Thoughts on finding help with analysis

Discussion in 'Strategy Development' started by AntelopeStew, Apr 20, 2008.

  1. I've reached the point in my system design efforts where I need input from a very capable statistician. My question is whether I should seek one within the trading community or from outside (e.g., academia).

    The argument for finding one inside the trading community is that the person will already understand the markets and the nature of market data. The argument against is that the person may bring a significant amount of bias to the effort based on their own research/strategies.

    The argument for finding one outside the community is a lack of the aforementioned bias. The argument against is finding someone who will be interested enough in the markets to bring sufficient energy to the task.

    Any insights are appreciated. Thanks.
  2. LOL

    in other words you need help developing an edge

    in other words you don't have one now

    in other words you want to team up with a guy like me to develop an edge


    it don't work like that, trading is a solitary thing unless you have a brother you can trust
  3. An argument against seeking one within the trading community that I foolishly omitted is that the community is filled with smug, over-assuming dicks who speak before they think. Thanks to "original" for demonstrating that fact. I can definitely say, without reservation, that I don't want to team up with a guy like you.

    Just to clarify, I have an excellent edge which is both theoretically sound and effective in practice. I believe a robust analysis would be helpful in order to build a set of filters which would facilitate its implementation as an automated strategy. I specifically want someone who DOES NOT have their own set of ideas. I want a clean slate who is well versed in statistical modeling.

    That said, I'm not seeking an analysis partner with this post; I am seeking advice as to whether it would be wiser to search for such a person within the trading community or outside it. Clearly, original's post is a vote to seek one outside. Other opinions are welcome. Thanks.
  4. I won't insult you back cause you must be new here

    and don't know who I am so it is forgiven

    let me ask you a question

    why can't you do the statistical studies yourself

    there is no shame in saying you can't do any programing

    in any case, if I needed someone to program for me

    I would use about 4 guys and do strategic division of labor

    you seem like a semi-smart guy

    you should figure out why

  5. I'm well aware of the importance of protecting my intellectual property.

    Programming is not an issue. I am actually a very good programmer. However, I know enough about stats to know I don't know enough. I could slap together some analyses, but unless I match my specific data with the appropriate statistical techniques the results could be meaningless and/or misleading.

    This could be seen as analogous to my designing and building a house. Let's say I have a fabulously creative idea for the design. I can draw reasonably well, so I could put together some blueprints. Finally, I'm pretty good at carpentry, wiring, plumbing, and other construction skills, so I could take a crack at actually building it. In the end, however, I would be unwise not to consult an architect or engineer about whether I incorporated the appropriate design elements in order to have a safe and long-lasting home.

    To sum up, I'm not excited about sharing my ideas with another person, but it seems like a necessary evil for my system to reach its full potential.
  6. Mschlau


    The simple and most likely unwanted answer to your question is that you should seek insight from both academia and from a trader. This way you can hedge against bias. I'm no quant but I do have some experience with stat arb, so you can PM me if you want to discuss any of this in detail.
  7. Bo_D_


    what statistical information are you seeking?

    i trade multiple mean reversion systems and i am no mathematician or a quant. but i have a system im happy with, and i dont have heavy complicated equations or analysis.

    is it worthwhile? if the answer is yes then i obviously have a lot of work to do.
  8. My opinion would be to go with an academic if you can find one who has time and interest. They tend to be passionate about their field, often enjoy sharing their enthusiasm with others, and tend to be quite happy with their world of theory and so would most likely not present a threat to your proprietary trading style, as I have found that few academics actually trade. Best of luck in your quest....
  9. just go to a university math department and post a "job" for 20 bucks an hour or whatever on a cork board.
    I'm sure you can easily find a grad student who could use some money who has more than enough stats knowledge for what you want to do.
    The big advantage to not using someone from the trading community is they will have no bias and simply look at the data you provide.
  10. Agreed, an upper level math student might be perfect.
    #10     Apr 24, 2008