Ernie Chan claims that statistical arbitrage is better than momentum like trend

Discussion in 'Strategy Building' started by GloriaBrown, Aug 2, 2013.

  1. Yes, no, no, yes and put these in any order you like.:)
     
    #121     Sep 9, 2015
  2. Trader13

    Trader13

    Gloria, why do you say momentum trades have better risk control than stat arb? A well constructed stat arb trade (admittedly, challenging to find) is inherently hedged.
     
    #122     Sep 9, 2015
  3. I mean monument trade needs much more and better risk control than stat arb.
     
    #123     Sep 9, 2015
  4. Trader13

    Trader13

    Got it. I read your post too quickly and misunderstood what you meant to say.
     
    #124     Sep 9, 2015
  5. In my own testing, I still cannot find two products that have 80% or higher correlate movement to trade, for example aust and nz dollars suppose to move in very somilar ways, but still too much trend movement and not a good pair for stats arb

    I look at gold price and few biggest gold stocks in local market, not move in similar ways at all. The second biggest gold stock even got "stop to trade and under investiage" for years.
     
    #125     Sep 10, 2015
  6. Trader13

    Trader13

    The great difficulty in finding candidates for stat arb (aka, pair trading) is why it has fallen out of favor. Traders who have found a small number of good candidates for this style of trading will regard them as proprietary as they have likely invested a lot of research to find them. Chan does a good job in describing the fundamentals of stat arb, but his market examples have not been sustainable beyond the time period he presents.
     
    #126     Sep 10, 2015
  7. yeah I think just few proftiable stat arb pairs in the world, and it is so competitive that without high tech, commission, hft advantage is hard to be profitable.
     
    #127     Sep 10, 2015
  8. ernie suggests to trade with 2 or more stuff with 90% or higher cointegration. I cannot even find any pairs with 80%, I don't think I even find any 70%. I mean in any time frame. May be there is working pairs in super fast time frame like MS time frame but that would not be profitable except having a lot of money or much lower than retail commission rate.
     
    #128     Sep 10, 2015
  9. Trader13

    Trader13

    Generally, stat arb is a slower trade that ripens in days/weeks (when it works). HFT time frames usually involve pure arb.
     
    #129     Sep 10, 2015
  10. Any pairs in the world that shows good amount of "stat arb is a slower trade that ripens in days/weeks" in recent years? I just want to learn.
     
    #130     Sep 10, 2015