The thing is, there are multiple markets not just one. So when you place an order to buy 100 shares of IBM, that order gets sent to many places to try to find the best offer to fill you.
If a HFT can see you submit that order to 1 market, and then beat you to the other 4 markets and buy up all the shares before you, that's frontrunning. That would be illegal if your broker did it, but HFT is not your broker, and just has the advantage of being faster than your broker.
What makes you say it's "not true"? What evidence do you have of that?
There are lots of things you are glossing over that complicate this picture quite a bit (like the fact that your broker is almost certainly executing on an HFT platform), but the biggest issue here is that no one can see your order until you place it on the first market.
Once that order hits the market it is public signalling data that should impact the prices on every other market. That some traders are more efficient at responding to that public signalling data than others is not an issue and in fact is a large part of why the markets are as efficient as they are.
As does purchased order flow, internal matching, preferred routing agreements, dark pools etc.
But given the simplistic nature of the discussion it is more accurate to say that in general HFT don't see your order before it hits the market than it is to say they do.
Okay, as long as people who actually want to understand the issue know that in many cases HFT firms can actually see orders before they become public although they don't actually do much to disadvantage a retail trader.
The thing is, there are multiple markets not just one. So when you place an order to buy 100 shares of IBM, that order gets sent to many places to try to find the best offer to fill you.
If a HFT can see you submit that order to 1 market, and then beat you to the other 4 markets and buy up all the shares before you, that's frontrunning. That would be illegal if your broker did it, but HFT is not your broker, and just has the advantage of being faster than your broker.
What makes you say it's "not true"? What evidence do you have of that?