strategy BarHunter DAX v1p
-
Hi
The South African concept looks interesting and I thought exotic but the index is surprisingly heavy in global names, perhaps there is a requirement for a local company to be used the trade there?
Looking at the back tests, while the larger data set is great for validating systems one of my pet hates is results that are shown over huge time frames as the equity curve tends to straighten out, you get a huge return and may lose site of what you are trying to achieve.
The system (for Dax at £1 per point) has a 300% return (based on £10K bank) over 10 years given the assumptions used about the spread and funding costs at the time and leverage of 1:20, however the index itself without any leverage is up around 220% over the same period without any leverage and presumably less risk. (I understand that you would not have needed £10k to open the positions)
I don’t have the data to calculate this but I would expect that of you had bought long term options on the Dax that it would have been possible to capture many multiples of the 220% return with an even lower risk than that of using CFDs/spreadbets
The system is profitable in periods and merits further work but in its present form it may not be the best use of capital given the time in the market and not showing a significant out performance of the underlying market in spite of leverage
Thank you for the latest improvement @Paul.
Do you get the same as me … the profit falls??
The attached equity curves are shown as …
- v1p + mindistv2 (6.23K)
- v1p +mindistv1 (10.6K
- v1p (10.6K)
@GraHal Yes results go down. Maybe it’s because it’s calculated correctly unfortunately? I’am still not sure.
Here’s a test update to the existing version. Basics is the same but it tests various ways of entries (and the original) with the minimum distance in mind. DAX 1 hour timeframe (100k bars) and this with spread=4 because the time of entry is expensive!
To cover a bit the minimum distance, I put breakpoint=5 and with the original settings it has the best results. No trend detection yet. (entrytype=1, barnumber=3)
If going for stoploss 2% instead of 1, results increase more.
added a nice addition to the trend detection system thanks to @nonetheless ! (from other topic)
Changed medianprice to totalprice and that worked good! Kept both.
// trend detection if tds=0 then trendup=1 trenddown=1 else if tds=1 then trendup=(Average[50](close)>Average[50](close)[1]) trenddown=(Average[85](close)<Average[85](close)[1]) else if tds=2 then bbup=BollingerUp[20](close) bbdn=BollingerDown[20](close) bbav=(bbup+bbdn)/2 trendup=bbav>bbav[1] trenddown=bbav<bbav[1] else if tds=3 then Period= 3 inner = 2*weightedaverage[round( Period/2)](typicalprice)-weightedaverage[Period](typicalprice) HULL = weightedaverage[round(sqrt(Period))](inner) trendup = HULL > HULL[1] trenddown = HULL < HULL[1] else if tds=4 then Period= 2 inner = 2*weightedaverage[round( Period/2)](totalprice)-weightedaverage[Period](totalprice) HULL = weightedaverage[round(sqrt(Period))](inner) trendup = HULL > HULL[1] trenddown = HULL < HULL[1] endif endif endif endif endifpic below with stoploss 1% comparable to first pic above.
with the original settings it has the best results.
You have over twice the number of trades as me and over twice the profit … any ideas how that can be?
here’s a new one. Since the 10 points difference is so important I made an adjustment to be dependent on the breakpoint number.
So if breakpoint =10 or higher, the minstopdistance=0
breakpoint=7, minstopdisntance=3
breakpoint=0, minstopdistance=10
If you select i.e. entrytype=1, then minstopdistance is ignored, which isn’t the best/reliable way.
pic is with stoploss 2% & with entrytype1, breakpoint5, barnumber3, trend4, spread4
Francesco thanked this postWe managed to place Stops this morning so maybe you have got the code and levels correct now Paul?
Paul thanked this postWorks well on TF = 5 mins also … just as you thought!
- You must be logged in to reply to this topic.
Summary
This topic contains 255 replies,
has 11 voices, and was last updated by 
2 years, 6 months ago.
Topic Details
| Forum: | ProOrder: Automated Strategies & Backtesting |
| Language: | English |
| Started: | 01/15/2020 |
| Status: | Active |
| Attachments: | 136 files |
