First, looking back 10 year from now, I think that Quantopian will be the single most important development in the world of finance. The barrier for me and a lot of strategy developers is the necessity of learning Python and having to re-create a fairly complex architecture needed to take full advantage of what Quantopian can actually do.
The typical reply to this argument is to hire someone who knows Python. But here is the problem - I am a strategy developer like you and not some big firm or office with deep pockets. I currently develop strategies on another platform - which is a combination of a GUI and some technical syntax - and I lease or sell these to family offices and funds. I don't have tens of thousands to pay out to a third party to 'just see if this works' better in Quantopian or not. I am of very low prominence in the investment world - yet I have made some highly profitable strategies over the past 5 years which a dozen or so funds, RIAs and family offices have used. Most of the strategy developers I know of are in a similar boat. We make enough to eat.
The second reason why hiring someone else doesn't work is privacy. If we really do have strategies that work well on other platforms - there is a risk that we lose that edge by floating the code around to third parties. If you want someone to code for you - you better pay them really well and have them sign all sorts of NDAs and have a serious trust factor. But it is much easier to say, "well this works good using end of day data so why risk the entire system by making it intra-day"?
Then why not learn Python? I started - but I am busy with other projects. To learn the necessary programming for what I have in mind (and what is needed) would probably take me a couple of years to learn in my spare time. I will likely take this route as will some other die hard strategy developers but that impedes the growth of Quantopian.
The answer, I believe, lies in having the Quantopian community come together and making a GUI interface in addition to the current Python-only setup. As well, there is a certain logic that needs to be built. The other institutional platforms for strategy development allows for the quick design of relative ranking systems of multiple factors. I think pipeline tries to address this a little.
Basically, the pieces are all there. The data and the programmers and the strategy developers. But a front-end needs to be developed for those who don't know Python. The concept of Pipeline needs to be taken a whole lot further. The question is - how do we get there? I am happy to donate my time and knowledge to help build something out. If there are others... maybe that is what is needed.
Comments from Quantopian users or management?