North Carolina
Experience: Beginner
Platform: NinjaTrader, Tradestation
Trading: es
Posts: 644 since Nov 2011
|
Backtesting is just one part of the overall process of strategy development, hypothesis falsification, etc. I think that market cognition is just as important. You do not hear many people talking about market cognition though but market cognition is what really drives the market. Market cognition is often much more difficult to backtest because it doesn't rely primarily on statistical verification but rather on forward thinking, analysis of other traders, and unknowable speculations.
Examples of market cognition:
(1)
A speculation is formed based on rational insight. Backtesting or hypothesis falsification (not verification) can be used to falsify either of these statements.
A. Interest rates are expected to go up in the near term. So, borrowers will rush to borrow money and deploy it before they go up.
B. Interest rates are expected to go up in the near term. As such, house prices will be pressured. Borrowers will be more selective.
(2)
Below is another form of market cognition which can not be verified by simple price stream, i.e. may have been made in 07 or 08.
A. US wages have been under pressure for many years. As such, credit is likely to be destroyed and this credit destruction is likely to cause many housing defaults.
Im order to turn (2) into an actual trade you will need a specific trading plan where backtesting or a term more generally useful, simulation could be used to figure out various risk/reward scenarios.
Unfortunately, even with a quantitative empirical framework of analyzing markets backtesting is merely part of the process, a very important part but just a part. You could learn Tradestation's Easylanguage. If you have a Fidelity account, you could try Wealthlab too which you might find easy to test as a beginner. Paper trading is useful but it is possible to paper trade for too long too. This is more of an issue with discretionary traders though.
But what is important to understand is that traditional backtesting can only verify the form, I think that a pattern or signal A in the market is predictive of B. It cannot be used to test more advanced hypothesis that do not follow that form or that do not occur with statistical regularity. Simulation however might be able to test some of those scenarios and does not require statistical regularity.
|