[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RT] Band Study (better explained)



PureBytes Links

Trading Reference Links

--- In realtraders@xxxx, DH <catapult@xxxx> wrote:
> > I promise I will show you the hard evidence(I am getting 
> > artificial intelligence program and I need to get comfortable 
with it 
> > first) in less than 6 months.
> 
> > HA:May be you are right, who knows, but it will be a new toy to 
> > experiment and I like anyway to do the research :-)
> 
> I think the Research Unit's point is that it would be better to do 
the
> research first and then tell us the results, rather than telling us 
what
> the results are going to be and then doing the research to back up 
your
> claims.

HA: Someone who use astro(more than me) showed me the results so I 
have an idea of what I will be be looking for so I just want to do my 
own research and I will post the results(positive or negative)I was 
promised i will get it in June. 
> 
> PS - statisticians will tell you that you need at least 30 samples 
for a
> valid test. 

HA: I am going to do the testing historically from 1885 for DJIA and 
others.

More than 30 is better yet but that's a bare minimum. None
> of this "it happened 3 times and was profitable 2 of them." Also, no
> fair saying "it only works when the market is overbought or 
oversold"

HA: I don't know if this will be the case, but I will present the 
data and certainly will test the corrolation with and without 
overbought/oversold condition and any other condition. Personally i 
want to get to the bottom line, i know astro is not the answer for 
everything but i see it helping me(if you think it could be placebo 
effect, it may be but it work for me and that what matter to me)
> without testing the overbought/oversold condition by itself, 
without the
> astro stuff, and testing the astro stuff, without the
> overbought/oversold condition. If the overbought/oversold (nothing 
else)
> test does better than the astro (nothing else) test, or vice versa, 
that
> will be a pretty good clue about where the winners are coming from.

HA: I agree

> I look forward to seeing the proof in 6 months.

HA: Ok but no personal attack please, I will present the data 
objectively and we will be able to test using different conditions.
I am thinking of taking, for example:

Profit/Loss trades using DJIA data between 1885-now and see how 
stochastic model perform then test the same data on MACD, astro,...etc
in this way I will be able to compare performace of 
indicators "seperately" by comparing apples to apples and oranges to 
oranges on the same data

Haytham
> 
> -- 
>   Dennis


To unsubscribe from this group, send an email to:
realtraders-unsubscribe@xxxxxxxxxxxxxxx

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/