There is no technical reason. If you start with zero dollars, it becomes a painful mathematical exercise to figure out how to calculate some of the risk metrics.

On a practical level, there's no live trading options where you can start with that small of a sum. So I guess I'm curious why you're looking to start with a smaller number?

Sorry to resurrect, but what about Robinhood now? That's less than $1,000. If it's a pain to start at zero, perhaps it could just be a min of 1? Obviously, I know very little about how the site's code works, but it would be nice to set lower than $1,000.

I don't have $1,000 to throw down currently, but I'd like to start sooner than I would. Would live trading with Robinhood account funded with less than $1,000 cause any issues?

Thanks,
Dylan

So yes with robinhood you can now trade low amounts of capital without incurring to much transaction costs. Still $1000 is such a small capital base that it will be hard to get more complex strategies to be reliable. The main reason for this is price frictions. If you are specifying that a certain security should be allocated a certain amount of the portfolio, it may be hard to get that exact amount due to the price of the security. Consequently your portfolio may not be constructed in the way you modeled.

This is kind of the bummer thing for retail quants -- we often have great ideas but not the money to either implement it or make a sizeable return in real dollars. If you have a great idea it may be compatible with the Quantopian fund.

Hi James,

Quantopian/Robinhood supports <$1000 accounts, correct? I was thinking of setting up simple algo, with $500, just to get a feel for things.

Thanks,

Grant

Grant,

I tried it with <1K. Worked just fine. The strategy I deployed should have had a minimum of 25K because of PDT (...and T+3), but it was consistent enough in back testing that I decided it would still be fine making only a portion of the trades, while getting rejected on the rest. What was really liberating about the process was watching trades get rejected, and then seeing that the account had remained in tact. It built up my comfort level for the future.

Frank,

Were you using Robinhood Instant? Or their standard account? And how does the PDT work? I've been mystified by the whole thing. As I understand, the PDT flag returned by Robinhood is as a protection to prevent being officially flagged as a pattern day trader, per regulations (i.e. Robinhood would have to report users on a regular basis, if they didn't have the protection in place, or maybe they are required to have the protection hard-coded). So what happens when your algo gets the PDT flag?

Grant

So from 3/1 until mid April I was using just a standard account. The only rejection errors I received were due to lack of capital. Basically the lack of capital rejections were acting as a shield against violating the PDT rules. Coincidentally I received RobinHood Instant the same week that the big glitch happened and all the algos were down. When I redeployed the next week with Instant, I started to receive the rejection error that was specific to PDT. The system seemed to successfully safeguard me against violating the rules. The only issue was that the rejections were only coming on exit trades. Entry trades were still being executed...so the algo carried trades overnight, which it was not supposed to do. I shut it down after a few days, and have not yet tried to code a new solution yet. It lost ~$6, but was worth it in terms of getting a feel for how things will happen in real time.

Is Robinhood Instant open to everyone now? Or were you in a queue?

I guess blocking selling but allowing buying makes sense. According to https://support.robinhood.com/hc/en-us/articles/210216703-About-Robinhood-Instant:

We define a day trade as the buying and selling the same stock on the same trading day.

So, you can buy all you want--it is the selling that is problematic. So the PDT flag to the Quantopian algo is not a warning, but actually blocks selling?

I was in a queue. I was able to get a friend to sign up early in the process, which boosted me to the top of the list. Beware when inviting people to join, the link on the invite looks like spam, and caused my friends to think I was pitching snake oil. Also inviting people employed in finance is typically a bad idea because they are worried about the implications with their employer. Also remember that you would be asking any invitees to add an additional 1099 to their tax return.

Yes selling is the trigger if you had already purchased the same security that day.

@ Grant

Correct. There are no minimums with Q/Robinhood.