> Top equities on US exchanges like NASDAQ and NYSE might be out of the scope for the average hobby quant, but a thinly traded penny stock, or a low-cap cryptocurrency on a smaller crypto exchange, might not be.
These kinds of markets tend to exhibit the kind of price moves and thin liquidity which make pulling off 0+ style trading very difficult.
Not really sure how this is supposed to be low risk -- I think they're omitting the actual interesting parts of the strategy (that are essential to actually making money).
If you want to be the first or close to the first in the queue on a price level, you're there before it becomes big. So then your "risk-free" closing strategy doesn't work, because if you get traded against, you will have to close out your trade at a loss, since the price level is gone (because it wasn't big yet).
But if you wait until the level becomes big, you will no longer be at the front of the queue, so the strategy doesn't work.
If the ask is at $8.03 and you are first in queue at $8.01 all the orders in $8.03 and $8.02 have to be filled before you get filled so you have plenty of time to cancel your order.
Assuming you mean bid. If the the size is only 10 @ 8.03 and 1 @ 8.02 and 8.01, then if someone is looking to sell 100 at market price it will be a problem. The sell order will go through many levels after filling yours and it will all happen at the same time.
It seems like the information being disseminated by the article is just the mapping between the name and the concept, since every participant already knows that they want to be at the front of big levels and can take from the orders behind them if they don't like the fill.
Yeah I was hired by a prop trading firm of old Chicago floor traders in 2009 and then shortly asked to come up with an algo strategy in Kospi 200 options, with no help from any experienced algo traders. I worked something out completely on my own which was very similar to 0+, it just follows naturally once you understand the dynamics of a microstructure with low fees and FIFO queues.
They really gloss over "getting to the front of a strong queue," which generally requires stacking to many levels of depth in the market (which adds risk), or finding some way to join or remain in weak queues before they become strong, which requires some other method of estimating a theoretical value. Estimating and tracking depth in the queue at each level was definitely a significant part of the more sophisticated code I wrote.
The article may overstate the importance of lower latency as well. I definitely was not the fastest at the time, running some fairly non-optimized Java when I knew there were some faster C++ players in the market, and I was still able to do relatively well.
Can't you start off by spreading your OCO limit orders 10-11 ticks away from the last price and then just wait it out? Those queues are usually very thin.
Right, so that's "stacking to many levels of depth." To add some color to my point about how it "adds risk," the day Kim Jong-Il died, the market spiked lower and I was instantly lifted on every stacked offer I had in every put. I managed to recover a bit but it still took out about a quarter of a year's PNL in one fell swoop.
Nobody is going to start the next rentec after reading this article, but for people new to the subject I think it is quite an approachable intro to some of the practicalities of market microstructure and how they affect trading strategies.
It's helpful, but even after all the work to be on all the right levels one still has a difficult prediction problem to decide when they don't like a fill. I guess the point is that this prediction problem (conditional on fills at front of big queues, should I dump it?) may be easier to make into a profit because your predictions can be of only moderate quality if your downside is just fees. In crypto d1, the fees are maybe small enough to support this, but other concerns make it difficult. As mentioned elsewhere in the comment section, one may learn that their downside is often not just fees.
Yeah, totally, it really just gives a flavour of how the structure of price levels and price/time priority, tick sizes, etc. can create exploitable probabilistic advantages, and I think many folk are not aware of even this.
The real environment is absolutely messier and more complex than described by this article. Still, it's a layer deeper than "speed is good", to paraphrase Gordon Gekko.
Yale finance could argue that the https://en.wikipedia.org/wiki/Perpetual_bond has priority with the oldest example being 1624 as the holder has the option to collect coupon payment. Currently, Yale has the opportunity to write up and sell the option on the coupons (back to the water board if they wanted to make it a little less unfeasible of being "perpertual") and turn the original "bond" into a zero coupon "bond"? (more of an annuity given the principal can never be collected - unless doubtfully they were written as callable).
Far more money has probably been stolen with the "heads I win, tails you lose" approach -- just put on massive leverage and stick someone else with the bill if it blows up -- than has been stolen by all the Ponzi schemes in history.
Slightly on topic, but could Elon turn starlink into a giant arbitrage opportunity? Once satellite - satellite laser links are up, it could be the quickest link between say NYC and Tokyo
The speed of light in vacuum is faster than in fiber optic cable, although presumably there are already microwave signaling networks like someone else mentioned
Yeah I'm pretty sure cables will always beat satellites over the same distance. One interesting idea I heard was to string stationary drones that would relay microwaves across the ocean, similar to what was done with the relay towers between Chicago and New York.
Light travels at ~0.8C through fiber optics, whereas it travels at very close to C through the atmosphere and C in space.
You need to make up for the extra ~200 miles going up and down, as well as the extra curvature of the satellites orbital shell vs the ground. But, it is doable, assuming you can relay the signal fast enough from satellite to satellite. Fiber optics needs to also be boosted on long routes, I'm not sure how that affects latency. But at the very least, speed of light up/over/down is faster than .8C for 7000 miles around the earth
Getting your fiber to go straight (in the great circle sense) from point A to point B can be quite challenging given the presence of mountains, water, and people or countries that control the land in the way.
For what it’s worth, hollow core fiber is available now, although at least the one vendor I found is very very expensive. I haven’t found an entirely clear answer as to the group velocity in hollow core fiber (it’s not in the data sheets), but it ought to be pretty fast. (The group velocity in vacuum-filled waveguide is not the speed of light in a vacuum — it’s slower.)
Right now, the fastest path from Chicago to London is to bounce radio signals off the ionosphere, which lands them in Lithuania, and then microwave back to London.
HFT may be the most obvious advantage of technology disrupting traditional oligarchical structures the world has ever seen.
A market that was previously dominated by good old boy networks of football players was made massively more efficient and the cost of entry driven to the floor purely based on nerds with math.
To the benefit of effectively every market participant besides said football players.
I find it interesting that an industry that has spent the last 30 years dedicating their best and brightest to panopticon building ad driven navel gazing can continue to act high and mighty about real innovation.
Otoh, that ad driven navel gazing ultimately delivered Generative AI which is likely going to lead to superhuman AI. What equivalent value did HFT deliver?
On the contrary, equities markets need the liquidity HFTs provide to function well. Before HFTs there were human market makers who played a similar role in equities markets, but charged far more to do it. HFT is a story of efficient technology bringing prices down a lot and shrinking the profitability of an existing industry.
I’m sympathetic to the view that a lot of resources have been poured into making trades happen in 1 nanosecond when 100 or 1000 nanoseconds would probably be fine for all practical purposes. The issue is that people keep finding that letting the fastest system get the trade actually does work better than alternative approaches.
It means people get better prices than they otherwise would.
US markets are pretty bad along many axes (e.g. brokers have an obligation to route flow to wholesalers in an effort to maximize price improvement, but this does not necessarily result in the best price on a per-order basis the way that a limit order book populated by wholesalers and other retail orders would. startup costs to get "fair access" to institutional flow and 0 captive retail flow are multiple millions of dollars because of exchange revenue models. retail maker orders will literally never interact with retail taker orders. etc.), but it's difficult to argue that retail would pay less in fees and spreads if they had to always interact with quotes manually created by humans.
I definitely don't know enough about trading. So please forgive if this is just naive.
From what I gather, the problem is matching sellers to buyers. Sellers list what they're looking to buy, and the price they're willing to pay. Buyers list what they're willing to sell, and at what price. Let's say each offer is valid for 100 ms (that's a reasonably short time span, no immense risk for the entity making the offer).
But I find it difficult to understand where speed (on the level of HFT) ever becomes a factor in this game?
If the HFTs looks at different exchanges and work as gel between them, making the trades between exchanges that otherwise wouldn't happen, that would make sense as a valuable business.
Is that their purpose?
Although, it still seems to me the slower buyer and seller would prefer to come to an agreement alone without the middle man taking part of the profit.
If the trade enabled by the HFT would've happened the next 100ms clock cycle anyway after the buyer and seller noticed each other and adjusted their prices, I don't see the value of the HFT trader. Seems they're just hacking the system.
Most HFTs are market makers, meaning they are willing to buy or sell any of the stuff they trade at any time. They provide the service of assuming risk that they don't really want. If some non-HFT participants show up to a limit order book, they could transact with each other, but it's more likely they will each transact with a market maker. Because of the way limit order books work, they cannot possibly receive worse prices due to the presence of the market maker's orders. If the best price each of them could receive was from the other's order, they would match with each other.
US markets don't work precisely this way for various reasons, but it seems like if the question is "What service, if any, are HFTs selling?" then the answer should be "They are willing to assume risk by taking the opposite side of your trade basically regardless or what your trade is" rather than a rant about US market structure.
"Trading in one market based on price information from another, such that the prices in the two places are in agreement with each other" is actually called out elsewhere in this comments section as a value-destroying activity that reduces liquidity and therefor at the margin causes less sophisticated traders to get worse prices.
If someone is regularly noticing things and adjusting their price in a tenth of a second, it seems likely they are engaged in some form of automated trading.
> Because of the way limit order books work, they cannot possibly receive worse prices due to the presence of the market maker's orders
The (HFT) profit has to come from somewhere. And I'm pretty sure they come from the other market participants. If that's the case, clearly some market participants are getting a worse deal than they'd otherwise have gotten.
At the time scale of seconds, trading is a zero sum game.
Perhaps there's something I'm missing?
> is actually called out elsewhere in this comments section as a value-destroying activity that reduces liquidity and therefor at the margin causes less sophisticated traders to get worse prices.
Yes that's how I also understand it. Is this description incorrect?
> If someone is regularly noticing things and adjusting their price in a tenth of a second, it seems likely they are engaged in some form of automated trading.
I'd consider automated trading and HFT to be orthogonal concepts. I totally see the benefit of automating trading, but I don't see the benefit of HFT. The 100ms number was selected small enough to be faster than any "real" "tangible" "physical" process in the economy, but still way slower than HFTs.
> The (HFT) profit has to come from somewhere. And I'm pretty sure they come from the other market participants. If that's the case, clearly some market participants are getting a worse deal than they'd otherwise have gotten. At the time scale of seconds, trading is a zero sum game. Perhaps there's something I'm missing?
Can you work through a specific example for me? Previously, before automated trading, specialists would make markets in specific stocks, and people who traded with specialists would pay larger spreads than they do now. I've claimed that "being willing to take on risk" is a valuable service worth paying for, and that most people in most markets pay very little for it now, as a result of their counterparties selling the service for very cheap. It is of course possible that some subset of participants could work out deals among themselves and end up with a surplus as a result, in the same sense that it's possible that you could stop going to stores and buy everything directly from suppliers by driving to their warehouses, or start your own stores or whatever.
> > [capturing surplus due to short-term inter-venue mispricings] is actually called out elsewhere in this comments section as a value-destroying activity etc.
> Yes that's how I also understand it. Is this description incorrect?
It's not incorrect, but in your other comment you said
> If the HFTs looks at different exchanges and work as gel between them, making the trades between exchanges that otherwise wouldn't happen, that would make sense as a valuable business. Is that their purpose?
Which precisely describes the activity that you now agree is value-destroying.
> I'd consider automated trading and HFT to be orthogonal concepts. I totally see the benefit of automating trading, but I don't see the benefit of HFT. The 100ms number was selected small enough to be faster than any "real" "tangible" "physical" process in the economy, but still way slower than HFTs.
I don't think this is reasonable. If you are engaged in automated trading and you occasionally place resting orders on a limit order book, you need to be able to cancel those orders very quickly some of the time. The only alternative is that you believe that capturing surplus due to short-term inter-venue mispricings is value-destroying and that *failing to let other people destroy value by interacting with your resting orders when you know that you would prefer that they do not* is also value-destroying. That would be a surprising set of beliefs.
As an aside, there's a proposal quoted elsewhere in this comments section for FBAs[0]. I don't have a sophisticated opinion on these, but my first impression is that if FBAs replaced much of the existing market structure then the need for anyone with resting orders to cancel their orders on very short notice could become dramatically less frequent.
HFT at the extremes doesn't just create no economic value, it actually extracts value from all other participants.
Frequent batch auctions are one possible solution to this problem[0] that I see as quite promising. The problem of couse is that exchanges profit from the HFT extraction via fees and the prestige generated by reporting higher trading volume.
Any real solution needs to solve both the micro-structure and the macro-structure of markets and make exchanges a publicly or collectively owned (or at least incentivised) utility.
It does the literal opposite thing: it provides value to every market participant (except the inefficient human market makers and commercial brokers it replaced).
What kind of things are you referring to? People writing poetry? Or someone helping their elderly neighbor shopping? Or misleading advertising?
Some things that don't create economic value are extremely good and valuable.
Bit it's obvious that spending resources to do nothing useful is objectively bad, even if it happens to redistribute wealth to you.
HFT is frustrating since it creates dubious value and eats up a lot of technical talent in the process
I don't think there's a reasonable way to stamp it out, though.. I guess if IEX's anti-HFT strategies were more ubiquitous it'd transform the HFT industry and maybe calm it down a little?
Thinly-disguised clickbait/advertising.
Reminds me of being pamphleted in the airport in the early 1990s.