squark62's  Instablog

WARNING!!!...WARNING!!!!

the SDS-SSO pairs trade model is topping out all kinds of parameter highs and some on the verge of being topped out.  if we continue on this trajectory unabated, we're likely to become un-tethered from reality and we'll be "lost in space."

here are some charts.  the chart below entitled chart_geodesic_SDS-SSO_20090724 is a snapshot of where equilibrium is along the geodesic.  the different colors represent different time series of data.  the purple is the path the parameter coordinates carved out from about mar-9-09 to 6-15-09.  march 9 started a the top right most purple dot, traced down to the apex of the curve and down symmetrically along the lower portion of the curve.  june 15 is the last lower rightmost purple dot.  then the blue dots represent the pull-back after june 15 to about july 15 which is masked by the leftmost yellow dot.  that's when the market rallied and re-traced in yellow dots all the way back to where june 15 was in less than 1/2 the time the pull-back took.  the spread of the points shows just how strong the upward motion in the S&P 500 index has been in the last 10 days.



the chart below entitled chart_geodesic_parameters_rates_of_changes_20090724 provides the indicator when the index has changed direction that is from bullish to bearish & bearish to bullish.  the crossovers at the x-axis are the indicators to change market sentiment.  judgment and other techniques can tell you about the same thing.  the trouble with most any system is being able to predict to a high level of certainty what the market will do next.  but crossovers are conservative and are generally correct.  another problem here is if the parameters cross each other, wiggle and cross each other again in short periods of time.  this happened recently and back in january & february for certain.  so crossovers can also be considered as no-man's land and the closer the parameters come to crossing each other, the less certainty of knowing short-term future market direction.  things to be aware of as with any models boundaries.



the chart below entitled scatter_plot_diff-theta_vs_diff-c_20090724 is a scatter plot (just like the geodesic plot is a scatter plot of it's parameters) of the daily rate of change of the geodesic parameters data plotted w.r.t. each others.  it's the same data above plotted independent of time.  the color notation time-series-wise is the exact same as in the geodesic chart.  the yellow dots started in the 2nd quadrant and followed a line basically through zero and are now extended way out in the 4th quadrant.  besides the apparently linearity of the data, notice how much farther the yellow data extended along the x & y axis versus the farthest point of blue data indicating how much leverage has been put in the market.



the final chart for this commentary entitled chart_vector-c_rates_of_change_20090724 is a time series plot of diff c defined as daily rate of change of geodesic parameter vector-c and it's daily rate of change defined as diff diif c.  theta is the other geodesic parameter that has been scaled down for relative comparison.  the data shows the weights of SDS & SSO are changing at there fastest rate going all the way back to late august-08.  this rally has the october-november crash beat in magnitude and the march crash.  note that highs of  this nature have been followed by steep reversals.  also note all the reverberation during october-november-08 a better part of 2 months to settle down.  the february-march-09 event was less noisy but took about 2 months to settle down.  since around early may-09, almost 3 months ago, events have been very erratic and building up intensity from peak to peak.  quite a perilous time for buy & hold investors in my opinion.



disclosures: i am currently 65% SSO & 35% SDS in all my pair trading accounts.  will start selling SSO gradually from here.