1) Place a trade. 2) Trade goes to internal dark pool. Maybe someone who paid for order flow fills it, maybe not. 3) In the absence of a fill on the internal dark pool, the trade is routed to the markets at large.

The only difference is who your anonymized counterparty is.

I don't think that is true, conceptually. Lets say that by going to the dark pool and being processed through to the exchange you lose priority on your order. Then your fill rate could be worse.

Not that I think this is a reason to chose to use robinhood or not.

1) Place an order. 2) Broker checks the darkpool - if a match is found, trade internally. 3) Route to the public market.

Step 2 can add latency, reducing your fill rate. So can playing music on youtube while doing step 1 probably adds more latency. And the latency in both cases is vastly lower than the latency you get from trading over 3g (robinhood is a mobile app, which to me seems moronic).

Another way you could see worse fill rates would be a customer places a limit order that puts them somewhere back in the book. This limit order happens to trigger some threshold for the order flow trading algorithm that causes it to back off of it's own orders (lets say it cancels some orders on the back of the other side of the book). This in turn causes a price shift that moves away from the customers limit order making it fill later.

Clearly we are in the realm of pure speculation and probably not something that is important to the customers of robinhood. I was just pointing out that conceptually, by selling their order flow, robinhood could impact fill rates.

My guess is that this is more than offset by the lack of fees.