Testing Projects Question #1

[b519b@xxxx]

Stephane or Tom,

Would either (or both) of you like to be the "lead" tester (or "co-
ordinator")? I've been part of such on-line testing before and it 
best if 1 or 2 people with some experience in testing/refining give 
informal direction to the project. Although I have done more hours 
than I can count backtesting (with other programs before discovering 
AB), I think I have a lot to learn about backtesting in general and 
using AB in particular. I would be an eager contributor to a testing 
project. 

As I see it the role of the "lead" tester would be to give overall 
direction and focus by: 
1. identifying the basic concept and/or initial code.
2. summarizing, from time to time, results posted by other testers.
3. (optional) making suggestions about possible variations to test.
4. calling an end to testing when it is time to move on to a new 
method. 

What do you think?

b

--- In amibroker@xxxx, "Stephane Carrasset" <s.carrasset@xxxx> wrote:
> Great idea to propose a trading system, with the purpose to improve 
it
> Have you tested it on a universe of stocks
> what parameters?
> buy and sell , open next day?, close of the day if H>ref(H,-), 
> cost?, stoploss?,Fixed equity or coumpound?, Bars in test?
> 
> I will be happy to contribute to test a trading system
> 
> SC
> 
> 
> > 
> > Here is an interesting system. I gravitate toward the simplicity 
> and
> > underlying concept of the system.
> > 
> > I think I have coded it accurately from an article in the most 
> recent
> > issue of Technical Analysis of Stocks and Commodities. No 
> guarantees.
> > The author seems to gravitate toward commodities rather than 
stocks 
> and
> > shows no stock related results.
> > 
> > Backtesting on my stock list shows a modest profit. Examining the 
> graphs
> > by eye (Indicator Builder) shows that there is a significant lag 
in 
> the
> > system (particularly on position entry). So, smart people should 
be
> > able to improve on it.
> > 
> > I would be interested in hearing your and other's thoughts and the
> > results of any experimenting.
> > 
> > Best regards,
> > -Tom McDaniel
> > 
> > 
> > 
> > /* Trend Detection Index */
> > /* Developed by M.H. Pee */
> > /* From Technical Analysis of Stocks and Commodities, October, 
1991 
> p. 54 */
> > /* Coded by T.L. McDaniel September, 1991 */
> > 
> > 
> > MaxGraphs = 2;
> > Period = 20;
> > 
> > TDM = ( Close - Ref (Close, -Period));
> > ATDM = ABS (TDM);
> > PTDI = SUM (TDM, Period);
> > ATDI = ABS (PTDI);
> > FDAM = SUM (ATDM, 40);
> > SATDM = SUM( ATDM, 20);
> > TDI = (ATDI + SATDM) - FDAM;
> > 
> > Graph0 = TDI;
> > Graph1 = PTDI;
> > 
> > Buy = iif (PTDI > 0 AND TDI >0, 1, 0);
> > /* Sell = Cross (0, PTDI); */
> > Short = iif (PTDI < 0 AND TDI >0, 1, 0);
> > Sell = Short;
> > /* Cover = Cross (PTDI, 0); */
> > Cover = Buy; 
> > 
> > Buy = ExRem (Buy, Sell);
> > Sell = ExRem (Sell, Buy);
> > Short = ExRem (Short, Cover);
> > Cover = ExRem (Cover, Short);