Super cool. Thanks for sharing this. I'll have to give this a try.
As for alternate approaches for go-live: have you considered using the LEAN framework from QuantConnect? It's quite robust, and even though it's 3rd party, it's open source so you can see under the hood.
Thanks! Yes, I took a look at the LEAN documentation.
While reading their docs, I got the feeling of a "convention-over-configuration" paradigm (which is something I usually don't like).
Also, I found native IBKR API's documentation more developer-friendly to me.
In the end, it's just a matter of personal preference. At least for me, if I need to invest time and energy to learn something, I usually prefer going to the source instead of learning abstractions. But that's me :)
Great article. Question : how much deviation would you accept between backtest results and live testing results to conclude that this strategy is ready to go live? Thanks
That is a great question. I'm usually more strict with trade stats (expected return/trade, win ratio, payoff ratio, etc) and less with overall metrics (total return, Sharpe, etc). That's because these overall metrics vary a lot throughout history.
Perhaps please do an article on stats that matter in back testing and (or) forward testing. When should someone take a strategy live? What metrics matter?
I'm trying to start with backtesting, can you please share what is your system for backtesting? are you using a python lib such as QSTrader? I would love to have a similar article about your backtesting system. Thank you for everything!
I've used most of the open-source backtesters... after I realized I was spending way more time tweaking them than testing new ideas, I decided to write my own system. That was many years ago, while I was still in my master's degree.
I'm still figuring out what the best format would be for sharing it.
Beautiful code <3
Thanks :)
Super cool. Thanks for sharing this. I'll have to give this a try.
As for alternate approaches for go-live: have you considered using the LEAN framework from QuantConnect? It's quite robust, and even though it's 3rd party, it's open source so you can see under the hood.
Thanks! Yes, I took a look at the LEAN documentation.
While reading their docs, I got the feeling of a "convention-over-configuration" paradigm (which is something I usually don't like).
Also, I found native IBKR API's documentation more developer-friendly to me.
In the end, it's just a matter of personal preference. At least for me, if I need to invest time and energy to learn something, I usually prefer going to the source instead of learning abstractions. But that's me :)
Great article. Question : how much deviation would you accept between backtest results and live testing results to conclude that this strategy is ready to go live? Thanks
That is a great question. I'm usually more strict with trade stats (expected return/trade, win ratio, payoff ratio, etc) and less with overall metrics (total return, Sharpe, etc). That's because these overall metrics vary a lot throughout history.
Perhaps please do an article on stats that matter in back testing and (or) forward testing. When should someone take a strategy live? What metrics matter?
I have a side question, so you are Carlos Mata? 🧐
I’m Carlos Souza. Carlos Mata is a good friend, though :)
He is a quant trader too… and he also shares some of his strategies. However, he is more focused on futures, while I tend to focus on equities
Both Carlos are very good!
Thanks heaps!
Your best post! Keep it up!
Thanks :)
Thank you for all the information you provide!
I'm trying to start with backtesting, can you please share what is your system for backtesting? are you using a python lib such as QSTrader? I would love to have a similar article about your backtesting system. Thank you for everything!
Thanks!
I've used most of the open-source backtesters... after I realized I was spending way more time tweaking them than testing new ideas, I decided to write my own system. That was many years ago, while I was still in my master's degree.
I'm still figuring out what the best format would be for sharing it.
But I will do it soon...
Cheers!