QIM Method

Discussion in 'Strategy Development' started by DT-waw, Jan 24, 2008.

  1. Quantitative Investment Management (QIM) quickly becomes one of the biggest CTAs in the world, rised $3B in the last two years. So i got interested in their trading methodology. Here's the article which describes it in general.
    from http://www.allbusiness.com/specialty-businesses/479588-1.html

    After much contemplation, Jaffray Woodriff settled on quantitative behavioral finance to describe his trading methodology. It is a mouthful, as is behavioral finance academics, artificial intelligence and evolutionary programming--all apt descriptions of what he does. It also is understandably complex considering he has designed software that has created and tested hundreds of billions of trading models. The best models are employed in Quantitative Investment Management's (QIM) Global program.
    [...]
    Woodriff's software, which is based on mathematics, utilizes artificial intelligence techniques that take a group of relatively simple inputs and create a vast number of trading models. Those models are tested throughout 25 years of market data in 40 different markets and the best 1,500 models are used in the program. The models are scored on performance and correlation to each other. They create an overall score, between -500 and 500, for each market. A score above 100 or below -100 will generate a buy or sell signal. It is a compilation of the all the models that recognize market patterns; some are trend following and some are countertrend, while others are neither.

    "The models are generated by computer code, they're not selected because I went out to try and find a countertrend model that would do great at long-term market turns. The models are screened algorithmically based on how well they predict the movement of markets at any time," Woodriff says. He ends up with many long- and short-term models. The composite signals tends to be the short- to medium-term type signals, holding trades an average of eight to 10 days.

    He separates his methodology from other managers involved in behavioral finance who create methodology on the psychology of market participants. "We take a much more systematic approach to finding patterns," he says. They will come up with a hypothesis that makes sense and then test it quantitatively. "While we have created a software program that goes out and searches for possible mathematical formulas for patterns, it is a much broader search," he says. "We take a very generalized approach that there are patterns in the market and we are going to generate potential patterns and see how well they do by generating billions of them and see what comes to the top."
     
  2. It seems like they have replaced the human development of trading models by the software which generates a massive amount of models and then they choose a combination of 1500 best ones in terms of performance and total correlation. Hm i wonder how much computer power it is needed to choose 1500 models according to the their lowest correlation score out of the corr. matrix which contains billions of models :eek:

    Also it seems like a monster optimization to me - to screen all possible models and their combinations and then choose the optimal set of them, without undertaking any intelligent insight into the internal logic (rules) of each model and without asking questions about the reason why these specific rules where implemented in the model.

    In my judgement such approach may have an edge over traditional trading systems development in terms of high chances for discovering patterns which aren't found by other managers who use "human" development.
    Comments?
     
  3. Good marketing .
     
  4. Is it only a marketing game without any connection to what they actually do? I guess they don't need such cheap marketing tricks since their flagship program targets institutional investors, min. investment is $20 million.
     
  5. QIM has raised money based upon great performance and great investor communication.

    So far they've proved themselves well. Give them credit
     
  6. Credit? There is something called the "discovery problem." They are only in the news due to their performance. There are large numbers of CTAs who muddle through, that you don't hear about.

    If they continued this for 15 years, then I might be impressed. These are the kind of outfits, that will suddenly lose 60%, and you will stop hearing about them.

    Nothing has been proven that they are any better than others. A few years of performance is only proof that past performance is not necessarily proof of future performance.
     
  7. MGJ

    MGJ

    A similar concept is available for sale, today, to the retail trader.

    (Vendor 1) uses genetic algorithms to write and optimize trading systems. "Finally! An algorithm that writes algorithms."

    (Vendor 2) generates millions and billions of trading systems by making random combinations of 2,500 built in setups and triggers and signals. Their high speed test engine backtests each of these systems, and presents you with the backtest results and system code for the top 1000 of them.
     
  8. Nattdog

    Nattdog

    sounds pretty much like the same stuff the other short term futures managers have been doing: statistical pattern recognition. For market wizard buffs, Monroe trout describes doing this in the second wizard book.
     
  9. Nattdog

    Nattdog

    another point is that they didn't quite start from ground zero in the business and get to where they are in a few years. both principles are highly experienced and have been in the game for a long time.. What they did was leverage that market/research experience and convert it into a successful, customer based business. many lessons here.
     
  10. bad start of the year for the firm, down over 7% through 25th January.
     
    #10     Jan 31, 2008