This post is a response to two questions raised by a reader. The original article can be found here.
How would the cycle dates and outcomes have changed if 3/30/04 had occurred in the middle of a cycle, and did not represent the true beginning of the first cycle?
Related questions include: A. If the true starting date for cycle 1 had been earlier than 3/30/04, how would the end date of cycle 1 be affected, and how would that affect the timing and the outcomes of all subsequent cycles? B. Are the timing and the outcomes of all cycles subsequent to cycle 1 an artifact of the selected start date?
I will analyze the question by examining four different scenarios. Each scenario assumes that the "true" starting date of cycle 1 preceded 3/30/04.
Scenario 1: The closing price of XIV on 3/30/04 was identical to the earlier purchase price. In scenario 1, the analysis and results would be identical to that of the original article.
Scenario 2: The trade was profitable as of the 3/30/04 close. In scenario 2, the upper sell limit would have been lower than that of the published model for cycle 1 only. As a result, the cycle would have been completed with a gain of ~133% prior to 11/10/04. However, because the next buy trigger would not have occurred prior to 4/19/05, the dates and outcomes for all of the nine subsequent cycles would have been unchanged.
Scenario 3: The trade was not profitable as of the 3/30/04 close, and the running loss was not worse than -30.0% on that date. In scenario 3, the closing trade for cycle 1 would have been later than 11/10/04 but not later than 4/10/05, with a gain of ~133%. Once again, because the subsequent buy trigger would not have occurred prior to 4/19/05, the dates and outcomes for all of the subsequent cycles would have been unchanged.
Scenarios 1-3 all have a different end date, depending on the "true" opening price. However, the result for cycle 1 would have consistently been a gain of ~133%. In all three scenarios, all of the subsequent cycles and outcomes would have been identical to those described in the article.
Scenario 4. The trade was not profitable as of the 3/30/04 close, and the running loss was between -30.3% and -34.7% on that date. [If the running loss had been worse than -34.7%, the previous cycle would have closed via a stop loss on or before 3/26/04, and 3/30/04 would have started a new cycle.]
In scenario 4, the first cycle would have been completed with a gain of ~133% between 6/14/05 and 7/14/05. This would have been followed by a trailing-stop cycle, and then three consecutive losing cycles ending in stop losses on 8/3/07, 11/9/07, and 10/7/08.
The next buy trigger following 10/7/08 would have occurred on 12/11/08, converging with cycle 4 in the article and in scenarios 1-3. Scenario 4 would have suffered large losses from 7/14/05 through 10/7/08. However, the timing and outcomes of cycles 4-10 would all have been identical, independent of the scenario.
The Likelihood that Scenario Four Represents the "True" Situation. In the absence of historical data prior to 3/26/2004 one cannot assert with absolute certainty that Scenario 4 is incorrect. Recall the assumptions: the position had to have been down between -30.3% and -34.7% on the 3/30/04 close.
Figures 1 and 2 show historical closing prices of the S&P 500 index and of the spot VIX. Dates range from 9/30/2003 through 3/30/2004, the six month period ending on 3/30/04. In both figures, 3/30/2004 is represented by a large black dot. The data for these plots were downloaded from here and here.
Figure 1. S&P 500 closing prices for the six months leading up to 3/30/04.
Figure 2. Spot VIX closing prices for the six months leading up to 3/30/04.
Inspection of the two figures reveals that the VIX was falling and the S&P 500 index was rising on 3/30/04. One would hypothesize that the original buy signal should have occurred earlier in March when the S&P was near its low and the VIX was near its high. If this hypothesis is correct, scenarios 1-3 would all be possible, but scenario 4 would be very unlikely.
Summary Answer to Question 1
Scenarios 1-4 all converge to the identical set of cycle dates and outcomes, beginning with cycle 4 on 12/11/08. Scenarios 1-3 all converge much earlier at the beginning of cycle 2 on 4/19/05. Convergence of all four scenarios on 4/11/08 suggests that cycle 4 and all ongoing cycles are independent of the choice of the starting date.
Even though scenario 4 is improbable, the steep losses that would have hypothetically been suffered in scenario 4 from July, 2005 - October, 2008 should serve as a solemn warning! There is no guarantee that a similar stretch will not recur in the future; and there is no guarantee that the model will "magically" work every time to save an investor from steep losses if/when this were to happen. Whenever an investor evaluates the economic and geopolitical situation as being "scary," my recommendation would be to sit it out on the sidelines, and to postpone opening a new position until the situation has resolved itself.
How would the results be affected if a position were opened on a "buy signal" that occurred mid-cycle, instead of at the beginning of the cycle?
Response to Question 2.
Markets do run in cycles, although the details are not so easy to predict. Models attempt to predict the beginnings and ends of cycles, but are imperfect in their attempts to do so.
The personal choice that an individual investor makes regarding when to open any position does not affect where the market is in its cycle, nor does such a decision affect where the market is going.
Therefore, if a person were to open a new position in XIV mid-cycle, they would be advised to adjust their expectations, based on where the market cycle actually is. The outcomes cannot be expected to be as good as if they had opened the position at the beginning of the cycle.
Furthermore, if a person were to open a new position when the cycle indicates that they should be sitting on the sidelines, they should do so with the understanding that their trade is speculative and could result in losses. This would have been true of new positions in XIV that were opened in the time frame from 3/5/07 to 9/29/08, following the strategy's trailing loss sell in cycle 3.
Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
Additional disclosure: A new position in SVXY could be opened if ongoing market conditions were to trigger a "buy signal," as defined in the article.