|
PureBytes Links
Trading Reference Links
|
> My question is for those people who have done that comparison. How
> much, and what is the best way to degrade daily results so that they
> approach the accuracy or validity of the REAL simulated results
> obtained via tick data if tick data isn't available?
* This will vary greatly depending on the system. The only way
for you to know for YOUR system is to do the study yourself.
* The best and only way to get *accurate* results is to run your
system with intraday data. Set the chart up as a daily chart,
but specify that the system should use a higher "strategy testing
resolution" (format the system, Properties tab). Tick data isn't
necessary; if you can get 1min or 5min it should give you an
accurate indication of what would have happened intraday.
Probably even data as coarse as 60min would be pretty accurate,
and better than what you'd get with EOD data.
> TS has the ability to use bouncing ticks.
Bouncing ticks are a guess, no more. I wouldn't trust my money
to a model that had only been validated with EOD data and a
bouncing-tick wet-finger guess.
Gary
|