Tag Archives: views

Corrected Markowitz and Sortino views for optimal expected return

If you have seen any of my views of Markowitz for portfolio optimization for improved expected return.  You see them here:

Testing of Python Markowitz Portfolio packages

Deep testing of Python Markowitz Portfolio packages

You can always search for more with this

Here are some corrections on both Sortino and Markowitz from some an expert on my Telegram PRIVATE group:

Corrected view

I just wanted to mention a couple of things related to your recent posts.

The first is that Markowitz optimisations assume the returns are normally distributed (which is never the case in finance), so don’t expect to achieve the ‘expected’ return. Also bear in mind that such portfolio optimisation is intended to be done over much longer periods of time; doing it over a week or month is likely an exercise in futility because you’ll constantly be fitting your weights to very recent data which is almost guaranteed to change in the subsequent period due to the nature of financial markets.

In fact, many practioners have come up with improvements to Modern Portfolio Theory (which is certainly not modern these days), and many have also found equally-weighted portfolios to be preferable (e.g. https://www.diva-portal.org/smash/get/diva2:694576/FULLTEXT01.pdf). In any case, minimum variance optimizations should be preferred to mean-variance (aka Markowitz); the ‘expected’ return will likely be lower, but it will also be much more realistic than that of the Markowitz portfolio due to the greater out-of-sample persistence of risk-related measures such as variance when compared to those that contain information about absolute returns (e.g. mean of the return distribution).

The second is in relation to the Sortino ratio. You mentioned that the difference between it and the Sharpe is that the Sortino accounts for volatility. This is not the case; the Sortino is identical to the Sharpe except that it does not penalise positive deviation. For me personally, I prefer Sharpe because the objective function that I’m maximising is stability of returns. If you’re more interested in absolute returns, then Sortino is still preferable. But again, measuring over 7 or 30 days is essentially a random exercise.

Combining best performing forex pair should hugely increase your expected return and Sharpe Ratio

 

 

NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!

Forex currency pair volatility long and short term views

Forex currency pair volatility long and short term views

I have 2 views of the FX markets with 11 months of recent of 1 min bar data. I have made videos in the past which is under my IQFeed playlist on Youtube. All of the video segments shows a consistent view of the markets for my short (intraday) view vs the long view. I still need to plug in the portfolio optimization peice in future videos  You can join my Quant Analytics section here

Order Analytics

Do note there is an accompanying video and sample data to download for my members found here

Volatile Forex intraday analysis for Oct 13

 

 

 

 

NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!

And now my additional views of Tim Sykes

And now my additional views of Mr Tim Sykes

I just posted this on http://www.tradingschools.org/

First I am new here, I want to say BIG UP to Emmett for putting out this fascinating website which I never knew about. As a researcher of automated trading technology, I want to inform anyone who maybe interested in these two tools I really like. First, JStock is an open source stock screener you download and hack . This tool could easily save you the $6k you don’t need to pay out to Mr Sykes. Here is a video playlist I have on my Youtube channel.https://www.youtube.com/watch?v=TEMH5fWPIOA&list=PLzrMTuGTDsMvN-np-w1bk9DafFyiGnXoE

Also you could use XLQ which Excel based so I did is on my Youtube as well.https://www.youtube.com/user/quantlabs/search?query=xlq

Next is Mr Sykes. I sort of respect him in certain ways where i have some of his stuff. It is very repetitive with about 75% of analysing charts. Personally. I would never focus on penny stocks but short squeezes scare me. I would find it very hard to automate the forensics accounting part but you can easily automate his strategy with Jstock. Just some tips for anyone who may be interested in my POV. Thanks all and Emmett for making this cool site.

http://www.tradingschools.org/profitly-fake/#comment-4167

I wait for the call from his lawyers

Join my FREE newsletter to learn more about my views on penny stocking with automated trading 

NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!

Sang Lucci views on options trading and live Youtube video for their Master Class

Sang Lucci views on options trading and live Youtube video for their Master Class

I mentioned in this in my Trading Club as a source to learn options

http://www.sanglucci.com/the-life-and-times-of-sang-lucci-part-1-the-stock-market/

http://www.sanglucci.com/professor-ann-christina-lange-on-swarm-theory-what-automated-strategies-are-working-and-what-the-markets-will-look-like-in-10-years/

http://www.sanglucci.com/options-master-class/

Thanks to the NYC Contact for helping out with these links

Join my FREE newsletter to learn more about options trading as I do

NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!

Strategies: Views on Optimisation Algorithms for Automated Trading

Strategies: Views on Optimisation Algorithms for Automated Trading

Thanks to Sholom for sending

http://www.automatedtrader.net/articles/strategies/33/strategies-optimisation-algorithms-for-automated-trading

 

Join my FREE newsletter to learn how we do optimisation 

NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!

Ranking views using entropy pooling in quant analytics

Ranking views using entropy pooling in quant analytics

Hi everybody,

I wanted to ask you about ranking views on expectations (as in http://symmys.com/node/158), because I encountered an unexpected problem. I apply the entropy pooling on a portfolio of N stocks, and I use only the simple-returns as risk-factors:

– View specified as a full-ranking ( E(r1) > E(r2) > … > E(rN) )
– Entropy pooling to get the probabilities in order to compute the objective (E(ri) for all i in (1, 2, …, N)) and the constraints (CVaR-controlled allocation)

Unfortunately, the numerical minimization always give expectations like E(r1) = E(r2) = … = E(rN) (the constraint always “hit the barrier”). When I set a view like E(r1) > E(r2) and leave the other assets free, I also get E(r1) = E(r2).

Thus, I went back to the two assets – two data points – two views (probabilities sum up to one, and ranking view between the assets), and I found E(r1) = E(r2) analytically.

Then I was able to find an analytical formula for the dual formulation of a symmetrized version of the KL-divergence, but the results (both numerical and analytical) are exactly the same as with entropy pooling.

Did you also notice this problem? Does someone have a solution?

 

==

The issue you encountered depends on how the ranking view relates to the prior distribution of the returns.
Let us consider the bivariate case. If your prior is already such that E(r1) > E(r2), and your view is in the same direction, i.e. E(r1) >= E(r2), then the posterior will be equal to the prior, and the view will be satisfied as a strict inequality E(r1) > E(r2) in the posterior.
If on the other hand with the above prior your view is E(r1)<= E(r2) and thus it contradicts the prior, then the posterior will feature E(r1)=E(r2), which is the closest solution to the prior that satisfy the view.

 

==

 

I got another conceptual question about the entropy pooling.

Suppose you have T observations of two risk-factors X and Y (that is (X_i, Y_i), for i in [1,2,…,T]) which are completely independant (take the dummy case of a two independent gaussians for example). Their prior (standard) esperance estimator is E(X) = 1/T * sum(X_i) = x and E(Y) = = 1/T * sum(Y_i) = y.

Let’s assume now that we want to implement a view on X such that E*(X) = x*, that is we get a new set of probabilities p_i (i in [1,2,…,T]) which are different (even if those deviations are very small) from 1/T. Thus, if we use those new probabilities for Y too as it seems to be the case in your case studies, we have E*(Y) = y*, different from E(Y) = y.

How is that possible since the two risk-factors are assumed to be independent? Is there a way to take the dependence structure between the risk-factors into account when applying the pooled probabilities within this framework?

==

 

In the gaussian case, you could use the analytic formula. If you apply the analytic formula just to x (ie. you apply it univariately), then it devolves to just whatever your view is. That means it will not impact y.

For the full algorithm, the reason why you don’t get the right answer is because by treating them separately you’re not letting the optimizer know that there’s no correlation between the two assets. The EP algorithm minimizes the difference between two distributions. If it doesn’t know what one of the distributions is, then how can it be expected to minimize anything?

All you have to do is just gather x and y (and z, etc) to a single matrix before applying the algorithm

==

 

OK, the gaussian case is obvious, but:

“If it doesn’t know what one of the distributions is, then how can it be expected to minimize anything?

All you have to do is just gather x and y (and z, etc) to a single matrix before applying the algorithm. ”

I understand you and this is basically what I did, but if you don’t express views on y (and z, etc), it won’t change anything. My problem lies in the fact that this algorithm modifies only one set of probabilities for all risk-factors. If you take two independent risk-factors X and Y with different dynamics (e.g. X ~ arch(10) and Y ~ garch(1,1) with independent innovations), and express a view only on X, how can you let the optimizer know about the distribution of Y? The idea of numerical entropy minimization is to set views as linear constraints of a convex optimization problem, but the risk-factors on which no views are expressed are not an input of the algorithm, or am I missing something?

 

NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!