Home > General Topics > Trading > So they gave me a Quant

So they gave me a Quant

  1. Quick background: running a CTA using technical models developed over time after coming from discretionary prop trading. Work with a few allocators who have seemingly been happy with performance so far.

    At the beginning of the year, one of the allocators asks me if I'll work with this "new talent" he's found and wants to cultivate. I agree under the conditions that he have nothing to do with my current programs, and then maybe we can develop something together. At the time I figured what the heck, the allocators paying him and I want to keep the guy happy.

    So the quant comes into the office for the past 2 weeks and so far all I know is that he is good at getting lunch and bad at making restaurant reservations! So now what?

    I've worked with quants before who were either introspective like rainman or hyper to the point that they were seemingly on some combo of crystal meth and jolt cola. But this guy is just a zombie, doesn't speak at all, and hasn't done a thing yet. I was expecting him to just come in and do his thing, maybe I'd pull his head out of the sand every once in a while, and teach him a little bit about trading.

    While this has been mildly amusing for the past few weeks, I'm now starting to grow slightly concerned if asked for a progress report. The allocator knows that I am not a math guy, so I'm not certain how to lead this guy.

    So what should I have this guy do? I've got access to good market data, but what relevant stats should I have him compile? My thought was to just have him calculate a bunch of stats and then try to find some statistical significant relationships that could potentially be traded off of. But where to begin?

    I will share the results of his findings with anyone willing to help me out with a starting point or a plan to follow. Provided of course that this guy can do it. PM if necessary.

    Man, the sh!t I get myself into lol. Thanks in advance.
  2. if your already successful.. why do you need a quant? i'm a skeptic

    you have intellectual property that is up for the taking.. i'd be careful..
  3. Sounds like total BS from a finance program drop out
  4. Quants are essentially a programming & datamining interface. If you can't explain what you need in plain English then you should both be shit-canned. Sorry to be blunt.
  5. This.

    Sounds like the guy is either useless or a spy. Have a significant talk with your allocator and make him tell you exactly what he sees in the guy that deserves cultivating.
  6. i agree with the interface part .... the guys a tool.. use him.. if i were you.. i would put the guy working on stuff that doesn't relate to my current strats.. use him to look for new ones.. you could as a trader i'm sure generates thousands of questions to ask data and get him to find the statistical significance to it.. the guys check isn't coming from your pocket.. you should put him on some stuff you have been curious about to see if you can discover something new.. meanwhile i'd keep your current edge to yourself.. look at him like a free research option respectively..
  7. Thanks for the replies guys. Caveman, my paranoia meter was the first thing that popped up, but my stuff is protected on a secure server offsite and this guy doesn't have access to anything. I'm not worried about theft of IP.

    For the rest of you, I don't need a quant. This guys was sort of foisted on to me and I initially said yes cause I figured this guy would be plug and play like the guys I had worked with in the past. I don't have anything I want or need done, I'm just concerned I'll be put in a bad position if this guy doesn't do anything.

    I've asked I'm what he looking to do, what areas he's looking to explore, etc. said he didn't know. A lot of you have PM'd me some cool project ideas but I think I'm just gonna spend the weekend on Wilmott and figure out some stuff this guy can do.

    Thanks to those who were helpful, and of course the usual eff you to those who weren't.
  8. This is exactly what I decided to do. I figured I'd cull a few ideas here and then share the results with the person who suggested it (them). Gonna give it a try next week. Will let you know how it goes.
  9. If you're a CTA (not a CPO) then they have access to you fills anyway. Why would they send in a quant to steal your IP? If you're paranoid then why did you let him in?
  10. They wouldn't and I'm not. That was an initial reaction, and response to caveman.
  11. Tell him to read a book, and figure it out himself. Not clear to me how you can really "help" him beyond that.

    He can start with Rishi Narang's "Inside the Black Box" for a non-technical, high-level overview.
  12. one key point is.. the guy isn't interested it seems.. thats going to be the big hurdle!
  13. There are tons of items you could give him to do. For instance, you could ask him to tabulate returns, distributions, statistics, etc.. and produce charts and tables to compare to other established CTAs. Certainly the results will help on the marketing side (well, assuming they are good to begin with).

    Then ask for feedback (slides/reports) on what similarities and or differences he finds and what areas can be improved. At least you can see if you are in the ballpark with heavyweight funds and can see if he's capable of doing some simple tasks for a start.

    If you don't want to divulge your ideas, find some good papers and ask him to reproduce the results (the process can be harder than it seems) in a code language you can work with. Take advantage of this resource.:)
  14. You have been wishing him to fail and go away from the start, which is a terrible and irresponsible thing to do to someone. He is picking up your vibes.

    Look for a more positive approach. If you cannot find one, terminate the arrangement now.

    Do not ever agree to something involving people that you are not really on board with.
  15. He could spend his whole life in JH methods!

    Provided he is really a very good quant!

  16. JH?
  17. Not true. Supposedly this kid is a quant genius, but for all I know he could be this guys bastard love child. All that aside, I wish I had some huge project for this kid to get into, I just don't. That's why I was asking for direction.

    Sure I can have him do a bunch of mundane stuff, run some excel programs figure out some correlations, etc. But what happens in a month when I'm asked what we been working on and I say some crap an intern would do?

    This guy is likely not with me forever, but probably going to do a rotation throughout this guys other funds. I didn't enter into an arrangement, this was forced on me. It's not a horrible thing, it's not like the guy asked to bang my wife.

    Regarding my attitude, I'm a 15 year market vet and I've got to bend over backwards to motivate some kid to do what he is supposedly talented at? Eff that. I do want to see this guy succeed, because of course I can take credit for it. So don't assume things about my approach.

    I've already gotten plenty of good responses and with a little research this weekend I think I will have something that should keep him plenty busy. And who knows, if he does it well I may even look to get more involved.

    Thanks Heech, for the book recommendation. I'm going to read it myself first.
  18. You could ask him to do a study of your trading results and make suggestions.

    Even better, give him at least an outline of how your reading approach operates, but a lot can be said even with just the results.............
  19. I would not focus his time on black box HFT, as that really involves computer infrastructure more than quant work. I would ask him to develop a futures trading model based on yield curve spreads. This is quantitative, interesting, and has lots of potential if he is really a genius.
  20. 1. Assign him a "project" to illustrate a curve fit:

    a. using TA of his choice (macd crossover or something basic) create a directional trading program and optimize over:

    i. one month

    ii. six months

    iii. one year

    iv. five years

    v. 10 years of data

    b. Each trial will include an out of sample (blind) performance report at least 20% as long as the training window

    c. Conclude the assignment with a graph of OOS performance vs. in sample returns for the various time frames

    2. Repeat the assignment using another TA, fundamental, spread, correlated, or other strategy.

    3. Present the most robust (least degradation OOS as % of in sample) and the highest performing (max return as % of max DD)

    4. Conduct sensitivity study on these two finalists by tweaking the parameterz and noting the performance change

    ...... that should keep him going for a week or two.