|
PureBytes Links
Trading Reference Links
|
John Nelson wrote:
>
> This posting provides some food for thought. If one could devise a
> function which accurately simulates the random and non-random aspects of
> markets using convenient data measures (price, volume, volatility) then it
> would be possible to train a neural net using other than historical data.
<snip>
How do you separate the random from the non-random? And, if a lot of
sellers come in at the end of the day in a bullish market and economy,
say, because a long holiday weekend is coming and they don't want to go
home long...then is that random or non-random? It might be random
because the reason they are selling has nothing to do with the market or
the economy,etc, but it might be non-random because the same selling
occurs ahead of every long holiday weekend, making it predictable. (If
predictability is an element by which non-randomity is defined).
Regards,
Monte
|