July 28th, 2016

Generate Positive Alpha

A short term stock trader has a choice, and that is to participate, take a position, or not. It is always his/her prerogative. Participating, taking action, is a deliberate act, that it be discretionarily done or delegated to a trading script.

A trading program will do what it is programmed to do, nothing else, and therefore, it is about the same as if its designer had made those same trading decisions except much faster, without hesitation or second guessing. A program, or a discretionary trader for that matter, will issue buy and sell orders according to their respective preset trading rules. Neither will invent a new rule on the fly, both will want some evidence that their prospective rules has the potential to generate profits.

Who do you think is responsible for a program misbehaving, or not doing what it was intended to do? I'll give you a hint. It's the one setting behind the keyboard. If you deliberately, even if at times unknowingly, design a program made to lose money over time, well, technically, that was your choice too.

If your program fails, the main reason should be for poor strategy design principles, misguided assumptions, and assured unprofitable trading procedures. Putting the market at fault is the same as saying you did not know what you were doing, didn't sufficiently analyze or understood the problem, or just didn't care! And those are not very good excuses. I would say: you simply didn't do your job.

Some try to extract short term logic out of chaos. Sure, there could be some, but by logic, if there was, you could or would find it, and your program would not blow up, would it!

A trading program that fails is synonymous to: you lost, generally a big percentage of your equity. It is how you identify a program that failed. Each iteration of these failed programs can be measured in lost capital, lost time, and lost opportunities. How many such programs are you ready to design before getting the message?

In trading forums, I often read: "... all trading programs fail." And surprisingly, these same people continue to promote their own trading programs. Is there a way to politely say: why are you promoting something that will eventually lose you money, your follower's money, or client's money since your program, by your very words, should fail too?

Because you have a trading program that is debugged, meaning that it runs without crashing all the time, does not mean that your program will generate above average return over the long term.

I don't think all programs will fail. Only those that are poorly designed. And there are many of them. Once, I took a miserably designed trading script. It was so bad you were almost surely guaranteed to lose, not just a little, but bankrupt your trading account. And to do so, all that was needed, was to keep it running. You can see those tests here.

The objective was to transform a very bad trading strategy into a profitable one. The article describe how it was done. I would recommend to anyone to do this kind of transformation exercise. There is a lot to be learned from it. In that article, the trading logic was also reversed since it was so bad. It went on the premise that if you are wrong most of the time, then doing the opposite, would make it right most of the time. Well, it doesn't necessarily do so as the article's tests demonstrated.

A trading script is composed of a set of trading rules based on conditional trade triggering mechanisms. It determines states, sets conditions, and logical rules under which specific trades can take place.

Overall, a trading program has a unique mission. Its objective, in the long run, is to generate money, not lose it. But profits are not enough, another directive needs to be applied. A trading strategy should generate positive alpha, meaning beat long term market averages. Otherwise, an index fund would be preferable since it would beat your strategy.

Why would one put an untested trading procedure to work is beyond me when it is so easy to backtest any of them.

Some trading strategy designers think that one strategy fits all, meaning that it can be applied in all markets, in all time frames, under all conditions. I find this a rather naive view of the problem. You see someone proposing this, run away. It is a telltale sign that you might not end your game with a profit.

Naturally, if you can't put your trading strategy to code, it should say everything, and classify it as discretionary. There is nothing wrong with using a discretionary trading method. One can find long term discretionary trading strategies that don't need to be coded, but still could if desired.

I think that in all circumstances, one should remain reasonable and consider common sense as his best asset and defense. What applies to one market might not so readily apply to another, they are different creatures.

If you present me with something that does not make sense to me, what do you think my reaction should be? What should I think about your premises or attempted logical dissertation?

Some traders give themselves and their programs too much importance. They see the game as them against this beast they call the market and give it personality traits as if it had a will, or a psychology of its own. Not at all, an individual is playing against, and also with the multitude of other players out there. People and machines confounded, we are talking of millions and millions of participants. And here is the surprise, if you look under the hood, you will find either the brightest people on the planet, or their software alter ego: their trading programs. What the other guy or machine is doing is not against you, they too are looking at their potential benefits in any trade they undertake. They too go long on positive expectancies.

Short term traders better learn to program and design trading strategies that can last even in the presence of other machines. Traders are now the minority, machines have taken over the space.


Created... July 28th, 2016 © Guy R. Fleury. All rights reserved

Alpha Project