Is that another way of saying that retail traders buy on the offer and sell on the bid while HFTs buy on the bid and sell on the offer?
This is market making which is the bread and butter of HFTs. This is the one thing they do which helps, not hurts, human traders.
They can do thousands of trades in the time it takes your nervous system to react and click the mouse. They parse a news headline many seconds before it ever shows up on the internet. They rent satellites to count how many trucks leave factories. They have access to person-to-person dark pool trades that never show up on the normal exchanges.
Algorithmic trading is a secretive black box, but who knows. These are just my speculations from hearsay and random research. If they can make a fraction of a cent from messing with you, they will. It can scale up infinitely. It's just software--there's no marginal cost to do so, provided they are properly hedged.
> This is the one thing they do which helps, not hurts, human traders.
Providing liquidity is their job, and it benefits the market as a whole. If market making by HFTs is a good thing, it’s difficult to malign those same firms for, well, making markets.
> They rent satellites to count how many trucks leave factories.
This isn’t their business model in the slightest. Quant funds certainly do this, but HFTs look for alpha in market microstructure.
> They have access to person-to-person dark pool trades that never show up on the normal exchanges.
Institutional investors can use dark pools to minimize market impact when trade large blocks. HFTs don’t benefit from having limited information on these flows.
> Algorithmic trading is a secretive black box
Algorithmic trading != HFT in the same manner that a rectangle is not a square.
> It can scale up infinitely.
HFT strategies don’t scale. ‘Scalable’ strategies support a large amount of capital. HFTs run high sharpe, low capacity strategies.
Either way, those hft firms definitely like retail order flow data. My guess is they're easier to "pick nickels in front of steamroller" kinda trades than institutional money which may cause extended one way moves that hits high frequency balanced traders adversely.
I'm also entirely talking out of my ass in the last paragraph. I don't actually know that any of what I said is true. Just speculation.
The order flow data is worthless (because the orders come from uninformed traders) the value is in filling them without the risk of being shortchanged (because the orders come from uninformed traders).
2) even if that were not true, hft works on volumes. So I contradict my previous speculative comment by saying this but volumes of institutional flows absolutely dwarves retail flows. You'd have much more opportunities trading around institutional money than retail money.
So if you've got a company and the city in which it's headquartered just gets a strong buy signal, sure that could be random but I'd imagine ...
Surely having the meta-data on who is trading and linking that to trades is the primary benefit?