Australian (ASX) Stock Market Forum

Quest for a NEW predictive BUY signal indicator

I like FX because it is a zero sum game, where short and long positions are generally in balance. Compare this to stocks which has a huge bias towards long positions (short selling can sometimes be banned). In stocks this makes the cascades from stop losses much stronger falling knives. In FX you don't have that problem and can get good reversals off spikes.

Tradeguider calls them Uptrusts and Tests. The EUR on the NFP ann was a classic the other night. Spiked up, the uptrust got sold into then it reversed, got bought up, reversed and carried on.

CanOz
 
Nice thread.

Cheers.

Sorry, don't use any indicators so can't contribute, but a good read.
 
Oh, what the hell...

Here's something for nothing. The following MS code replicates my indicator very closely. I'm not going to divulge the exact workings, but any competent coder will be able to tweak from this to get their own workable ideas:

Long:

Ref(Stoch(3,2) < 10 AND L < Ref(LLV(L,100),-1),-1)
AND C > Ref(C,-1) AND C < Ref(H,-1) AND
V > Mov(V,50,E) AND OBV() > Ref(OBV(),-1)


Short

Ref(Stoch(3,2) > 80 AND H > Ref(HHV(H,20),-1),-1)
AND C < Ref(C,-1) AND V > Mov(V,50,E) AND
OBV() < Ref(OBV(),-1) AND Ref(OBV(),-1) > Ref(OBV(),-2)

Been playing around with this, i should be able to post code for a scan in Amibroker tomorrow morning sometime.

I like it, its looking for trend ending volume in extreme overbought/oversold conditions. Well done MS, good thinking there. Seems to hold up well in choppy conditions.

Some ideas for trailing stops, countbacks, EMA X, ATR?

Initial stop could be the high/low that triggered the entry.



Cheers,


CanOz
 
The EUR on the NFP ann was a classic the other night. Spiked up, the uptrust got sold into then it reversed, got bought up, reversed and carried on.
Trading the news is difficult. I'm used to trading stocks not FX, so I could handle the first upthrust, but then my talent ran out after that!

Before the announcement (at 23:00 AEDT) the price was creeping down to a recent low, so I was fairly sure the first move for the ann would be up and I went long with three separate entries. Then bought in again on the first retrace. The first top spike in the 1:3470's was fleeting but I thought it would test it again, and I closed my positions when the momentum slowed the second time.

After that it was clearly going down but I didn't really know how to trade it, so I stopped trying after a little while. Trading is easier in hindsight lol
 

Attachments

  • nfp.png
    nfp.png
    70.4 KB · Views: 5
Much easier on the sim too. I went long as you did and got thrashed out when it retraced. Thinking, "crap i should have faded that" i waited until the selling stopped and the stopping volume showed up, went long and i'm still in the position, ON THE SIM!


FX Is just entertainment for me for the next few years. To me as a discretionary market its just too hard. Mechanically i can see some potential.

Cheers,


CanOz
 
“Predictive” – Possible???

My use of the term relates to the idea that in many cases there will be signs in the data prior to the surge, that the stock is about to surge.

This is apart from the unexpected changes due to announcements etc, etc and the general chaos.

To quote from Incredible Charts

"Twiggs Money Flow warns of breakouts and provides useful trend confirmation. It is based on the observation that buying support is normally signaled by:
• increased volume and
• frequent closes in the top half of the daily range."


But do these parameters signify more than a trend confirmation???

pic3mod.jpg

Note: The image is an ideal selection to illustrate the concept - not an ideal buy signal. Its just an example that shows the concept.

Referring to the image the closing price occurs in the lower half of the range as the stock price falls.
But in the three days prior to the surge the closing price moves up the range.

There are also changes in the relative volume (not shown) and in the total range (the relative size of the bar) + gaps, etc.

My draft indicator picks up the trend in these parameters and surges upward for 3 days prior to the price increase.

(Of course nothing is perfect blah, blah and you would use other info etc.).

The dream is that this is a general pattern that reflects modern market behaviour and patterns.
(VSA looks at similar patterns, though I have little experience with VSA).

An indicator that detects signals of a price rise in the days prior to the surge is “predictive” or that is the dream anyway.

PS Thanks to everyone who has contributed ideas, comments and observations, all which have been very useful and just what I was after in submitting the post.
 
Rusty99,

That indicator looks like a 'Slatyer Crossover' probably set @ 3 13 or you might have tweaked it slightly.

IF it's not the exact indicator then it's something similar

It might help 'time' entries better in the right market conditions, but I'm not a fan of it.

If it works for you then good, but i'll give you a few more months and
you'll start introducing other indicators overlayed over this indicator
because of all the false signals.
 
Rusty99,

That indicator looks like a 'Slatyer Crossover' probably set @ 3 13 or you might have tweaked it slightly.

IF it's not the exact indicator then it's something similar

It might help 'time' entries better in the right market conditions, but I'm not a fan of it.

If it works for you then good, but i'll give you a few more months and
you'll start introducing other indicators overlayed over this indicator
because of all the false signals.

Of course I use multiple indices and other data - I'll check out the slayer ?
 
Oh, what the hell...

Here's something for nothing. The following MS code replicates my indicator very closely. I'm not going to divulge the exact workings, but any competent coder will be able to tweak from this to get their own workable ideas:

Long:

Ref(Stoch(3,2) < 10 AND L < Ref(LLV(L,100),-1),-1)
AND C > Ref(C,-1) AND C < Ref(H,-1) AND
V > Mov(V,50,E) AND OBV() > Ref(OBV(),-1)


Short

Ref(Stoch(3,2) > 80 AND H > Ref(HHV(H,20),-1),-1)
AND C < Ref(C,-1) AND V > Mov(V,50,E) AND
OBV() < Ref(OBV(),-1) AND Ref(OBV(),-1) > Ref(OBV(),-2)

Here is the Ami code for the long side:

Buy = Ref( StochK( 3, 2 ) < 10 AND Low < Ref( LLV( Low , 100 ), -1) ,-1 )
AND Close > Ref( Close , -1 )
AND Close < Ref( High , -1 )
AND Volume > EMA( Volume , 50 )
AND OBV() > Ref( OBV() , -1 );


Now for the short side.

CanOz
 
Rusty,

Keep doing what you're doing. You won't find a 100% accurate indicator because at the end of the day, it is people who move markets and people can and will do unpredictable things. But your research will allow you to find exploitable edges. That is worth the effort.

For example, I developed my own indicator which I have facetiously called the Holy Grail. When it works, it works. When it doesn't, I know very quickly. I live trade with it.

Examples from LGL. Long/short entries are shown. Exits are not shown:

Here is the code in Amibroker:
Buy = Ref( StochK( 3, 2 ) < 10 AND Low < Ref( LLV( Low , 100 ), -1) ,-1 )
AND Close > Ref( Close , -1 )
AND Close < Ref( High , -1 )
AND Volume > EMA( Volume , 50 )
AND OBV() > Ref( OBV() , -1 );

Sell = Ref( StochK( 3, 2 ) > 80 AND High > Ref( HHV( High , 20 ), -1) ,-1 )
AND Close < Ref(Close , -1)
AND Volume > EMA( Volume , 50 )
AND OBV() < Ref( OBV() , -1)
AND Ref( OBV() , -1) > Ref( OBV() , -2);

and the chart.....
 

Attachments

  • Volume based Scan LGL.jpg
    Volume based Scan LGL.jpg
    209.7 KB · Views: 17
M/S some of the code seems incorrect---you've done this on purpose.

Post ed
Sorry my fault cut off the cut and paste too quickly.
 
Hey Can,

Hope it's useful to you. :)

LOL, yeah right!

I'm playing with a volume filter now and scanning US stock, so its fun anyway!

Just thinking of how to enter, where to out the initial stop, and how to trail a stop.

The volume filter makes a big difference on the US stocks, too many false signals on lower volume stocks without it.

Cheers,


CanOz
 
MS/Tech - You want to get out of this quickly if you're wrong, right? I'm thinking an entry on the next open, with an initial stop of the previous bars high or low. Got to think of a trailing stop now.

CanOz
 
MS/Tech - You want to get out of this quickly if you're wrong, right?

Yes. The high/low of the setup bar can be a good reference point.

The setup bar is the first line of the short/long signals. The remainder of the code for each side is the trigger bar. I enter on the close of the trigger bar.

Only difference is I use my own indicator instead of the stoch. But the difference in live trading is fairly minimal. I've used the stoch live as well.
 
Yes. The high/low of the setup bar can be a good reference point.

The setup bar is the first line of the short/long signals. The remainder of the code for each side is the trigger bar. I enter on the close of the trigger bar.

Only difference is I use my own indicator instead of the stoch. But the difference in live trading is fairly minimal. I've used the stoch live as well.

Cool, thanks for the help MS, i think i'll use a 25 EMA cross for the trailing stop initially, just to test. That's if i can get a system written around it today.

Cheers,


CanOz
 
Sorry See quote above.

Oh ok. Sorry, I didn't understand at first. Cheers. :)

Edit: Can,

I should say that the high/low of either the setup or trigger bar can be used as a reference. Whichever is the higher/lower. Of course, one can do whatever. Everyone trades differently which is also why I haven't indicated how I exit. The entries are what they are and can be adapted to different styles.
 
Top