Pair Trading Strategy Journal

Discussion in 'Journals' started by jonnysharp, Aug 18, 2008.

  1. magon

    magon

    On PTF they call spread to the difference between the stocks value and they show it on one of the charts, is this what you are refering to?
    In my opinion this value is no usefull at all. I call spread at the % from the mean value, and this is an indication of the money you can make on the trade.
     
    #1731     Dec 17, 2009
  2. waltbx

    waltbx

    The PTF "Spread" chart is the historic difference between stock prices. The spread you are talking about is spread of ratio to ratio mean in %. Understood.

    I'm still trying to determine what it means if the price spread is increasing, or decreasing, and what happens in the last two weeks of Dec when the spreads increase. I'm thinking it's not good.

    wb
     
    #1732     Dec 17, 2009
  3. magon

    magon

    If you enter the same amount of money per side, the smaller price stock would have a higher coeficient and the profit equation, and then this price spread doesn't have any meaning, I just care about % movement.

    Have anybody bought the autometed excel spreadsheet yet?
    Do you think is worth it?

    SOME PICS FROM TODAY

    BEN:TROW
    XRAY:MR
    GEL:CPNO
     
    #1733     Dec 17, 2009
  4. waltbx

    waltbx

    I don't see how it would benefit me.

    I've used a spread sheet to track all of the Twitter PTF alerts, since 08/20/09. At the beginning of November I began to insert my notes and a Yes or No if I would take the trade. Investing $10,000 on each side of the pair (about 40 pairs), the alert system profited $2,046. Not their best 6 weeks. However, because I manually analyzed each trade and rejected what I thought would be poor trades, my Yes picks from those alerts yeilded $4,916.

    So, using a strictly automated system would have lowered my yeild by $2870, 58%. I'm willing to put in the small amount of effort to enter my own trades from their alerts. And I have the time.

    The alert spreadsheet system does what it is designed to do: Make a decent profit for those who don't have the time, desire or ability to analyze trades.

    Walt B
     
    #1734     Dec 17, 2009
  5. magon

    magon

    I agree, I like to select the PTF Alerts and also I like to run PTF program and select my picks by sorting the pairs by the sigma (< -2 and > 2) then looking for correlation > 70% and RSI <30 and >70 respectively. I usually have from 4 to 6 pairs at a given time.

    I also look in the trigger screen , but I can find pairs that didn't stop at the 2 sigma band and are returning to the mean using the first method.

    After I get my candidates I use Quotetracker to check visually the cointegration in 1 year and 6 months comparing the 2 stocks on a chart.

    On stockcharts.com I like to use the 50 day moving average to see the trend on the ratio, I like it to be uptrending or flat also the fundamental P/E ratio < 0 and EPS ratio > 0.


    Thanks Walt for your feedback
     
    #1735     Dec 18, 2009
  6. May have been discussed already, but how does one measure if a divergence between co-integrated stocks is caused by informational shocks or misinformed traders.
     
    #1736     Dec 21, 2009
  7. magon

    magon

    killer04,

    This is a tought question, the best you can do, I guess, is just look for news about the stocks and see if there is a real reason for the divergence.
     
    #1737     Dec 26, 2009
  8. I have decided to make this a pair trade by buying a stock in the Consumer Staples sector. At 1st I was looking to go long HANS by buying an in the money call option or buying the stock and selling a covered call. However, I saw the JSDA already received a low ball cash offer of $ .30/sh, with the stock selling at $ .48/sh, the market seems to believe that the offer needs to go up. I don't see a deal getting approved at anything under $ .80 to $ 1/sh. At $ .48/sh, we basically have an option with no expiration. I have tried their products, and like them, but did not like that they had no diet drinks. Now that they are coming out with a diet soda, I decided to buy 5,000 shares.

    For GMCR, I bought in the money Jan put, that I will rollover if price does not hit my stop. Goal is for GMCR to go down in price and JSDA to go up in price. However, I will also be happy if one of them moves while the other stays the same. If GMCR falls, I plan to sell 4,000 shares of JSDA, but keep 1,000 in case they do get bought out at a higher price.
     
    #1738     Dec 30, 2009
  9. elit

    elit

    Could someone help me understand how to interpret calculation of Cointegration?

    I have a spreadsheet (found in this thread) and enter a time series (closing prices of stocks). Excel then calculates the cointegration and presents some values as a result. I don't understand how to interpret the resulting values, and I don't know wheter the stocks are cointegrated or not even though I have the calculation.

    Could someone please help me understand? I have searched google for several hours to understand the results of conitegration calculations but, I haven't found anything and I simply don't understand it.

    What are "critical values"?
    What is the "number of lags"?
    What is "test statistics"?
    "H0 : Rank<=x"?
    What is the difference between eigen and trace?
     
    #1739     Jan 1, 2010
  10. perhaps one of the best is:

    long or short 1 ES vs. opposite longs/short 1 YM...please study and evaluate this
     
    #1740     Jan 1, 2010