Ask HN: Is it feasible to do high-frequency trading as an individual?

I'll share some experience a friend had with this stuff. I don't have the low-level details, but this should be enough to make an informed decision.

He was working on what he termed "high speed automated trading" -- basically all in this realm of picking pennies up in front of bulldozers.

He's a very smart and talented guy -- Harvard Physics/Comp. Sci double major, etc. So we're not talking about an amateur throwing around some code.

He spent about 6 months and over $50k setting up his system -- he had developed some algorithms to trade spreads between different securites (on the CBOE, the NYSE, and NASDAQ). He built a relationship with a clearing agent, a direct broker, etc. He got machines in colo facilities as close to the exchanges as possible (one data center in Chicago, one in Jersey City, etc.) SLAs on low-latency DS3 lines, etc. The whole nine yards.

I'd call it a "pro-sumer" level setup -- everything done by one guy, but done basically as well as a bigger firm would set it all up.

Here's what he discovered after about a week of trading: He didn't have even remotely a shot at competing. Like, not even close. If he had 10ms pings to the exchanges, someone else had 5ms. If he got down to 2ms, someone else who was physically at the exchange itself had 1ms. He got killed with commissions -- even if he did make a few $$ on some trades, it vanished with commissions.

Why? Because he was competing against guys who paid NO commissions. The broker-dealers and clearing companies themselves had internal automated trading setups. They had deeper relationships and deals that traded commissions for a cut of the profits, etc.

In other words, he figured out pretty quickly that this is not an algorithm game or speed game - it, like many things on Wall Street - is a 'who you know' game. No matter how smart, how fast, how sophisticated you are as an outsider, the likelihood is that someone on the inside has a similar trade idea and can do it faster and cheaper than you.

Now, the corollary to that: My friend is smart, but he also tends to give up too easily ;) Clearly if you've got an idea for a better/smarter/more innovative trade, then you'll make money on it. But that's a trade idea, not a speed advantage.

So just recognize that there's potential to make money, but "same idea, just faster" won't cut it.

Seems like hosting a large hifi dataset in the cloud (eg rolling last 5 years) and charging a small fee to crawl it might be a good opportunity. Or maybe there's no margin in it when people are prepared to pay $$$ as you describe even if they're reinventing the wheel in the process.

Not that I'm a mathematical or economic genius of any kind, but I continue to be surprised at how primitive financial analytics seem. When people do find something interesting (eg Li's Gaussian Copula) they almost invariably make a fetish out of it and hurl themselves off the nearest cliff shortly afterwards. Economics faculties are as much to blame as anyone, I feel.

"In other words, he figured out pretty quickly that this is not an algorithm game or speed game - it, like many things on Wall Street - is a 'who you know' game. No matter how smart, how fast, how sophisticated you are as an outsider, the likelihood is that someone on the inside has a similar trade idea and can do it faster and cheaper than you."

For every wall st question posted to HN this is the answer. It's a lot of money in somebody else's backyard. The game's played in that backyard are always rigged against the outsider.

As already said both here and in previous threads: HFT is for big players; requires:

- hosting on exchanges premises to cut roundtrip delay (µS speaking). Also, very often, you'll be throttled depending on how much you pay. For entry fees, you'll be limited to ridiculous rate such as 20 msgs/sec.

- man years of development. Basically, when starting from scratch, we develop Direct Market Access gateways (i.e, custom/proprietary access to exchanges) in 3 to 5 man-months (using an in house framework). And they sit on top of other products that have dedicated teams working fulltime for years.

- upkeep: exchanges update their systems once/twice a year. Migrations take somewhere from a few hours to weeks, or a from scratch when they rewrite their complete API. If you can't keep up with the updates, you won't be able to trade.

And here, we're only speaking about trading. As highlighted, you also need market data, both retrieval and processing. Then and only then you'll be able to seriously start trading. (and then you'll want to do some back office stuff)

> The trading fee is close to $10/order

Deal breaker, especially for HFT.

Interesting thoughts. I had an acquaintance years ago who did a type of HFT. He built and ran a team of people for 5 years or so before 'retiring' with his FU money and traveling the world.

With respect to your news about a CEO, what he told me is that any stocks that had news or upcoming news were simply pulled off the table from trading. I always thought that there could be some value in parsing and understanding news, but from his standpoint it simply took too long. His other point is why deal with other variables when you don't have to. Find low news stocks and you find something that is more easily predictable.

That leads me to the next point he made, time. I asked him about using exotic analysis techniques and again he said most of their algos were variations of regression testing. Now, he didn't go into how they figured out the variables in their regressions, but he said the reason they stuck with those variations is because they were fast and they allowed them make decisions quickly before the window of profit closed.

Last I heard he retired at 30 traveled the world with his wife got bored and now runs a hedge fund. I don't know if how he was doing it 10 years ago would apply to today.