automised pairs traders, how r u doing?

Discussion in 'Strategy Development' started by man, Jul 21, 2005.

  1. man

    man

    self speaking title i hope.
     
  2. foe

    foe

    Not bad at all

    <a href="http://thetraderlog.com">The Trader Log</a>
     
  3. foe

    foe

    Nobody else?

    <a href="http://thetraderlog.com">TheTraderLog</a>
     
  4. Good, but compared to some of my other automated strategies, not as good.
    The vol divergence risk can be hard to swallow, at least for me. Sectors work
    better since vol doesn't change too much between days. The downside for sectors
    is finding good representive sector baskets, and manual effort needed to
    adjust them.

    Rufus
     
  5. man

    man

    care spend a word on trade frequency, whether you use daily data and what your sharpe ratio is?
     
  6. My system trades often (would almost classified as high-freq), betwen 30-80
    trades a day, for the pairs strategy. I use tick by tick data, in an universe of
    around 500-600 equities (including international equities), updating a
    custom in-memory pairs table / database. I don't compute Sharpe
    ratios often, last time I computed (mid-june), using weekly data
    (normalized to annual), it is 0.62 (nothing to write home about).

    Then again, it is not a classic pairs long / short system, it is more of statistical
    arb strategy ... so I am not sure if it answers your question.
     
  7. man

    man

    oh. yes that is perfect answer. i was so far reluctant to get into intraday with the pair strategy. though it looks like the only way to make money given current VIX-levels ...
    would not mind about sharpe 0.68 in this environment, though in the long run it should be higher IMO.
     
  8. man

    man

    forget my comment "should be higher". who am i to tell?
     
  9. foe

    foe

    Are you using a custom framework or something like Wealth-Lab