Marc Cohn's  Instablog

Marc Cohn
Send Message
I am a patent litigator with a background in physics and electronics. I enjoy studying quantitative, rules-based investment methods through rigorous backtesting and numerical analysis. I believe that patterns exist in the market that benefit trading -- the challenge is finding them!
Back To Marc Cohn's Instablog HomePage »

Instablogs are blogs which are instantly set up and networked within the Seeking Alpha community. Instablog posts are not selected, edited or screened by Seeking Alpha editors, in contrast to contributors' articles.

Comments (31)
Track new comments
  • mpd3892
    , contributor
    Comments (21) | Send Message
     
    Marc, how confident are you in this signal? I'd be very interested in your take. Thanks.
    30 Apr 2014, 02:21 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » My gut gives me doubt about this signal because ILF has had such a big run up (15%) in the last three months. But this is how the system works; we are supposed to by into the momentum. The system is not intuitive, which is perhaps why it works (i.e. people don't follow it because it feels like pissing into the wind). It is admittedly uncomfortable.

     

    The numbers, however, do not lie. I looked at all the 15% 3-month run-ups in the subject ETFs. The next month was up 63% of the time. Also, the ups averaged 6.54% and the downs averaged (3.87%). Based on these probabilities, our expected value is 6.54% * 0.63 - 3.87% * 0.37 = 2.68%. (The sample size was 124.) Statistically we look good.

     

    I was not very confident at all about the last EDV signal... but it turned out just fine. If I could take 2.8% a month every month, I'd be a very short time.

     

    So, we follow the system and let it do the worrying for us.
    30 Apr 2014, 08:50 PM Reply Like
  • Steadyprofit
    , contributor
    Comments (45) | Send Message
     
    Much appreciate you posting these updates. Looks like last months EDV selection was another successful selection by your modified GMR; well done.
    30 Apr 2014, 03:11 PM Reply Like
  • mbaril
    , contributor
    Comments (5) | Send Message
     
    Marc, was your correlation around 0.2?
    30 Apr 2014, 05:30 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » My correlation is -0.95. About as negative as it gets. I use a four-month correlation of the monthly SSO and EDV prices.
    30 Apr 2014, 08:52 PM Reply Like
  • mbaril
    , contributor
    Comments (5) | Send Message
     
    I used end of month close. What price do you use?
    30 Apr 2014, 09:05 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » End of the month close.
    1 May 2014, 10:23 AM Reply Like
  • mbaril
    , contributor
    Comments (5) | Send Message
     
    Marc,

     

    I don't see how you get -0.95 since SSO and EDV have been moving in the same upward direction for the last 4 months. These are my numbers and I get 0.68 for the last 4 months using the CORREL function. What am I doing wrong?

     

    SSO EDV
    Jan 95.14 97.39
    Feb 103.76 97.73
    Mar 105.31 98.83
    Apr 106.61 101.74
    1 May 2014, 05:15 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » Need to use dividend adjusted prices.
    1 May 2014, 08:58 PM Reply Like
  • mbaril
    , contributor
    Comments (5) | Send Message
     
    I understand now. It's the correlation of the monthly returns, not monthly prices.
    2 May 2014, 02:53 PM Reply Like
  • rscricket
    , contributor
    Comments (5) | Send Message
     
    Maybe I am missing something still.

     

    I went to yahoo finance and searched for EDV > historical prices > click 'Monthly' radiobutton. This gave me monthly data for EDV which was: 96.96,98.83,101.74, 103.58 for first trading day of months Feb through May (4 months).

     

    Exactly similar procedure for SSO gives: 103.73, 105.31, 106.61, 106.64.

     

    Both are moving in the same direction. So their correlation should be positive not negative (even if you consider monthly returns which I dont see why you should).

     

    Also SSO data above shows May 1 val as 106.64 and Feb 3 val as 103.73 (dividend adjusted). Its 3-mo performance is 2.8% not 11.68% as noted above.

     

    Help!
    6 May 2014, 08:45 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » You're not using the Yahoo data correctly. Even though it says Apr 1, that is referring to the last day of April, not the first. For "May 1" it is the current price -- check it today it will be different than the last time you checked.
    7 May 2014, 08:54 AM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » Also, your data points should be the Jan-Apr entries on the yahoo finance site. Take the LN of monthly returns and the correlate. Should be quite negative.
    7 May 2014, 12:12 PM Reply Like
  • rscricket
    , contributor
    Comments (5) | Send Message
     
    Marc, thanks for clarifying. As suggested, I got prices at end of month for SSO and EDV, calculated returns, LN'ed them and correlated them to get -0.99.
    7 May 2014, 06:46 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » Sweet! And y'all think I'm crazy. ;-)
    7 May 2014, 08:14 PM Reply Like
  • lingkui
    , contributor
    Comments (37) | Send Message
     
    Marc, your other strategy, based on your backtesting determined that 85-day look back (4 calendar months) is optimal while this strategy showed 3-month look back is superior. Is it possible to explain the discrepancy? TIA.
    8 May 2014, 10:30 AM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » Totally different ETFs.
    8 May 2014, 02:20 PM Reply Like
  • Dagarath
    , contributor
    Comment (1) | Send Message
     
    How do you LN negative returns?
    14 Jun 2014, 09:37 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » ? How do you get negative returns?
    16 Jun 2014, 08:53 AM Reply Like
  • Raj2020
    , contributor
    Comments (171) | Send Message
     
    Marc - Aren't the performance numbers listed raw numbers (meaning not adjusted for volatility as your model dictates)? Unless I missed an update about not using volatility any more.. Going forward, are you just comparing raw 3-month performance numbers to determine which instrument to invest in? - Thanks!
    1 May 2014, 02:53 AM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » I just use the raw performance numbers.
    1 May 2014, 09:12 AM Reply Like
  • OwnPuts
    , contributor
    Comments (24) | Send Message
     
    Can you explain why you used the correlation above a threshold between SSO and EDV as a cash stop?
    Would this only be when stocks and bonds are both selling off?
    Or are you also saying if both bonds and stocks are going up you would rather be in cash?
    2 May 2014, 11:43 PM Reply Like
  • olmendreef
    , contributor
    Comments (49) | Send Message
     
    Hi Marc,
    Great work. I have been reading a lot about M Grossmann and your approach and will implement it shortly.
    One question: would you consider adding ZIV to your ETF list? Of course, in this case, I would re-evaluate the ZIV every 2 weeks.
    Best regards.
    22 May 2014, 12:05 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » I tested using ZIV with a 2 week rotation and it did not work as well as the system as presented.
    22 May 2014, 12:41 PM Reply Like
  • Max1985
    , contributor
    Comments (11) | Send Message
     
    I tried to incorporate ZIV into a strategy very similar to this. It's very volatile and spread can be wide.
    23 May 2014, 03:57 PM Reply Like
  • mpd3892
    , contributor
    Comments (21) | Send Message
     
    Marc, I just read a previous article of yours on a paired switch using SPY/TLT based on an 85 day look back. Here we're using a 3 month look back and more ETFs. What are your thoughts in comparing the two systems? Why the different look back periods?
    25 May 2014, 01:09 AM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » The assets being traded are different.
    25 May 2014, 08:55 AM Reply Like
  • olmendreef
    , contributor
    Comments (49) | Send Message
     
    It could be interesting to backcheck the system with a take-profit (2,3,5% or end-of-month if not reached) based on daily prices.
    29 May 2014, 12:45 PM Reply Like
  • Marc Cohn
    , contributor
    Comments (88) | Send Message
     
    Author’s reply » I have done this. Take-profits do not help.
    30 May 2014, 07:56 AM Reply Like
  • rolandus11
    , contributor
    Comments (5) | Send Message
     
    Hi, is it fair, Lat. Am. 40 to compare with the MSCI Asia ex. Jap. . Would not DAXGlobal Asia (40 components, there exists a fund RBS Asia DAXGlobal, very good performance in May)) better?
    29 May 2014, 02:04 PM Reply Like
  • mpd3892
    , contributor
    Comments (21) | Send Message
     
    Almost 2 months into the current signal and the model has once again proven itself wise. I remember asking Marc at the time of the signal how confident he was in it. With little more then a week to go until the end of the month, it looks like we might get a new signal telling us to go to SSO. By my rough estimates (and they are rough) it looks like SSO has more than a 3% lead over ILF. Any thoughts?
    22 Jun 2014, 05:42 PM Reply Like
Full index of posts »
Latest Followers

StockTalks

More »

Latest Comments


Instablogs are Seeking Alpha's free blogging platform customized for finance, with instant set up and exposure to millions of readers interested in the financial markets. Publish your own instablog in minutes.