• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

Stripping down the robo-advisors: sparrow-brains inside

Hi guys,

have a look at my critical review of robo-advisors.

Summary:
  • Robo-advisors promise the risk profiling in a few easy steps, which is unrealistic both from mathematical and behavioral points of view.
  • The “optimal” portfolios are usually based on Markowitz-like models, which are inapplicable in practice due to their extreme numerical sensitivity to the market parameters estimation errors.
  • Robo-advisors lure investors with low management fees but minimizing fees and maximizing the wealth is not the same. Moreover, the compound costs are not so small in the long term.
  • A positive side: Robo-advisers do not (yet) foist toxic financial products upon you.
 
A follow-up about power and limits of the diversification.

Triggered by this discussion on NP.


Summary:
  • Sometimes (esp. to fool inexperienced retail investors) the diversification is claimed to be a silver bullet (even in a financial crisis). I show that in crises the diversification effect weakens significantly but still persists (esp. for “defensive” stocks).
  • I argue that in a normal (non-turbulent) market the diversification is very helpful in theory but also critically consider its applicability in practice.
  • The results that we obtained for the DAX / German stock market should be extrapolated with caution for other markets. You will also see why it is better to watch and know the market (rather than to blindly rely on quantitative analysis and common sense).
 
  • The “optimal” portfolios are usually based on Markowitz-like models, which are inapplicable in practice due to their extreme numerical sensitivity to the market parameters estimation
Yo, this is an interesting point. How are you optimizing your portfolio then? AFAIK MPT is still used in the industry.
 
Yo, this is an interesting point. How are you optimizing your portfolio then?
Kelly criterion is my starting point. But numerically it is as unstable as Markowitz, still it is useful to understand (and to quantitatively estimate) the danger of overbetting.
Moreover, since I am pretty risk averse and my Kelly fraction is relatively low, numerical instability is not so critical for me.
And I surely beat the DAX, at least so far, but since the time series of my returns on trades is pretty homogenious it is not implausible to assume that it will be so also in future
 
Kelly criterion is my starting point. But numerically it is as unstable as Markowitz, still it is useful to understand (and to quantitatively estimate) the danger of overbetting.
Moreover, since I am pretty risk averse and my Kelly fraction is relatively low, numerical instability is not so critical for me.
And I surely beat the DAX, at least so far, but since the time series of my returns on trades is pretty homogenious it is not implausible to assume that it will be so also in future

Do you invest in stocks outside of DAX? Utilize any non-equity securities? Risk-adjusted performance?
 
Do you invest in stocks outside of DAX?
Yes, but the share of non-DAX stocks is relatively small.
Not because I explicitely promised it but rather because DAX is my benchmark and I have to (more or less) stick to it.

Utilize any non-equity securities?
Yes, first of all the oil (currently it is pretty easy to trade with).
But also in relatively small proportions.
In attached pdf you can see all assets I ever traded and the box-plots of returns on them.

Risk-adjusted performance?
In either case better than by DAX, since I beat it BOTH in terms of return AND and volatility/drawdown.
The complete statistics is available here:
https://www.wikifolio.com/de/de/wikifolio/999ducks
 

Attachments

  • VasilyNekrasov_SomewhatBetterThanDUCKS_BoxPlotsOfReturnsPerISIN.pdf
    53.5 KB · Views: 11
Kelly criterion is my starting point. But numerically it is as unstable as Markowitz, still it is useful to understand (and to quantitatively estimate) the danger of overbetting.
Moreover, since I am pretty risk averse and my Kelly fraction is relatively low, numerical instability is not so critical for me.
And I surely beat the DAX, at least so far, but since the time series of my returns on trades is pretty homogenious it is not implausible to assume that it will be so also in future
Isn't Kelly just Markowitz?

"...So there, the tangency portfolio is the same as the Kelly optimal portfolio, up to a normalization constant, and without telling us what the optimal leverage is."

Quantitative Trading: Kelly vs. Markowitz Portfolio Optimization
 
Some time ago you claimed that you are making profit because you can tolerate high drawdowns. If this was the case, how are you still being able to show better volatility than the index?
 
Isn't Kelly just Markowitz?
Purely numerically mostly yes (but sometimes still no)
http://www.decal.org/file/1071

The advatage of Kelly is its economic insight. Markowitz is one-period model, whereas portfolio management is essentially muliperiodic.
And there is no plausible idea of "fractional Markowitz".

Some time ago you claimed that you are making profit because you can tolerate high drawdowns. If this was the case, how are you still being able to show better volatility than the index?
Did I?! A link to this post, plz!
 
Last edited:
That's what I said:
For example, I tolerate larger drawdowns for highly volatile stocks (otherwise I would cut the losses too quickly - i.e. just trade noise).

How do you conclude
that you are making profit because you can tolerate high drawdowns.
?!

What I mean is that (if the risk/reward is good) I CAN buy a highly volatile stock and tolerate a relatively high drawdown by THIS stock (but it doesn't mean that I accept a high drawdown for the whole portfolio). Moreover, if I buy a highly-volatile stock, its weight it portfolio will usually be small (Kelly tells me does).
Thus I neither miss the opportnities nor let the drawdown go out of control.
 
Here is a cocrete example: yesterday I recommended to buy K+S stock, which is very volatile.
By that time I, myself, held 4% of my capital in it, the position drawdown was about 5%.
Today the stock grew about 2.5% and I reduced in K+S my position to 3%.
I will likely reduce it to 2% by the next opportunity but this 2% will be held for the mid/long term and I will not worry until it comes to -20% drawdown (I expect it will not at all, but even if I can tolerate it).
 
Last edited:
Top