Bootstrapping an Ultra Low Latency Trading Firm, Part 1

From the article: "And if you are careful enough to save money (for example living at home saves about 3K a month when all is said and done) and not increase your expenses too much (no need to get a Porsche when a BMW or VW suffices), it’s not hard to save 200K by the time you turn 24."

Uhh, sure? 200K at 24, in savings?

So that's about 250K pre-tax. Say you graduate high school at 18, find a magical job that gives a high school grad $42,000 a year, and spend /not one cent/ for six years, maybe... (or extrapolate outwards).

Or maybe you go to college. Graduate, let's say, age 21... In this economy, how many college graduates walk out of school into an $85,000/year job and no living expenses...

Never underestimate the value of a silver spoon in your mouth... and never underestimate the power of denial about just how silver your spoon is.

Many of the items discussed in this pre-article seem way overkill for what the author seems to be trying to do. It doesn't take much to create an account with a reputable broker like Interactive Brokers that has a good live API for auto-trading (and support paper trading to test your code without using real cash). There are no upfront fees on many brokers and the commission is often minimal depending on the transaction. eTrade has similar APIs. Historical data is available at minimal cost through a sites like eoddata.com. Also, especially if you're just starting and experimenting with formulas, using your personal laptop/desktop with a live feed from IB is more than enough to act on live tick data for the symbols you're interested in.

As an ML/AI guy, I'm curious about the problems encountered by HFT systems. What I would love to see is some data, along with the desired outcome. Maybe also include the amount of time available to make the said decision. One could go further and institute a "Netflix prize" with such data.

Right now, I'm clueless about what kind of data these systems deal with, and what are they supposed to do.

Any HFT/ULLT people willing to share data?

I'd like to go on the record as the pessimist in the room. If you go down this path, I strongly believe that you're going to lose your shirt building a bankruptcy machine. In many ways, the markets are a mechanism for transferring your money to the pockets of people who started out with more money than you.

Here's another question I'm curious about. Given that you've decided to start an automated trading firm, how much better off are you going for microsecond timescales rather than, say, seconds or minutes?