Trade more sh*t!?

[drmark27]

Baiting newbees with quant lingo. Awesome. Every day another trash product. A quick glance at the site makes that apparently clear. What is not clear is what those guys are even selling. A lot of mumbo-jumbo but nothing of substance. Example in case, the article on deep learning. Big bait title and nothing that relates deep learning to trading, something I do every day and claim to have some expertise in.

Sounds like their general suggestion is to find a number of decent strategies in noncorrelated markets such that when traded together, a synergistic improvement in RAR results.

What do you disagree with, exactly?

One thing that comes to my mind is correlation approaching 1 in severe market crashes. If the strategies are backtested to 2008, though, then one could see how they fared (although this would be a sample size of 1). Really, the argument that correlation goes to 1 in severe market crashes is itself a somewhat hollow claim because I don't know we've had enough "severe market crashes" to make for a large sample size.

 

[TommyR]

if you have 100 thousand really sh1t strategies you can also afford to turn some off and on osccaionally

 

[ironchef]

@qlai,

Why don't you google volatility pumping. This topic was discussed here not long ago:

http://parrondoparadox.blogspot.com/2011/02/parrondos-paradox-stock-market.html

 

[qlai]

Why don't you google volatility pumping.

Hi, am familiar with the concept, but not the name. But I don't see how it applies to trading more sh*t. My personal opinion is that the authors don't have any good strategies (at least none they would be willing to share) so they are saying - oh don't worry too much that you can't build anything worthwhile, just trade lots of mediocre ones and one of them is sure to be in synch with the markets. Might as well throw darts at Wall Street Journal and use proper money management.

 

[ironchef]

Hi, am familiar with the concept, but not the name. But I don't see how it applies to trading more sh*t. My personal opinion is that the authors don't have any good strategies (at least none they would be willing to share) so they are saying - oh don't worry too much that you can't build anything worthwhile, just trade lots of mediocre ones and one of them is sure to be in synch with the markets. Might as well throw darts at Wall Street Journal and use proper money management.

There is one catch: Pumping works only if the assets are uncorrelated. e.g., you cannot do this with combination of options and underlying. I actually backtested this concept with different stocks combinations and that didn't work out. Why, I think the stock market, stocks were too correlated, they moved up and down roughly in unison.

 

[tomorton]

Trading large numbers of instruments simultaneously works in certain situations. But only in those situations. The only two I'm aware of would be trend-following and the MA cross-over.

Its accepted that prices only trend for a minority of the time, so for the strict trend-follower there is usually no reason to get into or stay in a trade. If you trade only EUR/USD or maybe if you add in USD/JPY and GBP/USD you will still often have nothing to do and no income. The income you make will be lumpy and subject to very significant draw-down. Trading 28 forex pairs smooths all this out, trade load and income stream, but it also increases ROI.

Same with the MA cross-over. I'm talking about the orthodox version where the exit signal is the reverse cross-over to the entry signal. Trading this on a basket of instruments will work and will give a good return BUT the markets chosen must be uncorrelating as far as possible.

 

[drmark27]

Trading large numbers of instruments simultaneously works in certain situations. But only in those situations... Trading this on a basket of instruments will work and will give a good return BUT the markets chosen must be uncorrelating as far as possible.

I've seen much discussion of correlation over the years but very little about covariance. Where does the latter fit in?

 

[fan27]

I follow a similar approach as laid out in the article but I have tools and approaches to help in the effort. I am currently trading 135 unique strategies in a portfolio where it would not make sense to trade any of the strategies in isolation. That being said, I have developed approaches to deal with some of the pitfalls of this approach (i.e. strategy correlation, etc.).

 

[tomorton]

I've seen much discussion of correlation over the years but very little about covariance. Where does the latter fit in?

Never heard the term covariance. Is it just correlation by another name?

 

[drmark27]

Never heard the term covariance. Is it just correlation by another name?

My mistake. I was thinking about cointegration--not covariance--which I thought pertains to correlation stability.