Comments on: basic economics of an algorithmic trading startup

By: @mccartjt

@mccartjt — Fri, 23 Dec 2011 01:46:47 +0000

Volatile Smile

The funny bit is that the grand daddy of all of this HFT/ Maths in finance was Dr Edward O. Thorpe. He was very gunshy before giving up his university position! Still its funny how even today with all the computers and hardware out there that PeterLBrandt is still able to make a living!

By: tito

tito — Fri, 29 Oct 2010 01:48:34 +0000

@volatile smile I'm guessing you're not spending that $250 in manhattan..

By: volatile smile

volatile smile — Thu, 28 Oct 2010 15:26:23 +0000

Very interesting that it was the business issues and not the trading strategy per se responsible for your woe.

I think this must be the real reason people go into academia (OK, perhaps some cowardice and lack of adventurousness) — business is just risky and tense. Working with partners, you could get screwed, it won’t turn out like you think it will, and in general you’re totally unsheltered from the world’s natural uncertainties.

Sometimes you just want a salary and to forget about all the “dreams” — aka delusions of grandeur.

By: volatile smile

volatile smile — Thu, 28 Oct 2010 15:09:21 +0000

$900 / month is frugal? I consider $250/month frugal.

By: tito

tito — Tue, 13 Apr 2010 11:51:15 +0000

@John Thank you, John. I think domain experience is always helpful. That said, having solid quantitative skills and creativity are certainly key. Simulating HFT is not so obvious as you make it sound, particularly when you need to simulate multiple venues, each with different pricing and latency characteristics. I don't know of any commercial product with such features. Quality data is also expensive to buy but especially to manage.

By: John

John — Fri, 09 Apr 2010 18:55:54 +0000

Hi Tito,
It’s nice to find your blog. Since HFT more depends on models and computers, it’s different to traditional trading. The old experiences are not necessary for this area. My question is if the experience is not important in this area, people could directly jump into HFT and they could learn a lot from their back-testing and simulations for HFT trading. They could also check their strategies and P&L by using historical data. Also we need good simulation platform to simulate execution and dark pools.

By: tito

tito — Sat, 03 Apr 2010 12:34:35 +0000

@George I'll look forward to hearing your ideas.

By: George

George — Sat, 03 Apr 2010 09:50:33 +0000

I have ideas and i will give you an email.

By: tito

tito — Fri, 02 Apr 2010 21:24:38 +0000

@George
Thanks for your kindly comment. I’m not sure what you mean by a ‘master mind group’. I think the real advantage of stratbox (or any proprietary platform) is source code access – you can make it do whatever you like. I think this is why essentially all of the shops I’ve been able to “see into” have developed their own systems – absolute flexibility. But it comes with steep costs and is a thing which can be easily enough stolen. Thus, collaboration which can take advantage of the key benefits of a proprietary system is tricky at best. Indeed, if you work for a proprietary trading firm that has their own platform, they will not only keep track of every bit&byte that you take off their machines, they will force you to sign a non-compete enforcing 6+ months of ‘garden-leave’ when the relationship is terminated.

If you have ideas on how to productively collaborate while protecting the interests of all the parties involved, I’ll be happy to hear them as it’s something I’ve spent a lot of time considering without great result.

By: George

George — Fri, 02 Apr 2010 19:21:20 +0000

Hi Tito,
its a pity that stratbox is not a selling product. It looks really great. Wouldnt you be interested to setup a master mind group using stratbox. I would be interested and I have have valid ideas when it comes to regarding regime switching. I have tested them already with tradesignal, but your platform seem to be more prepared to do what I need to do.

Regards
George