Hired Talent: What are the Costs and Risks?

A

achilles28

Like all good pigs, I want to maximize ROI for strategies that go largely unused, save a couple instruments scalped manually.

Automation holds great benefit: access to 30+ instruments during their regular trading hours, instead of ~2. Further, quick markets more suited to program trading can be accessed, instead of ignored.

My technical skills are few. I can write simple code and have done so for 3 strategies in AFL (Amibroker Formula Language). What I need is a competent programmer who can automate existing systems, and have them trade flawlessly, real-time. Here's a short-list of estimated needs:

- migrate strategies to C+(+)
- integrate strategies to various API's
- test and debug over several months
- ongoing monitor and administration
- ongoing system development (my own toys) and light optimization
- hardware acquisition and support
- security design and administration (can be outsourced)

The term of employment would be full-time and ongoing. Insofar as computer trading is used, I would need a full-time administrator.

The obvious risk is an outsider must be trusted with high-value, intellectual property. The temptation to steal, sell or transmit IP for their own gain will be very real. Besides non-disclosure and non-compete contracts, what other legal safeguards exist for intellectual property not subject to patent? Even such legal contracts provide no guarantee as theft or transmission would be very hard to detect.

To explore further, how do top Wallstreet Banks and Brokerages secure proprietary trading systems against theft and disclosure? They too must entrust outside help - programmers and quants - with full access to source code. Those employees could easily trade that code for themselves or sell it to anyone. What protection does the industry have against this? Maybe consultants could advise in this area, too?

As far as compensation goes, what is fair? I think the average is ~250K for system designers/quants plus a healthy bonus. Perhaps ~1 million a year. The exception: outside help is not required to lay the golden egg. Only the technical implementation of existing eggs. So what is fair in terms of salary? 150K per year? What about a bonus structure?

It would be nice if guys with industry experience/programmers/Wallstreeters could give some input. Thanks.
 
Quote from achilles28:

What I need is a competent programmer who can automate existing systems, and have them trade flawlessly, real-time.
More...

Never seen that happen. Take a look at the problems that IB's software has. Still lots of flaws.
 
Quote from Roark:

Never seen that happen. Take a look at the problems that IB's software has. Still lots of flaws.
More...

HFT trading accounts for 70% of the volume on US stock exchanges. Clearly, it can be done. I don't mean to sound snobby, but I don't want an amateur. I want a pro. And I wouldn't use IB. Not sure yet, but at that level, a prime broker.
 
Sure, you want a pro. That's why a big swinging dick that's rich as fuck asks for free advice on a message board instead of asking actual professionals. Of course you're rich, you use Amibroker. It's the preferred choice of mega-rich mofos.
 
Quote from Roark:

Sure, you want a pro. That's why a big swinging dick that's rich as fuck asks for free advice on a message board instead of asking actual professionals. Of course you're rich, you use Amibroker. It's the preferred choice of mega-rich mofos.
More...

This is a serious thread. If you've got nothing to contribute but hostility, please stay off it.

As far as the rest, there are pro's on ET. And Amibroker is fine for backtesting. Especially for laymen, like me. So who are the professionals I should ask?

And to clarify, the strategies are not high frequency. They are trend-following. Maybe 10-20 trades an hour, per instrument.
 
Amibroker is fine for playing around. But I found the route to code to IB was a mess and vowed to never go that route.

But you'll find that there is a huge barrier of entry into HFT when you go to choose the software, data and API you connect to. A lot of the better brokers wouldnt go near you without millions in cash to deposit and possibly, a business plan. They don't need somebody caught in a for loop flooding the Market with orders on a $25,000 account.
 
To clarify, my strategies aren't high frequency. I only referred to HFT to demonstrate computer trading is done successfully, and with great volume. Clearly, Roark thinks otherwise. My strategies are trend-following. Average hold-time = 2-3 minutes.

So where did you end up? How about any input about the level of programmer I need for the job? Pay? Security?



Quote from psytrade:

Amibroker is fine for playing around. But I found the route to code to IB was a mess and vowed to never go that route.

But you'll find that there is a huge barrier of entry into HFT when you go to choose the software, data and API you connect to. A lot of the better brokers wouldnt go near you without millions in cash to deposit and possibly, a business plan. They don't need somebody caught in a for loop flooding the Market with orders on a $25,000 account.
More...
 
Quote from achilles28:

To clarify, my strategies aren't high frequency. I only referred to HFT to demonstrate computer trading is done successfully, and with great volume. Clearly, Roark thinks otherwise. My strategies are trend-following. Average hold-time = 2-3 minutes.

So where did you end up? How about any input about the level of programmer I need for the job? Pay? Security?
More...

How times have changed. I remember a time when an average hold time of 2 minutes was considered high frequency. Now its considered trend following.



:D
 
Quote from achilles28:

HFT trading accounts for 70% of the volume on US stock exchanges. Clearly, it can be done. I don't mean to sound snobby, but I don't want an amateur. I want a pro. And I wouldn't use IB. Not sure yet, but at that level, a prime broker.
More...

Ok, so you want someone like me in my day job. I work as programmer / architect / team lead on highly complex systems, currently I am one of the leads of a risk monitoring project for a smaller energy trading company (1500 employees). We talk of serious work here - 6000+gb data to be monitored, we just ordered a couple of Exadata systems ;) We run a team of about 30 people at the moment in 4 different groups (of varying sizes - one has 2 people, another 4, and those are not the ones I am involved with...)

The bad news: I earn a lot ;)

20.000 to 25.000 USD monthly (depends on working days) as contractor in commitments full time 6+ months (i.e. the contract says minimum 6 months ful time which is 10 hours per day).

I work on autoamted trading in my free time, but that got stopped for this year - I simply dont ahve the energy after work to do even more market data stuff.

One thing you must be carefull here is - people which are good in IT earn a LOT. This is one of the jobs where there has not been a project shortage during the financial crisis - projects were happy to pick up the free talent. Average IT guys earn a lot less, but then they dont do financially criitial stuff without supervision. What you look for are the junior people we hire on our teams, because you wont be able to pay the seniors.

So, you are better going wih a student and working through the problems. Finding someone with a reputation and the known skill set will cost you a lot more than is economical for you. Especialyl dealing fully automatic in all edge cases, or having half automatic emergency mechanisms automatically kick in (like a sms to you about calling the broker and closing positions because there is a communication failure) is stuff that takes months to develop, properly test etc. Not even talking about the infrastrcture - you need a minimum of 3 systems for working in the professional way you say, including full support frmo your brokers. One for development, one for testing, one for production (and possibly 1-2 backups if you want this fully automatic under pretty much all conditions).

This is SERIOIUS stuff, a serious investment. I doubt it is cost efficient in your porfolio size. seriously. Go with some talent without experience and just pay for the losses when something goes wrong.
 
Thanks for that detailed, albeit presumptuous reply. Low six figures is within reason. Not sure why you assume I trade a small account? The purpose of the thread was to solicit feedback on the costs and risks involved with full automation. You outlined the costs well. How about security of intellectual property? What best practices are common in the industry to protect strategy integrity from sticky fingers? Like I said before, I'm definitely looking for a professional. Not a student. At the end of the day, I would need a capable hand for the long-term.

Quote from NetTecture:


So, you are better going wih a student and working through the problems. Finding someone with a reputation and the known skill set will cost you a lot more than is economical for you. Especialyl dealing fully automatic in all edge cases, or having half automatic emergency mechanisms automatically kick in (like a sms to you about calling the broker and closing positions because there is a communication failure) is stuff that takes months to develop, properly test etc. Not even talking about the infrastrcture - you need a minimum of 3 systems for working in the professional way you say, including full support frmo your brokers. One for development, one for testing, one for production (and possibly 1-2 backups if you want this fully automatic under pretty much all conditions).

This is SERIOIUS stuff, a serious investment. I doubt it is cost efficient in your porfolio size. seriously. Go with some talent without experience and just pay for the losses when something goes wrong.
More...
 
You must log in or register to reply here.
Back
Top