HFTs will bury you if you're high volume scalper. your gross avg profit per share is taken by HFT. Media says few cents shouldn't matter, it does to a daytrader when netting out to volume traded (100,000 shares x 2 cents = $2000). The reality is you are basically adding 1-2 cents cost to all your trades on top of whatever your broker charges.
HFT Myths
- Thread starter hft
- Start date
HFTs will bury you if you're high volume scalper. your gross avg profit per share is taken by HFT. Media says few cents shouldn't matter, it does to a daytrader when netting out to volume traded (100,000 shares x 2 cents = $2000). The reality is you are basically adding 1-2 cents cost to all your trades on top of whatever your broker charges.
In my experience/opinion, no it's not a concern. Here's the scenario that CNBC is painting. The market is 10.05 x 10.07. You put in a market order to buy, expecting to get 10.07. HFT sees a print on the fastest exchange, buys all of the 10.07s and offers 10.08. You then get filled on one exchange at 10.07 and the rest at 10.08, while the HFT was able to buy 10.07 and sell 10.08 back to you.
This dialogue is largely a result of misunderstandings between people who have never traded a share in their lives. Here's what really happens that they did touch on in 60 minutes before the boogie man talk started.
The market is $10.05 1,000 shares x $10.07 1,000 shares. You decide to lift that 10.07 offer and buy all 1,000, so you bid 1,000 shares at 10.07. You get filled for 200 shares at 10.07, meanwhile the HFT on the offer sees a print on the fast exchange and cancels all of his 10.07 offers. You are now the best bid and the market is 10.07 800 shares x 10.08 1,000 shares. So, no the HFT did not rob you in this situation. He just pulled his quote to avoid being adversely selected. A lot of times if you just bid 100 shares at 10.07 they'll cancel the remaining 900 expecting you to hit the other exchanges.
In my eyes this is just strategy, not front running and not detrimental to the markets. Complaining about it is almost like complaining that a stock with good news popped before you could buy it.
the scenario you describe is VERY rare nowadays. in the past, it would only take place when a large size trade would unwittingly be sent to a pool or internalizer with a lot of "toxic" participants involved who would detect the size and trade ahead on the lit markets. the two factors needed for this to occur were SIZE (think large percent of daily volume or x% greater than y depth of the book), and IGNORANCE. since then, much of the buyside has smartened up about how to work their orders and avoid getting poached like this, but imo, you can't blame traders for sniffing this stuff out. that's how efficienct markets work, and the economic literature has a term for it: price discovery.
however, _nowadays_, the much more common occurrence is instead of liquidity being removed from the nbbo it's now quickly canceled. the way this works is if the market makers _detect_ that there's a good probability that LARGE orders are going to be coming through they're not going to sit there and take it on the head. you see, the majority of the time incoming orders are somewhat random, small, and happen on either side of the spread. in this scenario a market maker will want to capture as much market share as possible, and due to the fragmented structure, will tend to over-provide liquidity on all the different market centers to be competitive. however, let's say he gets the full amount of his order stuffed on one route... well he's just hit his exposure limit for that price and gained valuable information in the process (potential big flow coming). so, he will try and quickly cancel his orders on all the other market centers so as to avoid being left holding the bag with excess inventory and an informed trader coming in strongly against him. of course, size traders have gotten smart about this as well and written algo's like THOR that try and stuff them on every single route, and the result has been less liquidity posted at the inside market and a much more finicky nbbo (canceling quickly and often).
putting value judgements on either of the participants above though i don't think is very productive. they each have their function. they're both participants that are trying to trade with each other and trying to keep their risk/costs as low as possible.
so, to answer your question directly. if you're trying to remove ALL of the liquidity or some multiple greater than what's displayed on the NBBO, then yes, you may pay higher prices due to the reasons described above, IF you don't break up your order. if you're considerably smaller than what's displayed, you don't have to worry about a thing, because you're the ideal customer for a market maker's spread.
it's $2200/mo for cqs/cts and nqdf/ntdf, which cover quotes/trades for nyse/amex and nasdaq listing exchanges respectively.
You have got to be kidding me. There is a certain size showing on "the market." You put in an order to take that size. Because of the way HFTs operate, you only get maybe 10-25% of the size you wanted. And you think that is a functioning market? That's a complete sham (well actually it is a 75-90% sham). No one in their right mind would claim that is how a functioning market is supposed to operate.
Show 100,000 on the offer to lure out a big buyer. Buyer hits the offer, the order is flashed to the HFT'er > they pull the offer and enter a buy order for 10,000 shares ahead of 100K buy order.
The market is broken, and hft prop should be shut down but this sounds like less of a problem with the market and more of a problem with your smart order router (if you are doing DMA) or your broker's SOR or with the order itself if you tried to take all of the liquidity from an exchange that didn't have it. Most smart order routers will split your order up and send it to the lit venues, hopefully ordered by latency/fees/size/price depending on the algo or routing strategy they are using.
In my experience the worst SOR I've seen was 75% hit ratio and 98% for the best.
it's not a sham, what you're describing is a natural consequence of a fragmented market design and that's EXACTLY how you would expect it to function. if you're going to allow multiple market centers to provide liquidity simultaneously, then you have to expect that market makers can and will use information they get from data/executions on other market centers and act accordingly (and invest in speed to know what's going on everywhere).
you could definitely argue that that's a weakness in the design of competing market centers, increased complexity, but saying it's a "sham" is assigning a moral/criminal judgement on a design decision. a design decision, by the way, that was inherently decided on because it was pro-competition. as a result it had the very positive effect of reducing trading costs in us equity markets to next to nothing (exchanges competing with each other for business), AND allowing increased competition from non-member market makers/traders who could now go head on with member firms (specialists, etc.). not saying there aren't negatives, there are tradeoffs with every design, but it's important to understand the history and context before any harsh judgements are made.
+1 .... this is what the "I'll start making money when they ban the HFT's and stop losing a penny on every trade!" crowd needs to understand. If you are a small retail trader using market orders, HFT RARELY affects you -
- Almost all retail market orders are internalized anyway, and executed at the NBBO you see on the SIP anyway (or a little better)
- Even if you are not internalized, for small orders, there is probably a single exchange with enough liquidity to cover your order, and the "smart route" should take care of you (if any "smart routes" don't do that, that is something worth discussing)
If you are trading bigger size, know your routes and ping dark pools, or scaling in and out with 100 share orders works for me
If you are trading really big size or illiquid stocks, you may be screwed
If you are trying to place limit orders inside the spread, then understand you are playing the HFT game. You may realize some savings, but you will almost certainly never "beat HFT" and get an average advantage better than executing at midpoint (the issue being, you will get filled on the losing trades and miss the winning trades)
One more note I'd add that is very very important that most people leave out, intentionally or not. When HFT "sees a print", everyone subscribing to that pricefeed sees the print at exactly the same time, on the public (more or less, as long as you're subscribing to the relatively inexpensive feed) pricefeed. Depending on the market, you being the orderSender/liquidityTaker would actually receive this information before the rest of the world (in many markets, you get the fill first, followed shortly by the market maker, followed later by the rest of the world). In the "worst-case" scenario of order level feed markets, everyone in the world would get the print at the same time.
Not to create another fishmonger analogy, but it's like you have a ticket scalping operation at 2 street corners. You (and every one else on street corner 1) see someone buying all available tickets at the current asking price of $100. You send a runner down the block to raise your ticket price to $110. Anyone else could do the same, you just invested in a faster runner.
This is how the pits operated until about a decade ago. Except those guys were raking lots more than fractions of a penny on each trade, but no one complained because they were humans with paper instead of a computer.
And complaining about stock popping on news before you bought is directly analogous to firms using direct news wires for economic numbers. Still the same scenario: information is released to everyone at the same time (set by an atomic clock and automatic dissemination out of a closed room in DC), and those who react most efficiently and intelligently are rewarded accordingly.
Not to create another fishmonger analogy, but it's like you have a ticket scalping operation at 2 street corners. You (and every one else on street corner 1) see someone buying all available tickets at the current asking price of $100. You send a runner down the block to raise your ticket price to $110. Anyone else could do the same, you just invested in a faster runner.
This is how the pits operated until about a decade ago. Except those guys were raking lots more than fractions of a penny on each trade, but no one complained because they were humans with paper instead of a computer.
And complaining about stock popping on news before you bought is directly analogous to firms using direct news wires for economic numbers. Still the same scenario: information is released to everyone at the same time (set by an atomic clock and automatic dissemination out of a closed room in DC), and those who react most efficiently and intelligently are rewarded accordingly.