This article would get an F in a freshman college class... It assumes a "loss" for an order that is simply not executed... And then chalks that imaginary "loss" as an HFT "profit". Virtually everything written about HFT is complete BS... Especially anything shoved out there by Nanex... To "promote" their "state-of-the-art" TAPE-DRIVEN real-time data systems.
I would think its a little different. HFT sees buy orders coming in on a stock. How do they see this? Because they are front running. They pay to extra to see order before any one else, plus they have servers collocated to get the information faster than anyone else. So once they see these orders coming in, they get their second advantage which is the ability to jump ahead of any other trader or investor by quoting a sub penny better price order which is then filled. They then probably get out a penny ahead or maybe they hold the trade if their software sees more buy orders coming in who knows. Of course they do need to compete against other HFT software, and deal with the fact that many retail traders that may have been trading for cents are either stopping or going for longer trades which reduces volume. However, for all of their advantages their strength, and their speed, are still based in a world that is built on rules. Because of that, they will never be as strong, or as fast, as *you* can be. We can see this when an error caused the meltdown of Knight Capital.
Agree with Rationalize on this one. Subpenny issues are from broker internalization, not HFT. Also, if you ignored everything Nanex wrote on markets and just stuck to appreciating their pictures as art you'd probably be better off.