Weight averaged ADXR strategy

gabri · 2017-11-06T10:14:15+00:00

This strategy is based on the modified version of the Directional Movement indicator I posted here. The strategy has to be backtested but it always gives above 70% of winning trades with a decent gain/loss ratio. The average gain per trade can be improved by changing the stop loss and the ADXR threshold. Blue skies. //tipomm=2 //x=5 - I normally use a number between 5 and 10 DEFPARAM CumulateOrders = False // Entering long ignored, ignored, mioDI, ignored, mioADXR, ignored = CALL "PRT - ADX e DI"[14, 2] c1 = (mioADXR > 50) c2 = (mioDI > 0) IF c1 AND c2 THEN BUY 3000 CASH AT average[x,tipomm](low) limit ENDIF // Exiting long IF longonmarket THEN SELL AT average[x,tipomm](high) limit ENDIF set stop %loss 15

Viewing 11 posts - 1 through 11 (of 11 total)

Author

Posts
11/06/2017 at 11:14 AM #51604 quote
gabri
Participant

Master
This strategy is based on the modified version of the Directional Movement indicator I posted here. The strategy has to be backtested but it always gives above 70% of winning trades with a decent gain/loss ratio. The average gain per trade can be improved by changing the stop loss and the ADXR threshold.

Blue skies.
```
//tipomm=2
//x=5 - I normally use a number between 5 and 10
DEFPARAM CumulateOrders = False 

// Entering long
ignored, ignored, mioDI, ignored, mioADXR, ignored = CALL "PRT - ADX e DI"[14, 2]
c1 = (mioADXR > 50)
c2 = (mioDI > 0)

IF c1 AND c2 THEN
BUY 3000 CASH AT average[x,tipomm](low) limit
ENDIF

// Exiting long

IF longonmarket THEN
SELL AT average[x,tipomm](high) limit
ENDIF
set stop %loss 15
```
Nicolas thanked this post

trend-e-mm-1509956523c48pl.png PRT-ADX-e-DI-1.itf
11/06/2017 at 11:15 AM #51631 quote

Nicolas
Keymaster

Legend

Hi Gabri, thanks for your contribution for the benefit of everyone here! About this specific strategy, I’d like to know a little more: _ what instrument, spread and / or fees? _ what timeframe? _ optimized settings: did you try WFA?

11/06/2017 at 12:31 PM #51652 quote

gabri
Participant

Master

Nicolas, I use stocks (italian, german and french), daily timeframe and no spread on the strategy I posted. I used this system for several months in real trading (manual) and got always more than 70% winning trades and better gain/loss ratios than the one published. I didn’t WFA. Gabri

11/06/2017 at 1:24 PM #51655 quote

Nicolas
Keymaster

Legend

Fine!, but how do you define what stocks should be a good candidate to be traded by the strategy?

11/06/2017 at 2:22 PM #51666 quote

gabri
Participant

Master

I pick stocks with high CSI or ADXR. I take this data from the indicator I posted.

11/06/2017 at 2:32 PM #51667 quote

Nicolas
Keymaster

Legend

I made some backtests and I found it works way better with a major trend filter, since this is a long only strategy with a sorta ‘mean reversion’ theory, what do you think? (trading only if Close>Average[100-200] ..).

11/06/2017 at 4:14 PM #51673 quote

gabri
Participant

Master

Nicolas, I was using it just for small and short trades and I am sure that there are better way to use it. Every input is more than welcome. How would you code it with your system?

11/06/2017 at 4:37 PM #51682 quote
Nicolas
Keymaster

Legend
I coded it this way: (trendperiod is an optimized variable), a 100 days period is a good compromise between performance and trades quantity..
```
DEFPARAM CumulateOrders = False

tipomm=2
x=5 //I normally use a number between 5 and 10


// Entering long
ignored, ignored, mioDI, ignored, mioADXR, ignored = CALL "Mio - ADX e DI"[14, 2]
c1 = (mioADXR > 50)
c2 = (mioDI > 0)
c3 = close>average[trendperiod](close)

IF c1 AND c2 and c3 THEN
BUY 3000 CASH AT average[x,tipomm](low) limit
ENDIF

// Exiting long

IF longonmarket THEN
SELL AT average[x,tipomm](high) limit
ENDIF
set stop %loss 15
```
11/06/2017 at 7:23 PM #51728 quote

gabri
Participant

Master

Nicolas, I think your addition it’s quite interesting but I personally think that launching a screener first (CSI or ADXR) and then running the code on the first 10-15 stocks of the list (building a portfolio) has better results. I quickly tested your version on a portfolio and seems to be carrying lower profits. I like the fact though that you have better gain per trade. Gabriele

11/06/2017 at 8:30 PM #51729 quote

Nicolas
Keymaster

Legend

Ok 👌 What about the rebalancing of the portfolio made with this kind of stocks pick?

11/06/2017 at 8:43 PM #51732 quote

gabri
Participant

Master

Good question. I don’t use too much this system anymore but I used to rebalance weekly. I always keep between 20 to 35 stocks in a short trading portfolio. When I sell I am not rushed to find a new stock to replace the old. I just run this screener (or some other screener focused on short term trading) once a week and if I find a good stock I add it.
Author

Posts