Mother of Dragons trading strategy…
-
There are just too many conditions over too many timeframes to all be satisfied at the same time … hence no trades.
In periods of high volatility maybe we get trades, but we would get old waiting? 🙂
Anyway attached .itf gives trades today and Monday this week, but only Forward Testing on Demo will show if it is any good. Timeframe 10 seconds, spread = 4.
Unable to launch in automatic trading and I can’t find wat’s wrong … If anyone can help.
I do. Just take 2 position s since 1th of feb.
Hello guys,
I worked on the DJ 5 min MOD v5.1 in Long only. I optimize most part of paramters and reduce quantity of timeframe.
I also tested the code with and without trailing stop, breakeven and exit at win/loss cut. I observed that the trades were cut very often too soon. The best comprise i found was conserve only the exit at loss.
You can find in 50k the results of basis MOD 5.1 and the version i worked. I also join the worked version in 200k
Not so bad in my opinion.
r0d0lphe thanked this postwhen you optimize something please tell the way you have done it,
how much IS data have you been optimizing on?
its not hard to get something to look good in the backtest.
79 trades over 3 year on 5 min TF is a very small sample of trades over that timeperiod in my opinon, hard to clarify an edge with small number of trades
Hello snucke,
I work in long only and with 50k data. I check how it works in 200k and result are very similar as you can see in pictures.
I change some timeframe : the 2h become 1 hour. I delete the time frame 30min and 10min. I conserve the 15 and 5 min. So i use only 1h, 15min and 5min. All the original parameters are optimize.
In the original code, there are many way to cut the position : exit in profit, exit in loss, trailing stop, breakeven, NewSL, …. I test one by one what happens (except the NewSL), and i observe that the position are cut very fast and give a % win trading very interesting but win less money. Most importantly, all this panel don’t stop the big loss. There is only the exit at loss how limit the big loss so i only conserve it in the new program i suggest.
The other big difference is the time in the market : 25% time with some hours by trade in the original code against 41% time and 4 days by trade in the new version. That’s with there is not so much trade in the new code. The trades stay more time in market to get the 2.6% take profite.
I also test the new code in 1M barre. I join you the result. It’s quite good except 2015. But in my opinion, market evolve so not sure it’s so much important result have to be perfect 10 years ago.
Hope it gives you more informations.
Enjoy
GraHal thanked this post
- You must be logged in to reply to this topic.
Author
Summary
This topic contains 522 replies,
has 50 voices, and was last updated by 
4 years, 10 months ago.
Topic Details
| Forum: | ProOrder: Automated Strategies & Backtesting |
| Language: | English |
| Started: | 01/21/2020 |
| Status: | Active |
| Attachments: | 195 files |
