I've read studies that those who provide liquidity to the market constantly scoop excess returns from those who want instant buy and sells. Theoretically, couldn't someone then setup like 20 buy orders using a One Cancels All Order, on well-ranked stocks, and then bascially constantly provide liquidity to others? On the flip side - could have sell orders always active on all stocks. Do you think this would work? Then worst case scenario, if it doesn't sell, you still have a high ranked stock (assuming your model is good).
Has anyone tried something like this? Is there a easy way to do that? Interactive brokers? Is this possible in a TFSA? Thoughts?
Michael Lewis's Flash Boys is a fun read and talks about that some and huge investments in speed different mkt makers have in stepping in front of orders they want to be in front of. IIRC The IEX exchange was started to provide speedbump to how fast orders can be placed and cancelled. (I have no insight at tall on any of it, but mention as it's a fun read on the topic).
High-frequency trading (HFT) firms use fiber optic cables to connect their ultra-fast computers to stock exchanges just yards away, paying a premium for proximity because the speed of light over short distances is as fast as they can get. Wire cables are simply too slow for their needs.
Most of this is well documented in Spaceman’s reference and other sources.
"The Hummingbird Project" is a great film about two people who set out to build a microwave tower connection between the Kansas Electronic Exchange and the New York Stock Exchange to gain a one millisecond advantage in trading—the time it takes a hummingbird to flap its wings. While the film is fictional, it’s inspired by real-world practices in the HFT latency race.
AI and automated trading likely add layers of complexity that I don’t fully understand, but one thing is certain: no one can beat the speed of light.
Its a common HFT strategy to layer hundreds of orders deep into the book to be first in the queue, then deterministic (e.g. orderbook imbalance, tradepressue autocorrelation) and correlated predictors (e.g executed trades of liquid futures coming from the microwave link) are used to cancels these resting orders.
In a nutshell, you will be adversely selected on all your resting orders (fills on passive orders are highly toxic),
Obviously the above is overly simplified and I left "classic HFT" 10 years ago so this also might be outdated.
Hmm interesting. Yeah, makes sense that the big guys on the other ends of the trades always will have the biggest advantage and and fastest technology.
Haven't seen the The Hummingbird Project - Sounds like a great one! I'll check it out.
One might argue if they're small enough companies with low liquidity we won't be play against the big guys... though I am seeing that renaissance technologies does trade a lot of the same stuff I do also, which makes me wonder if i'm going up against the medallion fund....which is never a good thing.
I guess more my question is --- do you think I could get a better fill price by placing 10 different low-balled orders on 10 different stocks and taking the first that fills, then if I were to place 1 order on 1 stock? I don't mean trying to do 1000% turnover, but still keeping with my turnover target of 200%, and if i'm placing the trades anyways and ranks all with the first 5-30 are similar return, then I wonder if that would add up enough to make a difference in the long run if I get each company 1-3% cheaper each time I rebalance.
In other words, at each rebalance when selling and buying, place a bunch of buy orders across 5x more companies than you actually want and just take the first 1-2 that fill in attempts to get a more favorable fill price.
I guess we might still run into the same issue you guys are saying with having that many passive/limit orders while not watching - where we end up on the losing side of a price move and lose money because they attract trades only when the other party has better information.
I think what SFO is saying is that you will not outsmart the market makers. In other words, try to get all your orders filled at your assumed prices.
You're assumption about small caps used to be true - long time ago. Technology has changed it so that even the big guys trade the tiniest companies because it costs next to nothing to apply their signals to every stock, no matter how small.
Yeahh, i guess you're probably right. Market makers will skim more out of us than we will of them. Probably better to get orders filled quickly than leaving too many limit orders passively sitting available.
Yeah I have noticed renaissance and other institutional ownership on these small companies, crazy.