An excerpt from an email from Duke to the LinnSoft Yahoo Group on 03/26/02

I have enjoyed the thread in regards to the merits or lack there of in the construction, interpretation and reliability of indicators. Statistically speaking it would take many life times to truly have the required sample set to validate the effectiveness of a single indicator (considering the numbers of parameters) much less the robustness of a set of indicators. Optimization as previously mentioned does have usefulness but also carries with it numerous pitfalls curve fitting being the perhaps the largest negative.

(Opinion starts here J )

I think in real world terms first knowing the construction of how an existing indicator is built is essential then adapting that indicator to the market or style of trading is second. One of the many mistakes that is prevalent in the field of trading much less technical analysis, is something that we all have done at one point or another and that is multi-collinearity of indicators. We confirm momentum indicators with other momentum indicators etc.

Since we don’t have a lifetime of testing to confirm the true validity of the tools we use the best we can hope for is to formulate those tools that have the highest probability of success (based on previous tests on in and out of sample data.) In other words, putting as much weight of the evidence in our corner as possible. 

We all know there is no Holy Grail of trading (yet that doesn’t stop me from trying) so the best that we can shoot for is consistency and discipline. This brings me to my last point. We all either can build or have seen a trading system that touts remarkable performance with of course little draw-down. Those that have the where with all to evaluate the system quickly find out that most are unrealistic in that they have been “optimized” with parameters that quickly fade in real world trading. Those that don’t usually become painfully aware that historic draw-down isn’t conservative and why disclaimers are attached that say past performance does not equal the future.

So then, if we can’t guarantee the performance of tools that we use then we had better control the risk that we apply with our tools. Bottom line, a good or great system of indicators without great money management is doomed. 

Jones Law, 

1.) Now what you are trading and why you are trading it.
2.) Control those variables within your control
3.) Trade outside of your comfort zone (counter-intuitive)

I have a few other tenets that I have stolen from other traders but these are the main points.

I know I am probably preaching to the choir and that my points may be a tangent in the current discussion. I will cease my ramblings and let others wax on philosophic! J

Again, I am enjoying the current topic. I look forward to other areas this may lead to in the discussion.

Duke