Can Trading Edges be Quantified?

I

InTheZone

Today I spent the afternoon playing around with Wealth-Lab's web site (http://www.wealth-lab.com)

I must admit it is pretty impressive. It's similar to TradeStation in that you can define a trading system and backtest it against a symbol. Unlike TradeStation, it can also backtest against a portfolio of symbols as well and show you the aggregate results.

There are a number of public chartscripts (ie, trading systems) listed on their web site (see http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/getpage?page=Top10New.htm)

The backtest results of these systems are impressive. However, I suspect many, if not all of them, have been optimized for the data.

I tried running a number of these systems on a single symbol, and quite frankly (which any of you can do yourself from their website), these systems did not do so well.

I found the same case in EasyLanguage/TradeStation as well. I programmed in a number of setups, and also found that they did not perform well consistently on individual symbols (unless of course I optimized the system).

Which leads me to believe that a trading edge does not lie in a mechanically-defined trading system. In other words, an edge is not simply a trading system that can be defined in Wealth-Lab's WealthScript language, or TradeStation's EasyLanguage.

There's an intangible element to a trading edge as well that simply can't be programmed in.

Perhaps I am mistaken. If so, I'd appreciate hearing what you think.

-- ITZ
 
You "stumbled" on the basic fact that sometimes traders never learn...that the market is filled with non-quantifiable sources that will never be "programmable."

We discuss "edge" in detail with our new traders...and some of the 'edges" are simply: market access, lower costs, information, education, etc. ...and not some mystical program.

Automation can help traders tremendously to a point, and alerts are a valuable tool...but I really doubt that there will be a trading system that will work consisistently.

IMO
 
If there is an intangible edge to trading it's only because you haven't defined it yet.
We like to believe that we have this divine intuition, trading or otherwise. This allows us
to feel more unique or better about ourselves that we have this inborn or highly developed
extrasensory perception that can never be understood by the analytic mind. Not true, our trading decisions
are nothing more than the sum total of a loose collection of fuzzy algorithms or fuzzy logic rules. We perceive a set
of market events unfolding, the patterns are matched against a set of previously remembered patterns and we react
with more or less a conscious awareness of the processes underlying our decisions. But make no mistake, those processes exist and
are operative, and because of this they can be defined perhaps with only a little introspection.
 
Originally posted by InTheZone
...There's an intangible element to a trading edge as well that simply can't be programmed in.
Perhaps I am mistaken. If so, I'd appreciate hearing what you think.
-- ITZ
More...

Most of the trading systems made public on Wealth-lab or anywhere for that matter aren't very good. The good ones you won't ever see but they do exist. Wealth-lab by the way is an awesome program. Why would anyone publish a good system on their website? I'm going to post some of my marginal systems if I get bored, but not my bread and butter systems.

:)
 
Originally posted by Rogue Trader
To have an entry point is to have discovered an edge. To define the exit point is to own the edge.
More...

.... assuming the methodology is positive expectancy of course!
 
Wealth-Lab is a wonderful site but I was unable to achieve similar results when testing the band trading systems. It would be helpful if trsding edge's if they exist were more simply delineated and shown to us "outsiders"
 
Originally posted by InTheZone
There's an intangible element to a trading edge as well that simply can't be programmed in.
Perhaps I am mistaken.
More...
yes, you are. you only tried a few systems, and arrived at the conclusion that all possible systems don't work.

from logical point of view this is incorrect: you cannot prove a general hypothesis ("all systems are flawed") by listing a limited number of examples ("system A, B, and C did not work").

- jaan
 
This is an interesting subject that involves two opposing view points. However, I like to think that there is a blend that could be achieved.

Having been in both camps (discretionary and mechanical), I understand (to some extent) the benefits of both.

Rather than belabor the issues, I'll present may case for discussion.

There are two components to trading IMO: stock selection, and position management. Stock selection involves locating candidates that represent a high profit potential (reward/risk ratio, expectancy, etc.). Using scanners/filters, we look for strong trends, specific patterns, unusual volume or volatility, etc.. This aspect of trading is discretionary, and improves with experience.

Position management involves managing additional entries and exits once a position is initiated. This is much more important to the success of any system than stock selection (if I were to rank the two, I would place 80% on position management).

IMO, position management does lend itself to mechanization, and would be self-adjusting for any market if volatility is incorporated into the exit formula.

The advantages here would be the following...

1) it provides a consistent, emotionless process
2) it can be backtested for effectiveness and compared to other approaches (if you keep the degrees of freedom to 3, there should be no concern for curve fitting)

There's an excellent book that most of us have probably read (Van Tharp's 'Trade Your Way To Financial Freedom') that discusses position management and test results of differing strategies, so I won't get into any details here.

Wealth-Lab would represent an excellent test platform, since you can limit the stocks, and the test period. I would NOT program for stock selection; I would only program for managing the position (setting the entry and exit points).

Any thoughts on this?

TIA, Ron
 
You must log in or register to reply here.
Back
Top