High Frequency Trading Strategies

http://jonathankinlay.com/index.php/2015/05/high-frequency-trading-strategies/

by chollida1 • 11 years ago

122 65 11 years ago

jonathankinlay.com

rmxt• 11 years ago

Google Cache link: https://webcache.googleusercontent.com/search?q=cache:NePDgA...

yelnatz• 11 years ago

Mirror. Thanks.

functional_test2• 11 years ago

Take what Jonathan Kinlay says about HFT with a grain of salt. He worked at a friend's firm briefly before leaving without making a single profitable strategy (although his excel models were profitable of course). While what he says in this article is in the ballpark of correct, he is not the best source of information on modern HFT.

omgtehlion• 11 years ago

In this article he mentions only obvious concepts. Hence I consider the article to be ~99% correct )

functional_test2• 11 years ago

Exactly. It's so general that he's on base this time, but it's dubious if he actually has the experience to go into more detail.

dataker• 11 years ago

Any recommendations for a good blog on modern HFT?

exelius3• 11 years ago

No; HFT is effectively a zero sum game, so trading strategies are usually kept proprietary.

minimax• 11 years ago

This one is relatively new, but it's the most insightful (and candid) that I have come across: https://mechanicalmarkets.wordpress.com

bcheung• 11 years ago

http://www.quantnews.com/

Rainymood• 11 years ago

Full disclosure ... lol, thanks for the heads up.

guelo• 11 years ago

How about you divulge some tactics then?

pdeuchler• 11 years ago

Take what this commenter says about OP with a grain of salt. He had a friend that hired OP and seems to be bitter that OP left before he could profit off of him. While what he says may be in the ballpark of correct, an anon poster making ad hominem comments is not the best source of information on the OP.

lifeisstillgood• 11 years ago

It is nice to realise that people who clearly have stellar resumes still get this kind of knocking when sticking their heads above the parapet - it makes me feel worrying about the stuff people could say about me should not stop me from putting my head up more often.

lordnacho2• 11 years ago

So what's this strategy that makes money even if you only get filled 20% of the time?

torbjorn• 11 years ago

Just making a guess here but it may be that you have several limit orders fail to be filled for any given trade but when you you place an order that finally does get filled it may still complete a winning trade despite being your 5th order placed for the trade.

HockeyPlayer• 11 years ago

buying the bid without it going offered there. Then you sell the offer and have earned the spread.

harmegido2• 11 years ago

This would be an "aggressive" strategy. You get a signal from some source (looks like the author worked in news events, but the signal could be anything). If the signal indicates, via your model, that the product you are trading will go up through the best offer, you send an aggressive buy into that offer.

If all works out well, you get filled for a buy and the old offer (which you bought at) is now at least the best bid. Do this hundreds of times a day and you're starting to talk about real money, especially since you can scale (to a point) by upping your quantity.

Sometimes you aren't fast enough though. That buy you sent into the best offer is too slow - it merely appears as a buy at what is now the best bid (or doesn't at all if you use an IOC - immediate or cancel), and you get no trade. Even if you only get filled on 20% of the signals, if they are on average good, then you still make money.

This is just one type of strategy though. Not all HFT strategies rely on aggressive signals, though a lot do.

noname1232• 11 years ago

Quick question, I don't work in the HFT field (I trade only options for fun on the side) but would anyone in the field care to comment whether a startup fund say, with a back-tested strategy would still be able to compete with the likes of HFT firms, provided that they have (1) sufficient capital, >$500K, (2) sponsored access/co-location to markets by proxy, e.g., Lightspeed, Lime Brokerage that provides the lower fee-structure plus the rented rack-space on the exchange data-centers, (3) historical and realtime quote feeds for forward and back-testing.

I imagine HFT entities who are broker-dealer agencies have even lowered fee structure; but it sounds like simple market-making on equities is no longer an easy business to be in due to the lowered spreads and crowded competitors.

So I'm wondering if a startup funds with the rented access could compete with the big boys in the options, futures, other derivatives and even bonds space where the latency still haven't caught with the equity market yet.

fasteo• 11 years ago

This is something that comes to my mind every time there is a press release talking about HFT and the milliseconds advantage they use to make a profit.

I would love some comments from a seasoned HFT (tech) guy.

kasey_junk5• 11 years ago

There exist in the trading space entities that look essentially like what the OP is suggesting. They are called prop trading firms, or quantitative trading firms, etc.

There are even entities that act like incubators for those entities.

lordnacho2• 11 years ago

There's a fairly big question of what human capital comes along with this. You will need some guys who are good at low latency programming.

Some strategies that I've seen also depend on having good agreements with brokers, ie you may see an opportunity in the data, but the costs of trading tip you into the red if your guy is margining you in a different way. I don't know what Lime et al charge you, but they aren't a likely name among the people I know.

There's also a question about what exactly you'll be doing in those markets. There are guys who make money market making options over the phone, without relying on colo for anything. Sometimes they'll have a quote machine that needs to be quick to pull, but they don't see it as anything other than a bit of extra money. Other options guys are totally different, relying entirely on speed to make their dough.

Also keep in mind you are only competing with HFT in as far as what you're doing is similar, ie there's a high chance you'll both want to grab the same trade.

fixxer5• 11 years ago

HFT represents a spectrum of strategies when described in the news. For the most part, I would argue you are describing a "quant trading shop" and would avoid the HFT label since it doesn't apply (HFT really means that you're gaming Reg NMS).

The type of shop you're describing might be able to compete, assuming you've identified a niche product with a margin that will still exist in 6 months. Backtesting is useful, but it is so bloody hard to accurately incorporate liquidity that I place little value on simulated results when it comes to short duration trades. You really have to jump in and test live, which represents the largest barrier to entry: established shops have the infrastructure & capital to try new ideas. When I was trading, we tested half a dozen new ideas every month in addition to "production trades".

High frequency, IMO, should really be used to identify hardware intensive strategies co-located at the exchanges and utilizing telecommunications strategies not readily available to the market. For example, FPGA and dedicated microwave towers.

Lime is excellent, btw, but you really need some volume to justify them. For options at your level of capital, Interactive Brokers and long gamma trades look more attractive to me. Many would disagree and people do starve waiting for long gamma, but my memory is biased after seeing the little group next to me lose $40m in a week on a stupid short gamma "sure thing" based on back tests and other such nonsense. Oh, what a fun time that was.

tptacek3• 11 years ago

When you say "HFT really means that you're gaming Reg NMS", are you saying that true HFT strategies all rely on multi-venue arbitrage?

noname1232• 11 years ago

Hello fixxer, thank you for your reply.

My understanding is the reason you advocate for long gamma vs. short gamma because unhedged short gamma trade (e.g., short straddle) could cause theoretical unlimited loss. Long gamma trades like long straddle or long option hedged with delta-neutral stock, you are facing daily theta decay and you are betting for either a rise in implied vol or realized vol.

However, the missing ingredient/secret sauce is (1) you have to figure out at price point you're going to take profit in your scalp; and (2) when implied volatility is considered cheap for the particular ticker you're trading. Much appreciated for your insight and let me know if I'm missing anything.

thibautx2• 11 years ago

I would say that even options/futures/and other products markets are just as hard to market make compared to the equities markets. In my opinion, you need some sort of edge not through hardware but through your model or your market outlook, combined with market-making/hft/statistical-arbitrage in order to be profitable and compete.

1) Capital in the magnitude of >$500k seems reasonable for a small operation

2) Lime Brokerage (a friend has started a hft strategy with a professor on campus and has tested out lime) is okay, I'd look into some other execution services as well

3) Yes you should be backtesting on tick-level data from multiple exchanges ideally. Be wary of lookahead bias and try to build some sort of execution model to account for how your executions might actually be filled and how your orders might affect the market microstructure

Ultimately, I wouldn't rely on pure hft and latency sensitive strategies...latency is always important but you should try to find your edge elsewhere.

chollida14• 11 years ago

As always, if anyone has any questions, please feel free to reach out, my email is in my profile.

I'm here to teach and I love consuming coffee:)

stevewepay• 11 years ago

One question I have is, when do you know to abandon a trading strategy that used to be profitable, but has not been profitable over the past N number of days. At what point do you throw in the towel and move onto another strategy?

harmegido2• 11 years ago

You have to find out why it stopped working. Was it an aggressive strategy that used to have a 100% fill rate, but now is only getting filled on 30% of the orders it sends out? Chances are that you've just become too slow for that trade. The only fills you are getting now are the bad ones. Time to move on or speed up.

Let's say the fill % stays the same though. Has volatility in your market dropped off a cliff the past few weeks? Perhaps your strategy is highly dependent on volatility. It'd be good to keep this one running, as it's likely volatility will return at some point. You want to diversify though and have other strategies that perform well in such a low vol regime, so you're not dependent on high vol.

ksk• 11 years ago

Disregarding the "cool" technical challenges, what are some of the positive contributions to society that only HFTs could bring about?

kasey_junk5• 11 years ago

"that only HFTs could bring about"

That is a very high bar to get over for any industry. For instance you could say the exact same thing about the internet generally and not come up with too many answers.

Another way to reason about it is A) how does HFT compare to the system it replaced? and B) what would we need to give up to not have HFT (if we decided for some reason we didn't want HFT, though I'd love to hear someone enumerate why we wouldn't want it).

A is trivial to answer. It is much, much cheaper for nearly everyone to trade with HFT than it was in the world of specialist market makers or pit traders. In the modern world there are much less middle men for every trade and they take a dramatically lower percentage of the value than the system they replaced. This is the obvious and expected outcome of replacing expensive humans with cheap computers.

B is also pretty trivial to answer. We would have to go to a monopoly exchange with no electronic communications into or out of it. I hope the downsides of that are obvious.

anilgulecha2• 11 years ago

I'll take this on.

HFT replaces "expensive humans" with specialist computers that try to site right next to the trading servers, so they can take advantage of the speed of light.

So may I suggest a simpler alternative: Run a trading system with three chunks of time: "get orders", "serve orders", "rest". These intervals could he an hour or 15 seconds. But these way a phone on the otherside of the planet could put in a trade along with the speed of light computers, and not be at a disadvantage.

All HFT is doing is licking the cream off the top of the market -- and not providing any value that a more fair computerized trading system would provide.

esaym• 11 years ago

I tried this with bitcoins and that didn't work out so well. All your math formulas can be pointing up, but you never know what the human element will do...

curiously• 11 years ago

Isn't HFT pretty much limited to those with deep pockets?

kasey_junk5• 11 years ago

Depends on what your definition of HFT is. But for every definition of it, deep pockets are relative & HFT is not any more capital intensive than lots of other industries & much less than most.

arbitrage314• 11 years ago

From someone who works in/near the industry, I can tell you without a doubt that the best HFT strategies are the ones you never hear about.

known• 11 years ago

Insider trading + Front running = HFT

RomanPushkin• 11 years ago

The article is so unprofessional. I can give you unlimited number of strategies based on historic data. And they won't work in the real world.

High Frequency Trading Strategies

Related Stories