% | $
Quotes you view appear here for quick access.

American Capital Agency Corp. Message Board

  • reits_r_us reits_r_us Feb 2, 2013 9:08 PM Flag

    OT......$$$.... GLAD....Aren't we??

    Hey Everyone,

    I have off and on,(like many here perhaps, and many more in the investment community, I know), been pounding away for a simple(is it possible) algorithm to predict PPS movement. I have in the past used a deductive method going from general themes(TA) and applying the same to specific stocks, but have in recent months and years switched over to more of an inductive paradigm looking at specific stock movements to reason out a general theme which might be applied to many securities.

    My search over this time has brought me to the particular movement of PPS on stocks with a 10% yield or larger and more recently to monthly payers.. There is a large though limited universe of stocks meeting this criteria. I have explored junk bonds, mReits, MLPs and BDC's specifically during this investigative period.

    I am returning, once again, to an old paradigm I struggled with during the early 1990's, and am seeking your help in hammering out a system, which pays better than buy and hold, and which limits the risk of buy and hold over the risk of this preliminary "system" I will present.

    I know there are others out there who have looked at this before. As I get older, if there is one thing I know is true, it is that "there is nothing new under the sun", except cool cars and iPhone Aps...;-)

    That was the introduction. Dealing inductively I present one specific stock and will wager that this method and its outcome on this single security may be applied to all securities within the universe I have defined above, and through generalization to all securities.

    The stock I choose is the BDC GLAD. I have traded GLAD throughout the past 20 years and have found it typical of a good yielding(9.3%) stock with a range of 25-5 over its history. More recently, since the 2008 crash, it has traded between 8-12 dollars. Its current PPS is 9.21.

    Setting the trade:

    I want to capture .05 and .10 cents or roughly 0.5- 1% of the securities PPS on each trade, by going Long the shares for a one day hold when the following criteria are met. (There is a short scenario, but that's after the dust has settled...;-)).

    1)I look for a higher high and a higher low from the previous day and choose to enter the trade on the close of the day following these met criteria with the following caveats:

    A) The days high, low and approaching close must be within 2.5% of the previous day's high, low, and close. IOW, if yesterday's high was 9.25 and today's high was 9.50(2.7% higher) then the trade is aborted. Just the high and the low have to be higher than the previous day, but all three(high, low, and close) must be within the 2.5% window.

    B) Deep breath...the difference between the high and the approaching close of the day, divided by the difference between the high and the low of the day must be greater than 20%. IOW, we don't want to enter a Long position on the high of the day. If you have today's high at 9.30 with today's low @ 8.95 , with today's close approaching 9.00. Then take (9.30-9.00)/(9.30-8.95)= .85 which is greater than 20% so the trade is on, for this criterion.

    2) Never trade on the EX date. Use the following day after EX as the indicator day for the next day. IOW, your first day of trading, post EX, will never be earlier than EX + 2. The corollary, never trade on the DBEX(day before EX). So your last trade of the month(prior to EX) will never be after EX -2.

    3) Hold for .10/trade on stocks under 12.00 when IV(VIX) is above 16... Use .05/trade when VIX is less than 16 on stocks under 12.00.

    4) Exit the trade with your profit/loss on the close of the following trading day from entry.

    Quick example:

    So far this month Jan/Feb using .05/trade GLAD has made .13
    Last Dec/Jan(1 month) using .10/trade GLAD made .40.
    Last Nov/Dec(1 month) using .10/trade GLAD made .11
    Last Oct/Nov(1 month) using .10/trade GLAD made .21

    Each month had approx 4 trades. That is 4 days exposed to the market/month.

    By employing these different caveats above I have unfortunately made it a little more difficult for the trader to make quick calculations but these I have found are necessary to avoid the large draw downs which can turn profits into losses.

    I welcome other ideas, revisions, scrap and trash the whole idea, etc. I really believe with proper discipline to not quit in the face of the loss of missed gain (or suffered loss), that there are methods which exist, on a consistent basis to take small crumbs from the market, which eventually lead to great reward.

    Let's Get There,


    SortNewest  |  Oldest  |  Most Replied Expand all replies
    • lnhirst Feb 10, 2013 10:54 PM Flag

      Hi Doc. How ya be? Regarding the trading of GLAD, I noted that the average daily volume is

    • "Hey Mmichaelr, I could use your skills with working out the details on the GLAD thread. Any help would be much appreciated."

      I did not know what you were referring to__until now. For some reason until I cleaned out my DNS cache__I was not seeing this thread.

      I have felt you have been on to something with these date centric pattern trades. Also have felt there may be a procedure (developed) that could 'fine tune' entry/exit and possibly indicate whether the pattern will even repeat in the current cycle. I will give this some further thought (I am so busy right now preparing clients taxes).

      But I will give you some off the top of my head thoughts:

      I already like the use of percentage(s) {or even logarithmic} instead of targeting a historical day. That is why I asked you about price envelopes. Not hard plotted ones (like joining high and low peak/troughs with trend lines (channels)) but something conceptually like Bollinger bands (percentage/deviation) plots.

      One thing is what do you have available to test and run a possible program script. A platform such as 'Tradestation' would make life easier. However, even myself since, semi-retiring lost use of such platforms. I just am not willing to pay for the proprietary data feeds. I now use Ms-Excel and Access. Windows macros are basically Visual Basic and combined with Java alot can be automated.

      One difference I can already suggest is to think of the monthlies (as opposed to quarterly) having a relationship cycle similar to a front month futures contract and the commodities 'cash basis'. That means possibly a 28 day cycle (20 trading days). Since you are talking about record/ex-dates and payment date; I have not counted the cycle days. But in my thinking the monthlies have to be considered a different cycle length than quarterlies. This is only theoretically significant in that with commodity TA we use a shorter cycle to divisionally arrive at our moving averages and oscillator lengths along with wave cycles expectations (trough to peak and inverse). For example a 20 and 40 day MA is considered long and 5/10 day short.

      “look for a higher high and a higher low from the previous day”__Obviously that is the basic definition of an uptrend. The one problem with that is looking at a day to day range differential. The true range (Welles Wilder Jr.) is considered a more accurate representation because it accounts for gaps between days.

      True range:
      True high: The high or previous close, which ever is higher.
      True low: The low or previous close, whichever is lower.
      TR=maximum (high or prior close) – minimum (low or prior close)

      Example given the following series of days:
      Day 1 Hi 57.93 Lo 56.62 Close 57.57 TR=N/A
      Day 2 Hi 58.46 Lo 57.07 Close 57.67 TR=1.39 (same as Range)
      Day 3 Hi 57.76 Lo 56.44 Close 56.92 TR=1.32 (same as Range)
      Day 4 Hi 59.88 Lo 57.53 Close 58.47 TR=2.96 (59.88-56.92) (Range =2.35)

      Like I said just quick thoughts!

    • Re: your addendum on TWO Oct 25th

      I'm confused. I thought the increase from yesterday to today of high, low, and close, was limited to 2.5% for trading. That is, anything below 2.5% was traded, anything exceeding 2.5% was not. If I have this criteria incorrect, then it would make considerable changes in my results.

      I also wonder if our historical data used is different. This would explain some of the results differences. For instance, my data for Oct 1,3,4,5 produced gains of .18, .07,07,.04 respectively, and yours for the same days produced .10 for each of those days. I am getting my historical data from Yahoo and Nasdaq.

      I have re-run the entire TWO sheet again from 30 Oct 09 to 31 Jan 13 with the following results:
      Total trades = 191 with Net gain = $ 2.85
      Nr of trades with gains = 104 with total gains = $9.27
      Nr of trades with losses = 72 with total losses = -6.42
      Nr of BE trades = 15.

      Total dividends for the period = $5.05

      Unless I have my criteria wrong, TWO does not seem to be a good prospect. The net gain is only a little over half what the dividends would produce. If the program added a feature to limit trades with losses (could stops be used?) the net gain could be improved a good deal.

      Thanks again for sharing your experience and trading thoughts. It has helped my thinking and trading.

      • 1 Reply to jess_1554
      • Hey Jess,

        You are correct. I was incorrectly using 25% instead of 20% on Oct 25th, then I stated 2.5% in the bad..

        As far as the dates of Oct 1, 3, 4 and 5. I was holding for .10 cents from the close of each of these dates to the following trading day.

        Oct 1st closed @ 11.77, holding for .10 was hit on Oct 2nd.
        Oct 3rd closed @ 11.93 holding for .10 was hit on Oct 4th
        Oct 4th closed @ 12.00 holding for .10 was hit on Oct 5th
        Oct 5th closed @ 12.07 holding for .10 was missed by a penny(you are correct) and the gain was .04

        So upon each entry, you immediately enter a limit order for 10 cents above the close of the day. I think that was the piece that was missing, right?

        Try again using that information for TWO. Thank you for working the numbers Jess and keeping me honest...;-)


    • Hey Doc -

      I find this latest post of yours quite interesting. Seriously...........

      1) Usually you post your investing "tombs" late on a Sunday evening/Monday morning. Is this change in timing of any significance?

      2) You are obviously not satisfied with your strategy, since GLAD fits, as you have posted before, your style when you replied to my last "take a peek" post. Are you using strategy and style is in the same sense?

      3) I have been trading in the market since 1976, and my strategy has remained consistent with what I posted on this MB this past April. I rotate in and out of sectors. IF I was to deploy any constraints (which I believe is contained within the definition of your style, (if I am understanding you correctly) then that would be part of the execution of a strategy - agree???

      Remember "POEM" from management 101? Plan, organize, execute, monitor & control......

      Could you please describe the difference between your strategy and your style to me and perhaps others on this message board so that we can better understand what you are looking to improve upon.

      PS - I had a Super Bowl party at my home earlier, and I must admit I lost a large wager I had placed on the 49er's, since they didn't make the 3.5 point spread. :-(
      However, the real surprise was when one of my guests (whom I have known for 5 years now) tells me that he sits on the BOD of a company along with the owner of the Ravens. He was laughing (must have been the wine) as I was whining :-)

      • 3 Replies to saltynacho
      • never ever notice the player that retires usually wins the superbowl? Did you like that last call?
        Did you here the news on European mach fixing today? What makes you think US sports is any different ? I made the right (bets).

      • Hey Salty,

        Sorry about your team losing. Maybe you can have your accountant expense the cost out under a 'promotional' line item, as you have the proof since the beneficiary you claimed was laughing.

        Regarding the wine's effects:

        "You are obviously not satisfied with your strategy, since GLAD fits, as you have posted before, your style when you replied to my last "take a peek" post. Are you using strategy and style is in the same sense?"

        A little slurring with the intransitive verbs there Salty. Poorly researched references also which historically falls out of your purview. You keep such good notes as a rule....;-)

        Let me help before the wine wears off:

        "IF I was to deploy any constraints (which I believe is contained within the definition of your style, (if I am understanding you correctly) then that would be part of the execution of a strategy - agree???"

        Here is a rare opportunity to answer your own question Salty. I give you, none other than, your own thoughts on this very topic:

        ""I too believe that KMP will experience a pull back........... however, I would much prefer to sit back and wait to see it (and the magnitude, as well), rather than betting on the come.
        Just my style.""

        So I ask you to answer your own question Salty, ""IF I was to deploy any constraints .... then that would be part of the execution of a strategy - agree???.......Are you using strategy and style is(sic) in the same sense?"""

        I hate to take advantage of you like that to me tomorrow after the "party" wears off...;-)



      • correction....... "tomes" (must be the wine) :-)

    • I will follow your experience with the concept with considerable interest, and will attempt to add to your data pool by backtesting other stocks which fit the criteria. A steady diet of crumbs sounds better than hoping for occasional windfalls.

      • 1 Reply to jess_1554
      • I ran TWO from October 2009 to present on a daily basis using your criteria, with the following results:

        129 total instances with a sum total of $11.2. The annual instance count was 2009=0, 2010=32, 2011= 55, 2013 = 7 so far.

        If the 3 months of 2009 are ignored, this works out to 3.5 instances per month, although the distribution was not even across time. However, the frequency is increasing during the past year or so. The majority of instances were for $0.10 (95 of 129), but nearly all of the $0.05 occurrences were during the last five months. I will look at the frequency and distribution further.

        The average PPS during the total period was $9.96. So $9,960 for 1000 shares (used for a total of 129 days, average 42 days/yr) would have produced $11,200, or an average of $302 per month with zero losses.

        I think you may be on something here, Doc. I will look further.

    • adios doc ?

21.41-0.17(-0.79%)Mar 30 4:00 PMEDT