Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
Hi guys,
This is my first post on this platform.
I am a beginner on algotrading and trying to develop some strategies to trade crypto.
So I need more experienced people to guide me. I did some backtest, optimization and after that I decided parameters and ran my strategy for my full data.
For 48 crypto pairs, after 4113 trades my strategy have 3.60 RR per trade and a 22.5% win rate. SQN is 9.85. This figures are for only long positions.
My question is; what else parameters should I check to become sure about that strategy. (I used 12h timeframe)
Can you help answer these questions from other members on NexusFi?
Have you back tested and optimised on all your available data or just part of it?
As in did you keep back some data that wasn't used in the back test and optimisation to use for forward testing on out of sample data to see if those results are similar to the back test results.
So if you have say four years of data, back test and optimise for the best settings on the first three years then run those settings on the most recent last year of data and see if the results are comparable. Then try the other way round and optimise on the last three years and test on the first year.
Also do a Monte Carlo analysis of the results where your equity curve is rerun with the order of trade results changed so they are mixed up and you can see what would happen, ie, the maximum drawdown if the test starts during a period of unfavourable market conditions.
Both these things will check the robustness of your results and help to avoid over optimisation and provide more confidence for when you do actually start running the system on live data.
I won't say anything else because that pretty much covers my knowledge of back testing as it is something I looked at a while ago, read a few books on and tried out some testing, but then realised I needed to be testing on Daily bars and longer term trading to see any decent test results, and concluded I didn't have the account size for that so didn't pursue it further at that time.
Your stats look good so hopefully somebody can offer you some more experiential advice. Good luck.
You do not win as a trader, you just get to play again the next day. If that game doesn’t appeal to you then you should not trade. Gary Norden
Interesting approach!
You tell us that you use the 5 fold Ichimoku Kinko Hyo in a 12h time frame.
As I am using Ichimoku in all my trading - I have some detailed questions:
Ichimoku is a cool indicator with live, lagging and future lines. Can you tell more about which part(s) of Ichimoku you are using for the tests?
Because that influences a lot the prediction of a price development. Second it makes back testing very very difficult when you are not aware of the different times that indicator uses and shows in a chart.
Second: As a statistic expert expert I am not sure if you can treat the results out of history as Ichi does not show 3 different time lines in the chart...
On my side I am using patterns instead to reveal development in the chosen instrument.
If you are interested just search for my journals - there are 2 with bitcoin as well.
The most recent Ichi I used was the DAX trade of today with a take profit on the Chikou Span. It worked. But if you look back on a chart - you will never find the turning point again - as the Ichi doesnt offer a reading hindsight.
Hope to get some more insight into your using Ichimoku!!
Happy New Year
GFIs1
PS: Backtesting and walking forward is a waste of time anyway. You need a real time approach for every instrument and indicator. Best to open a journal here to discuss it with the best members here on the fio board. LIVE!
I have nearly 4 years of data. I optimized parameters with running on 3 years of data. After deciding the optimized values of parameters, I ran it for all 4 years data. Because I thought that 1 year data is not enough to test but I ll try to run it for the rest of the data as you said, thank you.
Monte carlo simulation is my next step, that is in my agenda. I hope it will go well on Monte Carlo. Thanks for your advises.
Yes, that is exactly what I said.
I'm using all 5 components of Ichimoku (tenkan, kijun, chikou, senkouSpanA, senkouSpanB).
I wrote some code on python for backtest, so I'm not using charts. Code is calculating values of these 5 component for each bar and then applying my rule set. I set up a very tight stop loss (%1) with a maximum 1% risk of capital, that is why my win rate is low like 22%. But I think average risk reward of results is promising.
I will try to test it with live data during next month, then I can share results of that if you interest.
Thanks for your comments.
Thanks for your details. I am just wondering if you are considering the lagging and future data on the components besides the actual one - this doesn't come out from your answer. If so - CHAPEAU!
Good trades!
GFIs1