.025^{i-1}
So when the i goes to infinity, the expected value goes to infinity.
However, the price is most likely to be somewhere around 0.
Also thanks a lot for that book recommendation, I appreciate it.
Julian White
>hurr durr anything can happen That's what Monte Carlo simulations are about, dumbass. Anything can happen, and you're trying to figure out the likelihood.
Fucking retard.
Brody Johnson
bayes says $5.7173698e+86
your model is not realistic.
Austin Collins
this is a GS interview question.....just sayin....on inteview 10.
Dominic Sanders
Statistically: 270 dollars, inflation included.
William Nguyen
Actually, using the central limit theorem I just found that the probability of the price being 600 dollars is just 0.0013.
Also here it is with 500 trials.
Funny how so many Veeky Forumsnessmen seem to have trouble with elementary problems.
That's where I got it from, the right answer is 0 indeed.