Pair Trading Strategy Journal

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

  1. An index is not a tradable instrument anyways. Why not using DIA?
     
    #2691     Jan 23, 2012
  2. nip

    nip

    QUOTE]Quote from sals3r0:

    What is the full symbol on Google Finance? I am unable to find such a symbol (^DJI) - the closest what I can see there is "INDEXDJX:.DJI",
    which does not have data download link anyway.
    [/QUOTE]

    Just a beginner here and appreciate the replies, but a little confused atm.
    Yahoo finance has ^DJI as the Dow Jones Index. The PT I’m running allows other indexes like ^SSEC,^AXJO = Shanghai and ASX200,
    no problems at all with them but not ^DJI
    Cause it’s a symbol found at Yahoo Finance thought it would be accepted by PT, but not to worry

    Ta

    QUOTE]Quote from total_keops:

    An index is not a tradable instrument anyways. Why not using DIA? [/QUOTE]

    Thanks total_keops I should’ve thought of that one.

    P.S. I haven’t traded any pairs at all yet, just feeling my way round and now understand no indexes can be traded just this one had me stumped,

    Thanks guys
     
    #2692     Jan 24, 2012
  3. sals3r0

    sals3r0

    PTL website works with Google Finance data, so if you are not able to find your symbol on Google or there is no "Download to spreadsheet link", it won't work.
     
    #2693     Jan 24, 2012
  4. Amit,

    You are correct. I'm a beginner. So far just made few investments but never traded. Thanks for pointing this out.

    Thanks
    Sachin
     
    #2694     Jan 26, 2012
  5. You can effectively short using a swap. Talk with your prime.
     
    #2695     Jan 29, 2012
  6. CA04

    CA04

    Are Yobo or JohnySharp still around and pairtrading?

    Just curious why some of the really successful guys stopped posting in this thread.
     
    #2696     Jan 29, 2012
  7. Pairtrading works in concept and also in practice with a lot and I mean a lot of work. There is no free lunch right...

    My guess is Johny and Yobo and others are making a killing working 10-12 hrs a day and don't have time (like us) to post to this thread....If you read this topic, at one point I think Yobo said he had 77 pairs open..and thats what you have to do to come out a winner but having 77 pair open means you do nothing by math and research and some more math...
     
    #2697     Feb 6, 2012
  8. #2698     Feb 6, 2012
  9. God this journal's really lost value.

    PairsTradeFinder, referenced as PTF previously, cannot ever give you an accurate backtest because it doesn't have any fundamental data to set a split-adjusted series that won't distort the pairs.

    Also, negative correlation is necessary for pairs trades. I haven't looked at this thread in years, but anytime I do it just staggers me nobody realizes only a highly negatively correlated instrument is suitable for pairs trades.

    If your correlation is not greater than 0.9, or less than -0.9, there is not good chance that the pairs trade will succeed.

    If you're looking for data without this distortion, Wealth Lab Pro has good pairs scripts that you can adjust, backtest, and modify, but it seems the only real factor involved in most of this thread and the value it held to most of the participants was awareness of such correlations. I see too many trades in here without a high enough correlation for pairs trades.

    77 pairs? Are you kidding? I mean, pairs trades in their ripest form only happen whenever there is a highly negatively correlated event like merger arbitrage that's so fleeting in time that whenever you do identify the correlation it is probably to late to put on a good pairs trade.

    Why don't you guys look at leveraged long and leveraged inverse ETFS? Those are the pairs trades with guaranteed negative correlation. I mean, I've worked to find ways to approximate daily data on a long term, and it is not that we cannot create this synthetic data, but you people don't know the first thing about pairs trading and that is that you must have a high correlation.

    Maybe I'm too harsh, but this thread really bothered me that the emphasis was not on taking only high probability trades rather than on pointing out less than optimally correlated securities that might be suitable pairs trades.

    If they aren't >0.9 or <-0.9 do not put a pairs trade on.

    If you don't have software that has split adjusted series that don't fuck up your program whenever they split, do not put a pairs trade on.

    If you haven't done any backtests on the right securities, do not put a pairs trade on.

    If you find yourself with 77 pairs when it is simply impossible for all of those to have a high enough correlation with each other to ever be suitable for pairs trades, do not put a pairs trade on.

    If your emphasis did not involve the discipline of highly correlated assets, negatively or positively, then you have mistakenly falled for random spurios correlations that have no basis for any pairs trades.

    Pairs trades are only for highly correlated assets with those characteristics. If you don't have a security that is either correlated by a factor greater than 0.9 or a factor less than -0.9, do not put a pairs trade on.

    I think the so called "succesful" traders got blown out by lack of research, when even the most cursory examination of these methods implies that only in the case of a reasonable amount of time of purely highly correlated price movement should be the only time you should ever use things like PTF, Wealth Lab, and any other correlation analyzer to put a pairs trade on.

    The next one of you I see announcing your latest pairs trade, I might come back here to yell at you so that I can emphasize my point that this thread is a failure because it did not get the first thing about pairs trading.

    If you don't have a security with a 0.9 or higher or -0.9 or lower correlation, do not put a pairs trade on.

    I can't believe the investing public doesn't get this. I've emphasized this for years. Most pairs trades in this thread are from spurious correlation or cases where correlation was not even a consideration in the suggestion. Pepsi and Coke can be popularly classified as pairs trades when they are anything but because they are not correlated, and, even worse, they are not negatively correlated.

    I can go on and on here, but I'm in this thread previously making these exact same points. There aren't 77 pairs in this market in stocks to base pairs trades on. I suggest looking only at quantifiably long term correlated and negatively correlated trades with enough correlation for the trade to have significance and to even be classified as a pairs trade.

    If you can't identify at least a 3 month period of 0.9> or -0.9< correlation, DO NOT PUT A PAIRS TRADE ON!
     
    #2699     Feb 6, 2012
  10. You getting angry like this is silly. Even if every pair trader who participated in this discussion is complete moron there is no need to come out swinging like this. What is with you people. Get some medication or something...
     
    #2700     Feb 7, 2012