Many traders today are endeavoring to construct a trading methodology from stone knives and bearskins.
In the future, any implementation of walk-forward analysis (WFA) which cannot change from IS to OOS and back based on TA signals, as well as conduct retrospective analysis based on arbitrarily defined historical date ranges will be considered crude, archaic, and dangerous. The market changes behavior (mode) when it pleases. Technical analysis (TA) is our means of measuring such change (sometimes paired with a well-developed intuition). WFA which cannot respond to TA is WFA with a passing relationship to reality at best. I include here under the umbrella of TA the analysis of current system performance, which itself can work off of many metrics, including specified or optimizable # of trades. In other words, static WFA will give way to dynamic WFA (and almost certainly already has). When this is productized for the retail trader is another question entirely. --- In [email protected], Howard B <howardba...@...> wrote: > > Hi Herman, and all -- > > The issue determining the length of the in-sample period is not the number > of trades, but the number of data points it takes for the model (afl) and > the data (OHLC) to become synchronized. > > On Fri, Feb 19, 2010 at 6:51 AM, Herman vandenBergen <psy...@...>wrote: > > > Hello Howard, > > > > I cannot help but wonder why the OOS periods in optimizations aren't based > > on the number of trades, instead of some arbitrary time period.
