Great Video, I have realised today's videos are just fancy, having lots of slides just to explain a single concept, while old videos explain things clearly within a min.
having a hard time understanding the concept of this model but this man just explained it to me in a simplest way. this is the best video tutorial i've ever seen! Dr. Stander, you are the best!
That's an interesting model. But of course it will always be missing the "drift" caused by large institutional traders with billions of dollars of assets under management. Who buy and sell whole sectors or large groups of stocks with buy and hold strategies for various lengths of time.
The GBM equation Xt = X0 * exp (mt + s* B) , could some one explain why this doesn't explode since exp of any mean + std * random number is going to be a reallly large number. ie (exp (11+1*.2) is > 59K
I dont think this is correct. You are claiming that X_t = X_0 *Exp( mu_hat *t -sigma_hat*B_t) is geometric brownian motion? Geometric brownian motion or GBM is the solution to the SDE , dSx = Sx *mu dt + Sx*sigma* d Bx where Sx= Sx_0 *Exp( (mu - sigma^2 / 2) *x + sigma d*Wx) ( check wiki) . Also, the linear drift system is better genearlized by fractional brownian motion i.e. d B_H(x) which is defined by mandlebrot-vonness and where H is the hurst index. Fractional stochastic calculus would better suit the purpose here since it not need be gaussian. The only real application of the linear drift system would be if the dWx was gausian and often its not. Lastly, you are using log returns which have an implied bias of the gassuian, I find percent change works better in calculating returns.
this is exceedingly inaccurate because it is well known that financial market prices are distributed with fat-tail distribution, not with normal distribution that is unfortunately assumed here.
Great Video, I have realised today's videos are just fancy, having lots of slides just to explain a single concept, while old videos explain things clearly within a min.
Have to admit, this is THE BEST video I have watched that explains the Brownian motion clearly. Thank you!
having a hard time understanding the concept of this model but this man just explained it to me in a simplest way. this is the best video tutorial i've ever seen! Dr. Stander, you are the best!
Dr. Stander, this is the best, straight to point and most well explained video about Geometric Brownian Motion on RUclips. Excellent job, congrats.
So good. The best explanation online.
brilliant video that explains brownian motion almost effortlessly! 👏 very helpful
Wonderfully explained everything
That's an interesting model. But of course it will always be missing the "drift" caused by large institutional traders with billions of dollars of assets under management. Who buy and sell whole sectors or large groups of stocks with buy and hold strategies for various lengths of time.
very didactic! cheers from Brazil!
Thank you so much.It was very helpful to understand brownian motion.
Well done, clearly explained
Hi, Prof . Its really interesting modelling. Could u shows us the coding please
Always wondered what this was in chem now im balls deep in quantitative finance with no intention to change careers lol
Great Video. Would you be able to explain how to simulate geometric brownian motion in R?
This is a great video. Please may you tell me how you did the codes. Please may you send them to me!!!
Grazie mille!
So we don't need itos lemma?
He didn't tell why there cannot be mixed values for mean and standard deviation (variance); in his assumptions
Many thanks.
Is there a link to the R code in the video?
Have you found it?
The GBM equation Xt = X0 * exp (mt + s* B) , could some one explain why this doesn't explode since exp of any mean + std * random number is going to be a reallly large number. ie (exp (11+1*.2) is > 59K
a mean of 1,100% would be too much for a stock. innit?
Can someone please provide the codes
I dont think this is correct. You are claiming that X_t = X_0 *Exp( mu_hat *t -sigma_hat*B_t) is geometric brownian motion? Geometric brownian motion or GBM is the solution to the SDE , dSx = Sx *mu dt + Sx*sigma* d Bx where Sx= Sx_0 *Exp( (mu - sigma^2 / 2) *x + sigma d*Wx) ( check wiki) . Also, the linear drift system is better genearlized by fractional brownian motion i.e. d B_H(x) which is defined by mandlebrot-vonness and where H is the hurst index. Fractional stochastic calculus would better suit the purpose here since it not need be gaussian. The only real application of the linear drift system would be if the dWx was gausian and often its not. Lastly, you are using log returns which have an implied bias of the gassuian, I find percent change works better in calculating returns.
if you work with GBM, then log prices are a must
this is exceedingly inaccurate because it is well known that financial market prices are distributed with fat-tail distribution, not with normal distribution that is unfortunately assumed here.
I can tell you for free that the markets are not on a random walk.