strategy BarHunter DAX v1p

Viewing 15 posts - 16 through 30 (of 256 total)

← 1 2 3 … 16 17 18 →

Author

Posts
01/16/2020 at 7:24 PM #117013 quote

Francesco
Participant

Veteran

Last update with your adjustments @paul on nikkei. Tested on 200K, spread on 8.

NIKKEI200K.jpg NIKKEI200K2.jpg

01/16/2020 at 8:46 PM #117016 quote
Paul
Participant

Master
Every year positive on the Nikkei! I’ve seen a lot worse tests. Can you post your changed settings so I can have a look?

On the south african 40.

I increased the spread to 20. As with the nikkei I want to take a different approach to handle that, but that’s for later.

So it’s basically the file you got, activated trailing stop, stoploss 2, profittarget 2

break=20 & only one bar is analysed for long & short in range to 23
```
once steps=0.1
once minatrdist=4

once atrtrailingperiod    = 14   // atr parameter
once minstop              = 10   // minimum  distance

if barindex=tradeindex then
trailingstoplong     = 8   // trailing stop atr distance
trailingstopshort    = 8   // trailing stop atr distance
```
It looks good without WF and it still looks good with WF. (often it breaks down completely) A lot of time the 3 bar is the one, with very few exceptions.

Ofcourse could be better, but for 1 variable in range to 23 and with this simplicity its great base!

edit;WF

1e bar 11

2e bar 3

3e bar 3

4e bar 1

all others; barnumber 3

So it’s pretty consistent

Francesco thanked this post

Screenshot-2020-01-16-at-20.29.15.jpg Screenshot-2020-01-16-at-20.31.39.jpg Screenshot-2020-01-16-at-20.52.09.jpg
01/16/2020 at 8:58 PM #117020 quote

Paul
Participant

Master

couldn’t resist, the difference to WF other method.

Screenshot-2020-01-16-at-20.56.47.jpg Screenshot-2020-01-16-at-20.57.04.jpg

01/16/2020 at 9:59 PM #117023 quote

Francesco
Participant

Veteran

I’ll take a look later to the last update of South Africa, the results looks amazing!
Here you have the .its of NIKKEI 1H, i’m not an expert in coding, not even remotely; so probably my attempts to diversify the strategy are futile, but I’m just trying to do something productive to stimulate development 😀

Paul thanked this post

BarHunter.itf

01/17/2020 at 10:27 AM #117043 quote

GraHal
Participant

Master

Any of you guys got attached Reject at 4:00 am this morning on the V1p version on the DAX?

Or at any time, any version, any Market?

What is the fix please? How to code it into the Strategy?

Would 2 x different min distances be best … 1 x for Market Hours and 1 x for Out of Market Hours?

Thank You

Paul-10.jpg

01/17/2020 at 11:43 AM #117048 quote

MAKSIDE
Participant

Veteran

Hi Grahal,

I will test also this strategy asap

but i think the ddmax is too important..

01/17/2020 at 11:52 AM #117050 quote
GraHal
Participant

Master
I changed below to 12 (from 10) to see if it makes any difference?

For info: I set below temporarily to 100 (to test the boundaries) but the Long trade was still executed at 4:00 am on backtest.
```
once minstop              = 12   //10 // minimum  distance
```
01/17/2020 at 11:58 AM #117052 quote

GraHal
Participant

Master

MAKSIDE wrote:

i think the ddmax is too important..

Are you referring to Maximum Drawdown?

Too much / large at around £1k (for lot size =1) over 100k bars over around 4 Trades?

01/17/2020 at 12:35 PM #117063 quote

MAKSIDE
Participant

Veteran

Yes,.. for me, everyone manages their risk. In general, I try to have the smallest drawdown even if it is to have less performance. PRT is not good for backtesting by drawdown result.. maybe one day..

01/17/2020 at 12:43 PM #117068 quote

MAKSIDE
Participant

Veteran

and.. for me, the perfect system is to have a drawdown as little as possible and to be very very little on the market. But nothing is perfect 😉

01/17/2020 at 12:45 PM #117070 quote

Francesco
Participant

Veteran

@makside Gotta risk it for the biscuit 🙂

01/17/2020 at 2:47 PM #117083 quote
Paul
Participant

Master
@GraHal Your are quick in testing! It’s a good find but not so good in terms of results at first glance.

As I understand it, the error occurs when the stop is getting placed but is too near to the current high/low.

you can try this. If it still gives an error, increase the minstopdistance slightly.
```
if longtrading or (longtrading and shorttrading) then
minstopdistance=10 //minimum stop distance between stop and current price
if intradaybarindex=barnumberlong then
breakvaluelong=high //experiment!
if high-close<minstopdistance then
breakvaluelong=close+minstopdistance
else
breakvaluelong=breakvaluelong
endif
endif
endif

if shorttrading or (longtrading and shorttrading) then
if intradaybarindex=barnumbershort then
breakvalueshort=low //experiment!
if close-low<minstopdistance then
breakvalueshort=close-minstopdistance
else
breakvalueshort=breakvalueshort
endif
endif
endif
```
01/17/2020 at 3:30 PM #117088 quote

grimweasel47
Participant

Senior

Hello – I’m just trying to test this but I keep getting error ‘backtest exceeds limit of repetitions for Walk Forward optimization’ – but I haven’t got WF selected??

Thanks

01/17/2020 at 3:37 PM #117089 quote

Paul
Participant

Master

@grimweasel47 Hi, Adjust the repetitions to 3, even if you don’t use them. That should fixes it.

01/17/2020 at 6:59 PM #117106 quote

GraHal
Participant

Master

Paul wrote:

As I understand it, the error occurs when the stop is getting placed but is too near to the current high/low.

Yes above is the reason for the Reject error.

I have added the code-fix you kindly and quickly provided and I will run a version without the fix and with the fix to validate the settings etc.

Many Thanks
Author

Posts