Kelly Criterion for Multivariate Portfolios: A Model-Free Approach

[yetanotherquant]

Just put an updated version on SSRN: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2259133
Have a look!

 

[yetanotherquant]

Wow, 3000 downloads on SSRN!

 

[yetanotherquant]

More than 4000 download and soon it gets into 1000 most popular papers!!!
I have submitted it to a prestigious journal.
If accepted, I will have to remove it from SSRN. So guys, hurry up to download!

 

[yetanotherquant]

A prestigious journal - I was pretty naive!

1. It's simply NOT TRUE that Kelly strategies cannot be used with n rather than one asset. An example is a paper published in Management Science in 1981 by Hausch, Ziemba, and Rubinstein.
(my remark: the whole paper(!) is on how to use the Kelly criterion in case of n assets, not just one)

2. As far back as the 1970s there were two papers in the Journal of Financial and Quantitative Analysis coauthored by Ziemba that dealt with the multivariate case.

3. See also THOMAS M. COVER,IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. IT-30,N O.2 , MARCH19 84 An Algorithm for Maximizing Expected
Log Investment Return 369
(my remark: I cite this paper and explain why my algorithm is better than that by Cover)

4. There are glaring errors in the abstract.

The reviewers felt that you did not have significant knowledge of the literature and that the paper was not very useful for practitioners.

With this type of feedback, I cannot consider your publication.

Thank you for giving us the opportunity to consider your work.

Yours sincerely

Frank Fabozzi
Editor
The Journal of Portfolio Management

Why do I make it public?! Well guys, as a good case study for all of you, who want to make an academic career. My case is not an exception, all friends of mine who managed to get published in peer-reviewed journals wasted too much time corresponding with incompetent editors and envious peer reviewers.
(By the way, I had no problem to get published in applied journals for professional).

But if you really have a passion for research, do it and publish on SSRN or arXiv, like Grigori Perelman did. Look, two years ago I had 4000 downloads, now I have more than 7200 (and quite a lot of positive feedback). I highly doubt that I would have so many readers in "prestigious" JoPM.

BTW, currently I apply Kelly criterion to a pretty complicated gambling system.
My closed-form approximation works poor in this case (since tails are too heavy) but my Monte-Carlo Grope works fine even in 1000-dimensional space (in my paper I analyzed a portfolio of only 7 stocks but claimed that there is not curse of dimensionality - and it is indeed so).

 

[Ryo Eng Lian Hu]

Just put an updated version on SSRN: Kelly Criterion for Multivariate Portfolios: A Model-Free Approach by Vasily Nekrasov :: SSRN
Have a look!

Thanks for sharing, reading...