Eliminating Repeated Trades

[GRANDEUR 10]

Recently, I tried to figure out how one can backtest day-strategy on smaller timeframes in order to eveluate an optimal stoploss. I was given a piece of advice to use a function 'entriestoday(date)<1'. It works fine to me.

Then I tried to implement something similar to an Hour-strategy. I tried to backtest it on 5-min timeframe. But, unfortunately, 'entriestoday(date)<1' doesn't work properly in that case.

Taking into account all written above, please help me to recode the following simple strategy so that I can backtest a stoploss amount:

 

inputs: x(100);

 

buy next bar at h stop;

setstopposition;

setstoploss(x);

 

sell short next bar at l stop;

setstopposition;

setstoploss(x);

 

Thank you in advance!

[Tams 20]

...

Taking into account all written above, please help me to recode the following simple strategy so that I can backtest a stoploss amount:

...

Thank you in advance!

 

 

if you can rewrite the above in the following format, I might be able to give you a hand:

 

1. write out your thoughts ONE logic PER LINE

2. write out your logic ONE action per LINE

3. write out your action ONE sentence per LINE

4. start a NEW LINE for every sentence

[GRANDEUR 10]

Sure)

 

There is a 60-min strategy:

Buy at the high of the previous 60-min bar on stop.

Sell at the low of the previous 60-min bar on stop.

Anytime there is an open position, place a stoploss.

That is it.

 

I want to eveluate the best stoploss value.

It's impossible to eveluate it on 60-min bars.

Various stoploss values must be tested on smaller timeframe bars.

What code will let eveluate an optimized stoploss value?

[Tams 20]

Sure)

 

There is a 60-min strategy:

Buy at the high of the previous 60-min bar on stop.

Sell at the low of the previous 60-min bar on stop.

Anytime there is an open position, place a stoploss.

That is it.

 

I want to eveluate the best stoploss value.

It's impossible to eveluate it on 60-min bars.

Various stoploss values must be tested on smaller timeframe bars.

What code will let eveluate an optimized stoploss value?

 

 

h[1] = high of the previous bar

 

EasyLanguage works on bar-by-bar basis; it doesn't matter what is the bar resolution...

[Tams 20]

...

 

I want to eveluate the best stoploss value.

It's impossible to eveluate it on 60-min bars.

Various stoploss values must be tested on smaller timeframe bars.

What code will let eveluate an optimized stoploss value?

 

 

what do you mean by "impossible" ?

 

your stop loss is too small?

it triggers too soon?

not triggered at all?

etc...

[Charlton 10]

Sure)

 

There is a 60-min strategy:

Buy at the high of the previous 60-min bar on stop.

Sell at the low of the previous 60-min bar on stop.

Anytime there is an open position, place a stoploss.

That is it.

 

I want to eveluate the best stoploss value.

It's impossible to eveluate it on 60-min bars.

Various stoploss values must be tested on smaller timeframe bars.

What code will let eveluate an optimized stoploss value?

 

You are using a position based stoploss, so if it is set to say $500 and the trade goes against you then it will be stopped when the total position has made $500 loss, regardless of the data compression (time period of bars used).

 

I suspect that what you are trying to do is to set the stoploss to be based upon the price movement expected in the time period (data compression) being used e.g. to base it upon the bar range or the ATR.

 

If that is what you are trying to do then it is probably easier (from a visualisation viewpoint) to use SETSTOPSHARE. This sets the amount per share or contract. The important point is that it does not need to be a fixed amount.

 

You can have an input or variable called, say, SL_AMOUNT. Then you have the code

 

SETSTOPSHARE;

SETSTOPLOSS(SL_AMOUNT);

 

This will set the stop loss per share/contract to be the amount the input or variable contains.

 

So you could set it to ATR

 

SL_AMOUNT = AvgTrueRange(10);

 

 

This would set it to the average true range based on 10 bars for the time interval (data compression) chosen in the plot. In other words the ATR and hence the SL_AMOUNT and hence the stop loss would vary according to the data compression.

 

You could also use an IF statement to set it to specific values based on bar type and interval

 

e;g. IF BARTYPE = 1 AND BARINTERVAL = 10 then SL_AMOUNT = 5;

 

This would set the stop loss amount to 5 for a 10 min intraday bar.

 

 

BARTYPE values are

 

0 = TickBar

1 = Intraday

2 = Daily

3 = Weekly

4 = Monthly

5 = Point & Figure

6 = (reserved)

7 = (reserved)

8 = Kagi

9 = Kase

10 = Line Break

11 = Momentum

12 = Range

13 = Renko

 

Personally however I would recommend using something like a volatility based stopped such as ATR which would automatically select stoploss levels at a basis appropirate to the time interval being traded. You don't have to think about it then.

 

Finally another possibility of course is to set a stop loss based on 1 time interval but to use it in a stategy based on another interval. For example you could have a strategy entered onto a 1 min chart, which uses an ATR taken from a 10 min chart, but that's a whole new thread on multiple data feeds and potentially GVs, ADEs and so on, which I don't want to go into at present.

 

 

Charlton

[GRANDEUR 10]

Thank you for your replies!

But it seems to me that you haven't understood what I was asking for. Let me express my question by images.

On the first chart there are 60-min bars. And the code for the strategy used is:

buy next bar at h stop;

setstopposition;

setstoploss(500);

sell short next bar at l stop;

setstopposition;

setstoploss(500);

 

On the second chart there are two data feeds: 5-min and 60-min. The code for the strategy is:

buy next bar at h of data2 stop;

setstopposition;

setstoploss(300);

sell short next bar at l of data2 stop;

setstopposition;

setstoploss(300);

 

So I want to trade on 60-min bars but I want also to verify wether my stoplosses are not executed inside 60-min bars.

 

On the third chart is a magnification of chart #2. One can see there that there were several trades inside one 60-min bar that shouldn't take place.

[Tams 20]

a picture is worth a thousand words !

[Tams 20]

[GRANDEUR 10]

Tams, thank you for your reply! But it seems to me, there is still some misunderstanding.

'Position limits' is used when you don't want to make buy #1 and afterwards buy #2, for examle. It doesn't solve the problem in my case: buy #1 then comes a stoploss (so there are no positions at the time), then comes another buy #1 and there comes another stoploss and so on ALL INSIDE ONE 60-MIN BAR.

I want that there would be no more than one buy/sell trade and one cover-trade inside a 60-min bar.

[Tams 20]

Tams, thank you for your reply! But it seems to me, there is still some misunderstanding.

'Position limits' is used when you don't want to make buy #1 and afterwards buy #2, for examle. It doesn't solve the problem in my case: buy #1 then comes a stoploss (so there are no positions at the time), then comes another buy #1 and there comes another stoploss and so on ALL INSIDE ONE 60-MIN BAR.

I want that there would be no more than one buy/sell trade and one cover-trade inside a 60-min bar.

 

 

I would suggest you to draw a flow chart to illustrate your logic.

 

the branches should include ALL THE PERMUTATIONS.

 

[Charlton 10]

Tams, thank you for your reply! But it seems to me, there is still some misunderstanding.

'Position limits' is used when you don't want to make buy #1 and afterwards buy #2, for examle. It doesn't solve the problem in my case: buy #1 then comes a stoploss (so there are no positions at the time), then comes another buy #1 and there comes another stoploss and so on ALL INSIDE ONE 60-MIN BAR.

I want that there would be no more than one buy/sell trade and one cover-trade inside a 60-min bar.

 

If you follow Tams advice of drawing out the logic think about:

 

(a) How you might count whether an initiating trade took place during a bar

(b) How you reset the count to zero when a new bar forms

© How you know if there are currently any open positions in play during this bar

 

 

Charlton

[GRANDEUR 10]

Guys, I'm sorry for my incomprehensible style of writing. I thought I expressed myself clearly but seems it was not.

So, here is another attempt of mine.

[Tams 20]

Guys, I'm sorry for my incomprehensible style of writing. I thought I expressed myself clearly but seems it was not.

So, here is another attempt of mine.

 

 

your picture is not showing.

 

you can load the picture directly to TradersLaboratory.

Just click the Manage Attachments button under the Advanced reply window.

[GRANDEUR 10]

Oh, thank you. I clicked 'Insert Image' and have already got tired of looking for file sharing services.

Here it is.

[Tams 20]

arh... that's very good description.

[Tams 20]

For TradeStation, get the ebook EasyLanguage Essentials Programmers Guide (free)

 

read up on Look-Inside-Bar-Back-testing (LIBB).

 

 

 

For MultiCharts, use Backtesting Precision. (same function as LIBB)

 

You are back testing on 60 min bars,

and using a finer resolution (e.g. 1 tick data) as back testing precision.

 

 

Edited December 14, 2009 by Tams

[GRANDEUR 10]

Tams, I am using Multicharts but I don't see the 'Backtesting Precision' tab in my 'Strategy Properties'.

Maybe, this tab appeared in a new version? I have 5.0 Beta.

[Tams 20]

Tams, I am using Multicharts but I don't see the 'Backtesting Precision' tab in my 'Strategy Properties'.

Maybe, this tab appeared in a new version? I have 5.0 Beta.

 

 

.......... time to upgrade !

[GRANDEUR 10]

I' m sorry for delays in my replies. I got a newer version of MultiCharts. It's become much easier with that new feature but it's not all obvious though. It doesn't work properly with some of my strategies. Now I'm trying to figure it all out for myself.

Tams, thank you very much for your help!

[Tams 20]

I' m sorry for delays in my replies. I got a newer version of MultiCharts. It's become much easier with that new feature but it's not all obvious though. It doesn't work properly with some of my strategies. Now I'm trying to figure it all out for myself.

Tams, thank you very much for your help!

 

 

.......... you are welcome !