Try backtesting a portfolio of strategies in Ninjatrader. Let me know how that goes for you.
The missing feature from every back tesing platform
- Thread starter fan27
- Start date
Works quite well for me. Running fully automated, multi assets, CTA program on it.
Perhaps the version 8 is better than what I was using in 2009, but since I do much hedging and options, Ninja trader didn't back then or now, it also didn't allow to plug in when reports around the world were due and I use automation to trade these reports, and before spread intraday relied on making an indicator to day trade options and that was herky jerky. Plus, at the time you could only put in 500 symbols. But they put out what general public requires.
I am not trying to knock NinjaTrader and am glad it is working for you. But to state that their are no limits with NinjaTrader is WAY off the mark. The platform does not support portfolio level back testing as in testing multiple strategies simultaneously and being able to set risk controls based on the activity in all strategies. See this post:
https://ninjatrader.com/support/for...ktesting-multiple-strategies-in-sync?t=110189
Count yourself as fortunate in that you found an inexpensive retail platform that meets your requirements. I had to create a research platform from scratch as I have specific requirements that could not be met otherwise.
Ninja does not support options, even 8 as far as I know. So if options is what you need, but not be for you. Same for 500 symbols.
I understand what you mean, when I said there is no limit, I meant c# coding. You can program anything you want. I achieved what OP in the link you posted was asking, by merging results via custom code.
Don't get me wrong, NinjaTrader is excellent for most traders whether longer term or scalping, I bought a copy long ago. But once you step outside of what the masses require, you have to build your own. Yes, the 500 limitation is way too small for automation, but for most it is just fine.
Good trading all
I met someone a while back who automated 100s of symbols via R. Perhaps it will work for you.
I already have mine in place, hired in late 2009, took them years to get what I needed. Mixture of python and couple other languages don't remember, one does only data crunching etc.. Have over a dozen systems automated and running. Many have said I over paid, most likely, but I got what I needed and that bottom line. Now have partner and he does all the programming, I don't have the skills any more to program, health does ruin areas, but I design risk management in ways I can understand and programable. Except for scalping where entries have to close to perfect, signals generally don't need perfection when you hedge 30 minute or larger timeframe bars. So I do mixtures of options and futures to hedge stocks/etf's. And I still manually trade new ideas to see if worthy to get programmed. But generally always working on ways to lower drawdowns of intraday and longer term, if you not losing much, profits take care of themselves, all we can really do in trading is control losses and not profits.
Short post as I am on holiday (as you know) but I could not resist:
As long as you have a static methodology how you evolve new strategies over time you backrest over any dataset of your choosing. So, for example, if you decide to evolve new strategies every 6 months or upon a certain data trigger then have your back tester do that and generate your performance and risk metrics off generated trades. In market parlance this is called model re-calibration. Whether you do that manually or fully automated makes no difference. In essence your trading strategies are just another adaptive set of model parameters. Is that somewhat answering your question or did I miss the mark?
As long as you have a static methodology how you evolve new strategies over time you backrest over any dataset of your choosing. So, for example, if you decide to evolve new strategies every 6 months or upon a certain data trigger then have your back tester do that and generate your performance and risk metrics off generated trades. In market parlance this is called model re-calibration. Whether you do that manually or fully automated makes no difference. In essence your trading strategies are just another adaptive set of model parameters. Is that somewhat answering your question or did I miss the mark?