Reading a trade with VSA: A working example

[Gringotts Bank]

Yes you did and I recognize that, but you also said VSA systems won't give you simple buy and sell arrows. This very simple system, still unoptimized, does it all for ya! I'm sure you could get 50%pa out of it with a bit of extra work.

If anyone has AB and a screaming fast PC, can you please optimize this (below) or even for 10 minutes?

//!!Simple VSA - Gringotts Bank, Aussie Stock Forums!!//
OptimizerSetEngine("cmae");
SetBacktestMode( backtestRegularRaw );
SetOption("MaxOpenPositions", 100 );
PositionSize = 20000;
PositionScore = 100+V;
a = TEMA(C,Optimize("A",14,1,100,1));
aa = a<Ref(a,-1);//downtrend established
b = C>O;//
d = V-Ref(V,-1)>0 AND (V-Ref(V,-1))/Ref(V,-1)*100>Optimize("B",50,1,200,1) AND V>JurikJMA(V,600,-500);//+ve volume spike
Filter = Buy = aa AND b AND d AND C*V>500000;
Sell = V>Ref(V,-1) AND ROC(C,1)<0 AND (V-Ref(V,-1))/Ref(V,-1)*100>Optimize("C",40,1,200,1);
AddColumn(V*C,"IVT");//


And I agree. I'm not a systems guy. I much prefer gut feel. I won't be using this system, but others might enjoy it. It's just an intellectual exercise.

 

[Gringotts Bank]

It works in all market conditions. Have a look at 2008 - very nice profits going long.

Some people will say it's stupid to post up the bare bones of a really good system, but as you know, sticking with a system is bloody hard. 99.9% of people won't have the discipline to trade it, me included. That's why I say i won't be using it myself.

But if anyone wants something that will return 50%pa, without spending 10 years developing a system, well here it is. Free!

 

[tech/a]

What did you run it on--Data base?

 

[Timmy]

Thanks both of you for posting.

Great posts! I think many others are interested too!

Yep, good trade and good thread, thanks Tech.

 

[Gringotts Bank]

What did you run it on--Data base?

ASX 2007-2011 blue chips, so need not worry about slippage. The final code I used is quite different, because I worked on it for a few hours last night. Here's the equity curve.

 

[tech/a]

Must get my Amibroker up and running.
Got it but never cranked it up as I use metastock.
If I had it going I could couple it with Tradesim and run some Montecarlo.

 

[Gringotts Bank]

If I use Positionscore = Random(); and run it heaps of times, all the results are all >25%, and mostly above 45%, but yeh a monte carlo would be a good test for it.

The code is so basic, you could almost transfer it straight across to metastock.

It lends itself quite well to 'trading the equity curve'. It would be the sort of thing I'd use if ever automation was possible for retail traders, otherwise I couldn't be bothered.

You can bet there are lurkers reading this who have gobbled up all this info for free and run out the door, hoping no one says another word!

PS. still running them manually, just got an 18%, but holding up well with >40%

 

[pavilion103]

Not sure if this is the best place to post it, but wasn't going to start an entire thread for this.

I've been doing testing over hundreds of charts and am becoming more and more familiar with VSA bars/setups.

I don't have any ambition to try what I'm about to suggest, but I wonder what the results would be from a system that entered on 'no supply' bars (before the close of day) and then exits 1 or 2 days later. It is amazing how often these bars result in price shooting up the next day.

Most trades I enter on seem to be on no supply bars. But I'm wondering if there was a system where you enter on a large number of these opportunities, regardless of the background and held for only 1 or 2 days, whether this would yield positive results.

Emphasising, I am purely curious, not interested in actually doing this.

Do people think it would be profitable?

 

[Billyb]

It works in all market conditions. Have a look at 2008 - very nice profits going long.

Some people will say it's stupid to post up the bare bones of a really good system, but as you know, sticking with a system is bloody hard. 99.9% of people won't have the discipline to trade it, me included. That's why I say i won't be using it myself.

But if anyone wants something that will return 50%pa, without spending 10 years developing a system, well here it is. Free!

Gringotts,

I dont have the Jurik JMA code so I replaced it with a normal EMA which I presume would give reasonably similar results.

GREAT, in fact FANTASTIC results with no trade delay. Perfect equity curve, I've never seen anything better.

But when you you make it realistic and incorporate a one bar trade delay - it loses money.

Did you find the same?

 

[tech/a]

PAV



Ive had 3 FTSE trades tonight and currently opened a 4th.
See if you can pick them? And my stops.



Ill post up answer in an hr.

 

[tech/a]

Off to bed. Had a good night.

 

[pavilion103]

I didn't get a chance to have a crack before the answer was put up! Ah well.....


What hours does the FTSE trade Australian time?

I'm potentially starting a new job, and this could mean I can set up an account with IB. I'm wondering if it would be a good idea for me to begin trading after work (if the market is open then).

Is it much different to trading equities? Obviously the same VSA principles apply.

What are the differences?

 

[Gringotts Bank]

Gringotts,

I dont have the Jurik JMA code so I replaced it with a normal EMA which I presume would give reasonably similar results.

GREAT, in fact FANTASTIC results with no trade delay. Perfect equity curve, I've never seen anything better.

But when you you make it realistic and incorporate a one bar trade delay - it loses money.

Did you find the same?

Hi Billy, I didn't see this post till today.

What would stop you entering at close same day? Assuming you had live data and could scan at close for entry to appropriate stocks.

 

[tech/a]

Hi Billy, I didn't see this post till today.

What would stop you entering at close same day? Assuming you had live data and could scan at close for entry to appropriate stocks.

GG
The problem is that when testing we have the advantage of hindsite.
So EVERY trade will be shown as a winner with the entry on the same day.
Those that dont qualify wont be shown.
In real time however most buy triggers cant be confirmed until close out.
Just have a look at your problem with your method.

Often a buy triggered during the day wont be there at close.

 

[Gringotts Bank]

GG
The problem is that when testing we have the advantage of hindsite.
So EVERY trade will be shown as a winner with the entry on the same day.
Those that dont qualify wont be shown.
In real time however most buy triggers cant be confirmed until close out.
Just have a look at your problem with your method.

Often a buy triggered during the day wont be there at close.

Yes but if a system says 'buy at close', and the scan is performed at close (4pm) and the SP is well within the bounds of the system's requirements, then that should allow one to enter at the closing auction without affecting the reliability.

 

[Billyb]

Yes but if a system says 'buy at close', and the scan is performed at close (4pm) and the SP is well within the bounds of the system's requirements, then that should allow one to enter at the closing auction without affecting the reliability.

But your buy criteria relies on the Close price from THAT day. You do not know the close price in real life until AFTER the close. IMO this means real results probably wont align with testing results.