Technical Analysis has evolved with the trend towards more quantitative methodologies. The discipline is grounded in structuring objective rules and testing them for efficacy. The first portof-call has always been the back test or system test, but is that the right place to start? What are the pitfalls of mining historical data for ideas?
The issue with an all-encompassing back test is that there are far too many variables to test for it to be considered valid. The tests usually become an alchemy of entry rules, exit rules, position sizing, stops, starting capital, and the list goes on. The result, while seemingly robust, is often difficult to repeat when implemented.
In this presentation, Mathew will look at a quantitative approach of running multiple tests to break the process into discrete steps. Starting at the signals, then the trades, then the full system test. Each step needs to be thoroughly examined to be sure that the results are statistically valid. We’ll also cover some statistics, in plain English, and discuss how to build an accurate sample distribution to compare against our test results.
For any Analyst who wants to research new ideas or be able to present ideas that can stand up to rigorous quantitative scrutiny, this presentation will help you see that not only is a quantitative approach achievable, but as Technical Analysts, we are in the best position to drive quantitative models.
