Podcast: Play in new window | Download
Subscribe: iTunes | Android | RSS
Jason speaks with James Thomas, Director of Research for Headlands Technologies, LLC, about the business of high-frequency trading.
Subscribe to TechZIng
techzing tech podcast
If you're a hacker, you'll probably like our show ;)
I was sceptic at first, but I like the “one interviewer” format a lot. There’s much less fighting for attention this way.
Good show and thanks for explaining the vocabulary to those without any trading experience.
@Stanislaw Pitucha – Thanks, I’m glad you liked it. 😉
I can imagine that the “one interviewer” format may take some getting used to for some of our listeners since not everyone likes change, but Justin and I both feel that this approach will work better overall. We’ll try to talk a little more about it on this weekend’s discussion show.
Very cool show. Could have been a little more technical, but still great.
Maybe you could reach out to some other HFT shops?
BTW, Justin, I just noticed the new Pluggio website (a little late, I know). Looks really professional now, congrats!
Perhaps you could lend some of that sense to the TZ website? Could help with convincing potential HFT interviewees, you know… 😉
@Axure – Thanks, I’m glad you liked it.
I would have really liked to have gotten more technical, but James made it pretty clear in our pre-interview discussion that he wouldn’t be able to talk about what they did specifically, which kind of ruled out most of the technical questions.
Interesting interview. I didn’t complete a Ph.D but close. Anyhow, I guess my interest in search engine technology somewhat overlaps with data mining a bit. This interview got me thinking, maybe I should apply data mining algorithms to figure out a way for machines to assist humans to do better SEO 😉
Talking about machine learning, it border on AI I guess. There are just too many variables in this and it is hard to make work unless the computer is powerful enough to sort out all possibilities, IMHO.
I have a thought on how to fix the overfitting issue. Maybe just using some common sense, would correct the over-reliance on algorithms and machines. It seems to me, a lot of academic type of folks would over-analysis certain issues where a simple and straightforward solution is just as good enough…
An interesting interview on an interesting subject.
You touched upon some of the criticisms of high frequency trading, but I thought I would bring up one critique of it that wasn’t mentioned. In high frequency trading there is a lot of money to be gained in being the fastest to make and act on knowledge, and doing so certainly increases liquidity and improves market efficiency. However, improving efficiency and liquidity may have diminishing returns to society as the time horizons they are working on shrink below human perception. It isn’t clear that having a stock reach it’s ‘true’ value a millisecond faster than it would have otherwise brings much benefit to society even if it makes a firm a lot of money.
The concern then, is that the money to be made from certain types of high frequency algorithmic trading may be dislocated from the good it provides to society. Incredibly talented people and vast amounts of computing resources are drained from the productive economy to work on a problem that may not provide actual value to the wider economy.
Hard to prove one way or the other — but interesting grounds for debate.