No, they don't want to spend the legal effort writing small fry contracts at this point in time.

You also have to consider that your algo will grow in value.

Making buttloads of money isn't a requirement for the contest. It's more important that it acts like someone's bank account with a high interest rate - constant gains even if slow, low beta, and low drawdown

Yeah, but with the amount of shares that can be traded without really unfilled orders, the profit is about 0.00001%. How are others doing it?

I suspect they would prefer you extend that algorithm to a basket of several hundred stocks.

Max -

You might try running it at the maximum capital it'll support (gross leverage 1.0), doing a pyfolio tear sheet (one can be run without revealing the securities), and sending it to them directly, just to get feedback from the horses mouth. Even if you can't enter the contest, you could run the algo using Quantopian's simulator. After 6 months, update the tear sheet with the out-of-sample data, and ping Quantopian again. If it is a bona fide money maker, why would they turn you down?

I have a long-short algo I've been working on that trades ~50 stocks (relatively large-cap). It still has a problem with some orders not being filled by the end of the day, when backtesting. Once Quantopian releases there tradeable universes, I guess I'll give those a try (see https://www.quantopian.com/posts/the-tradeable500us-is-almost-here). Based on my tinkering, I'd agree with Simon that you'll need 50-200 or so stocks at $10M+, otherwise it'll be hard to get it to work and not have the returns be choppy.

Thanks very much for your help - will do that!