Selecting Markets to Trade

Chris Murphy · Post by **Chris Murphy** » Wed Jul 23, 2003 10:06 pm

Through out this forum I have read that many of you preselect your markets to trade based on the markets characteristics. My gut tells me that this is not the right approach. When selecting markets to trade based on how trendy they have been in the past and then back-testing you are curve fitting your system. I could use a cycle trading strategy and could make a fortune in back-testing if I preselected my markets based on those that exhibited cycle behaviour. Please explain how what you are doing is not curve fitting the data. I do not mean to offend anyone but I see no rational explanation for how you can say that because in the past the market was trendy it will be so in the future. I view trendfollowing as a way to scalp profits from the markets due to the efficient market theorists who believe in a normal distribution when in fact the tails are fat. Based on my theory ( of course I didn't event this ) of non-normalized markets if their were an infinite number of markets and you trend followed them all you would make an infintie amount of money if one used proper money management and this would occur with an amount of volatility and risk that would be so small as to converge to zero with each additional market traded. This suggests that one should trade as many markets as is reasonably possible and goes against preselection as some of you suggest. I am contemplating using this as my Masters research topic so please contribute to this thread if you have the urge

Chris

Hiramhon · Post by **Hiramhon** » Wed Jul 23, 2003 11:21 pm

How about we email our thoughts straight to your thesis advisor and cut out unnecessary layers of middlemen?

Chris Murphy · Post by **Chris Murphy** » Thu Jul 24, 2003 9:16 am

I'm not recycling ideas Hiramon if thats what your suggesting. Just trying to raise a debate and view all sides of opinion, that is what this forum is for. Also, I've never said this is for my thesis, I'm doing a self designed course and am trying to decide on my topic.

MCT · Post by **MCT** » Thu Jul 24, 2003 10:12 am

Youâ€™ve brought up a topic that doesnâ€™t receive the attention it deserves.

This suggests that one should trade as many markets as is reasonably possible and goes against preselection as some of you suggest.

I find â€œbroad diversification to be a hedge against ignorance.â€

Chris Murphy · Post by **Chris Murphy** » Thu Jul 24, 2003 10:22 am

[quote]
I find â€œbroad diversification to be a hedge against ignorance.â€

damian · Post by **damian** » Thu Jul 24, 2003 10:55 am

Chris, first you said:

A:

if their were an infinite number of markets and you trend followed them all you would make an infintie amount of money if one used proper money management ....... This suggests that one should trade as many markets as is reasonably possible

Then Menelik said:

B:

I find â€œbroad diversification to be a hedge against ignorance.â€

MCT · Post by **MCT** » Thu Jul 24, 2003 10:55 am

Also:

When selecting markets to trade based on how trendy they have been in the past and then back-testing you are curve fitting your system. I could use a cycle trading strategy and could make a fortune in back-testing if I preselected my markets based on those that exhibited cycle behaviour.

Couldn't agree with you more. As Nassim Taleb points out "Mathematics is a tool for meditation not computation." That message seems to have been lost to most traders.

enigma · Post by **enigma** » Thu Jul 24, 2003 11:05 am

Hi Chris,

I think your use of the term diversification needs to be more specifically defined. As some of the other threads have already pointed out, one can diversify across markets, time frames and systems. For example, if one is a passive investor and applies the buy and hold strategy, broad diversification will give you a portfolio that mimicks the market index, as in your example. In this respect, we are left with the systematic risk of the market that we cannot diversify away.

On the other hand, if we apply a combination of systems that allows for shorting, we might not require that many different assets/stocks to provide the diversification or risk level we seek.

Regarding market efficiency (I know there are many interpretations out there) I think that a market can be efficient and trend at the same time. Also, a trending market does not necessarily imply fat tails and likewise, fat tails do not necessarily imply trending markets. Fat tails, however, do imply an inefficient market.

Lastly, regarding market selection, I think the problem is the same as most other financial issues, namely we can only estimate the future with past data (insiders info aside), which may not necessarily be the most rational thing to do, but diversifying across many assets may not be the most practical thing to do either, especially if you're a small trader. I believe the majority in this forum choose their markets not just by the trendiness in the past, but also by finding a compromise between the issues of practicality, liquidity, who the main market makers are and other personal reasons. After all, trading is just as much of an art than it is a science, if not more so the former

Anyway ... just my two cents.

Best,
Louis

MCT · Post by **MCT** » Thu Jul 24, 2003 11:05 am

damian

I missed that

Chris Murphy · Post by **Chris Murphy** » Thu Jul 24, 2003 11:07 am

Damian, I see how my post was confusing. Let me clarify. My agreement was in terms of the stock market. The more stock you own the more you become the market and your performance will thus mimic the market. In terms of commodities I don't see how this applies unless it is being suggested that the more markets you trade the less ROR you get because your margin requirements go up. This makes sense to me, in order to maximize expected ROR/Risk their will be a limit to how many commodities you would want to trade because at some point the ROR decrease will not be compensated by a significant decrease in risk or volatility ( the efficient frontier so to speak). However, I'm just suggesting that in theory you could eliminate volatility in ROR if you could trade an infinite amount of commodities. This won't maximize ROR but looks to me in theory to provide you with a riskless ROR. Essentially I'm suggesting the opposite of what LTCM did. I'm going to read "Fooled by Randomness" as my next book and from what I've gathered I may be suggesting what Taleb suggests in his book.

Bollinger · Post by **Bollinger** » Thu Jul 24, 2003 11:34 am

Chris,

The real problem with trading a larger number of markets is that you run into margin requirements that are too high. If you trade 20 or 30 major futures markets at a high enough risk level that you would historically see a 40% or so drawdown you might have a margin to equity ratio of something like 20-30%. But if you trade everything under the sun you see an interesting phenomenon -- equity volatility decreases because of the diversification, so to achieve the same performance (or even close to it) you need to turn the risk/leverage level up. The result for a 100 market portfolio can be margin requirements into the high 40s and even 50+ percent. This is unacceptable -- you subject yourself to the realistic possibility of a margin call in a run of the mill drawdown.

So if you want return you're stuck with trading the smaller number of markets, and you have to pick them somehow. I'd suggest that with nothing better to go on you select from the abundant number of markets that are both highly liquid and that have performed well in the past. If you prefer instead to take the shotgun approach and aim merely for massive diversification, go ahead and throw in the US Dollar against the Hong Kong Dollar, Appalachian Coal and Tokyo Silk and let us know how it goes.

Neal Stevens

Chris Murphy · Post by **Chris Murphy** » Fri Jul 25, 2003 11:50 am

Bollinger, your response makes a lot of sense to me. You have to strike a balance between ROR, Volatility/Risk, and Margin requirements that you feel comfortable with. Because you can't trade everything under the sun you pick what markets have worked well in the past, not necessarily ideal but better than the alternatives. Does anyone dynamically change what instruments they trade based on market conditions? Ex. corn was unprofitable last year while european bonds were winners so you drop corn and add euro bonds.

MCT · Post by **MCT** » Fri Jul 25, 2003 12:40 pm

I trade anything that meets my criteria: 1.Interest Rates 2.Currencies 3.Individual Stocks and Stock Indices 4.Energy 5.Metals 6.Commodities.I tend to view my system as a supportive and firm partner. I take care of the qualitative work, which I believe is essential for a dynamic portfolio, and my partner [the system] takes care of the quantitative and risk management aspect of our operation. All markets are traded the same with the same system. Im not sure if that answers your last question â€¦

Chris Murphy · Post by **Chris Murphy** » Fri Jul 25, 2003 12:45 pm

Yes, that does. Is your market selection then subjective or do you have some general rules ( speaking of things more complex than simply enough liquidity and volume )?

Kiwi · Post by **Kiwi** » Fri Jul 25, 2003 6:05 pm

Reading the questions I'll offer a further view that is basically complementary.

You need volume to get liquidity (see list below - no sorry seem to have deleted it but just look at the aberration portfolios or run a filter for the top 25 by volume and open interest). You might add 3 or 4 overseas contracts with good reputations for trending that are often uncorrelated with the high volume US contracts (see the aberration global portfolio). You cant switch in and out of them on an annual basis because as you'll see when you look at historic data many of them have a number of bad years in a row before a big trend - which you cant predict

You can also skip some bad markets (Lumber, Juice, Pork Bellies) and be careful of some others (Coffee because the brokers will create slippage for you

, and Natural Gas because the contract is large, indexes because they dont trade well with trend following systems). You may also drop some potential markets because your system is never in tune with them (possibly sugar and cocoa) but dont drop them if the system makes a little money and they are uncorrelated with your most profitable markets so they smooth out your equity curve.

John