Lean Engine
th988163
Posts: 2
First post here, hope it doesn't come off the wrong way...
My biggest issues so far with using this site are having to run a back test to debug(and the vagueness of the debug results) and the quality of documentation. To be clear, this site is far better than anything I could put together, and i'm sure there are countless hours doing so, so please understand that the criticism is constructive.
That being said, why not use the lean engine for the site? It's open source, it is intuitive, it can use C or python, it has a good debugging module, its documentation is thorough, it is downloadable onto my computer so that I can do the coding and debugging on my own machine before bringing it to the site (which would likely save a ton of bandwidth and CPU resources for doing what the site does best... provide data), all of the little scripts here that I seem to have to write to reproduce simple things are already done, it would free up the dev team here to support their traders rather than build/maintain a framework, etc.. I've been toying around on another site that uses the lean engine and the difference in ease-of-use and learning curve are night and day. I want to keep going on this site, mostly because Paul's voice is so soothing on his youtube videos compared to the young yuppies on the other site (and the relationship KTG has with SMB), but the 9 hours to learn python thing doesn't include the 9 days in seems to take to sort through your code and figure out what you're doing wrong with the API.
Comments
Appreciate the feedback, particularly about my voice
All of your frustrations are frustrations I have expressed myself. Unlike many other systems out there, this system is in live use at trading firms. This limits even more our ability to make the kind of significant changes you suggest. (Lean is also a competitors product so.... ).
Let me know if there is anything in particular tripping you up or which was confusing as you learned how to use the service. I am always available and want to make the user's learning curve as smooth as possible.
I'll post specifics as I get them. I think I am expecting too much from the systems. I am coming from easylanguage and thinkscript for trading, both are pretty easy to pick up and work well for certain things like making indicators. I have previously coded in C#, C++, and VB.Net. Each system I use ends up having limitations, and every time I try something new i'm hoping to find the one language to rule them all... and keep getting disappointed.
Limitations up front, our backtester was developed in house for a intraday prop house so it is focused more towards the "event" intraday type trading rather than Quant/Fund Multiday hold. You can certainly create multiday portfolio trading algos, it is just more complicated and you have to think more about how you are going achieve it, in order to make it run as fast as possible.
But you cannot really lose, at minimum you will come away from the experience a reasonable knowledge of how to program in Python and that will undoubtedly benefit you for the rest of your life.