I want to show that if $B_t$ is a Brownian motion then $t B_{1/t}$ is a Gaussian process, i.e. that it has increments which have the normal distribution.

It seems like a trivial fact, since the distribution of $B_{1/t}$ will just be a downscaled version of the distribution of $B_t$ while $t B_{1/t}$ is scaled back again. However, I don't have any ideas about how to prove this in a more rigorous way.

I am somewhat of a newbie when it comes to Brownian motion so I would appreciate seeing a proof of this property.


Let $X_t=t\,B_{1/t}$. Every linear combination of the random variables $X_t$ is a linear combination of the random variables $B_t$ hence it is normal. Thus, the process $(X_t)$ is gaussian, in particular every increment $X_t-X_s$ is gaussian.

More generally, let $Y_t=a(t)B_{c(t)}+e(t)$. For every functions $(a,c,e)$, the process $(Y_t)$ is gaussian.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.