If a strategy is an attempt to single out a particular risk factor then a basket of strategies weighed appropriately should give us nothing else but the performance of the underlyer. What is the reason for any improvement in the statistics then?

Hi Dimitry, I am afraid we start from different assumptions. If I understand correctly, for you there is no alpha, for me there is plenty.

Very interesting, Francesco. I'm experimenting the same kind of combined strategies on Forex market, with the same improvements in sharpe ratio and drawdowns. To obtain the combined equity, I'm trying two different rules: one asks that all the systems agree before doing a trade, the other simply sums the results of the different systems.

Hi! I'm a trader and I am just wondering if both of you have actually traded your algorithms with real money. Are they profitable? In my experience, real life does not replicate

theoretical results.

are you equal weighting the strategies? I usually equal vol weight them, like a naive risk parity. um do you have a volatility target?

Awesome stuff. I am doing the same by combining mean reversion, trend following and pairs trading both in stocks and futures. Very simple systems, not great on their own, but robust. But combined have greatly improved sharpe. Got the idea from a book and never looked back since.

To rephrase my question, if you hold a unit of a risky sequrity you employ certain amnt of capital for a period of time. to compare returns with any strategy properly you either align returns and compare the capital at risk or start with equating the capital utilised and compare the returns. The difference will manifest your (possible) pick up in the alpha.

I adopt the same approach (diversification) with the portifolio of strategies I trade: there are several for each traded instrument, and some few traded instruments (hopefully more in future). The benefits in term of performance are evident, even though they tend to saturate quickly, at least in term of benefits for the same instrument (to diversify the traded instruments shold be much more beneficial, I think).

The return on invested capital is a key factor to evaluate the performance of the portfolio, at least to me, together with the volatility of the expected return normalized on a time basis. The challanging things are:

- How to optimize the portofolio (optimal strategies combination). Done by combining the trades of the strategy and analyzing the resulting equity line. The number of combinations "explodes" soon...

- Margin estimate for ROI calculation (I trade futures with daily strategies, that may enter in different days / directions)

And several others...

Nice one

Diane Tycangco yes the strategies are live. In real typically the strategies perform less well than in backtest for obvious reasons but that being said you can create a portfolio of strategies that makes money of course.

Seine Yumnam, thank for your message. The strategies in the chart are not vol weighted, but I agree with you they should be. Good point.

Well done. I am backtesting portfolio selection and weighting based on market regime. Have you done anything similar with your trading strategy portfolio?

Francesco Landolfi yes! Simple naive risk parity is powerful and robust. There is one more thing I like to do while looking at strategies and that is to see if the return the portfolio is delivering can be attributed to known risk factors. The easiest return based analysis will be to run OLS regression of your returns to factor returns of momentum, value, growth, size, market, etc. This will give you the alpha as well as the percentage of returns explained by this factors. The reason why we do this is that if 90% of your portfolio is explained by momentum+value, for instance, with very little alpha, you might as well buy 2 etfs that tracks value and momentum. This gives a deeper understanding of what risk premium we are harnessing. Fama french database has all the data you need!

I don't follow any specific strategy just observations of how a stock moves and I make money.

Francesco Landolfi Asking out of Curiosity .. Do you optimise the strategy per pair or do you run a portfolio optimisation ? Is it a different strategy per pair or the same one on all ?

Seria interessante colocar o seu sistema sob supervisão na MQ5.com ou Myfxbook para todos ver o seu desempenho.