Quantopian Algo Community Grows to 160K Members

Quantopian Algo Community Grows to 160K Members

 

They recently shut down live trading capability. Not sure of the end goal.

 

Wasn't there a headline six months ago where Steve Cohen's fund bought an algo from someone there for something like $2 million? I'm sure that has helped with signups.

 

160k and not 1 decent algo?

 

What's that all about?

 

That's a lot of free intellectual property Quantopian has access to.

 

Not really. 99.9999% of those algos are useless.

 

Not sure, but their infrastructure wasn’t very cutting edge IMO. They restrict sub-minute data. Big joke in the age of HFT.

 

In theory Quantopian has no access to your algorithms. I agree that shutting down the live trading was annoying, but I still think they've got the best business model and the best financial backing in what is mostly a very shitty product segment (compare and contrast with collective2).

GAT
(I don't work for Q but I should disclose that they've invited me to speak at a few of their conferences so arguably I'm potentially biased)

 

I'm a disillusioned Quantopian fan. They had a ton of promise, and as a fellow entrepreneur I want to root for their success. I participated quite a bit in their early days, even won one of the first 6 month performance contests. Although I wasn't looking for funding I can see that they lost a lot of people who were serious about it because their backers couldn't decide what it was they wanted. First it was absolute performance, then it was alpha, then it was minimize drawdowns, then it was do all that plus make an arbitrary X trades per day. At the same time, they had almost no finance expertise, which showed whenever you dealt with them regarding anything finance related. Spectacular software devs, but it sucks to be weak in finance when you're a trading startup.
They've kept pushing off and pushing off adding options, and they killed live trading which was the only way you could test any short strategy since there's no way to incorporate borrow fees into the algorithm. Given that their backers require both long and short trades before backing anyone, that's a pretty big problem. I pretty much burned through all the stuff I could test there, and now just use it once a month or so when I come up with an idea that their data set can test. Still holding out hope for them, but not nearly as excited as I once was.