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)
Any way in MC to optimize on a list of instruments?
Anyone found out a way in MC to optimize a signal in MC on a list of symbols so that it would optimize the signal with different inputs for each symbol?
In NT, despite NT's all shortcomings, this can be done but I havent found a way in MC to do it.
Can you help answer these questions from other members on NexusFi?
The problem arises when I have quite a generic strategy that I would like to run on a lot of stocks and I would like to find on which stocks it works best with small variations in settings for every stock.
I wonder are you trying to optimise on many symbols at once just for convenience? I assume this is the case and you are trying to optimise on a large number of symbols and don't want to have to load up each chart and run optimization manually for each one.
I am pretty sure you can't do this in PBT, but would be a handy feature I would vote for Simpler to implement would be some kind of scripting capability in PBT. Your task be a cinch with a shell script provided backtester could be run from the command-line.
I am not doing it for convenience, its not really for optimisation purposes. Can be more viewed as data-mining, based on these results I know at which symbols I should start looking at more in depth.
Scripting possibilities would solve it, was even thinking about adding the logic into the signals themselves somehow but that might be a bit too hard.
Ok I did think of some outlandish hacks but no idea if they will work without knowing more specifics. You might be able code it so there is different optimization variable for each symbol and the logic for choosing the variable is set in the strategy code. Still messy though
Another crazy idea is a windows keyboard and mouse macro to automate the repetitive task of optimizing each symbol on the chart
Tnx, I had something like these messy hacks in mind
For the time being I settled for this solution:
1. I modified the strategies a bit to be more generic.
2. Those parameters that didnt allow more generalisation i split up in 3 intervals.
3. I run these more general strategies in PBT as a backtest in a couple of batches, due to the parameter interval splitting.
4. Those symbols' results that look interesting I start to look at more in depth and analyze further.
PBT doesnt allow for BID/ASK testing etc. but this works as a crude acid test.
I feel quite satisfied with the solution and will use the approach in the future for similar problems.
But on the side I started also looking at other software platforms like openquant, Matlab etc.
A workaround I used in a previous life was the following.
In the strategy, add a line to print the results ("MessageLog" or equivalent), for instance under the following format:
Then apply the Porfolio Backtester to the whole portfolio, with an optimization of Parameter1 and Parameter2 (with the option "all parameter values" [I do not remember the name] and not "genetic optimization").
Then do not take into account the results of the optimization in the Backtester, but import the content of the log (all the printed lines) into Excel, and analyze the data the way you want.