Tuesday, April 26, 2011

4/26 Addendum

This postscript regards cycles.  As you may already know I have spent a lot of time the past two weeks improving my existing cycle optimizer and running my machines into the ground generating new cycle baselines for the markets that I trade.  The optimization is still running but is complete for SPX, DJI, ND, bonds, and dollar.  There are several points regarding cycles:
1) Cycles seem powerful, but the optimization process makes all past occurrences look good.  Optimization is a "cheat", in that while it twists the probabilities in our favor going forward over the next year or two, eventually the errors inherent in the cycles themselves make the tool less valuable.
2) The cycle optimizer weights certain cycles according to probability but I have found that cycles that appear in the future may or may not manifest themselves proportionally with price movement.  Thus, a cycle that is relatively shallow but extends for 6 months may result in a price movement out of proportion to the cycle(s) amplitudes.
3) Cycles do not rest upon any physical causation but are merely statistical correlations with price.  This makes them less reliable than I1 or the other sentiment gauges that utilize real world metrics that are not price-related. 
However, used in conjunction with the sentiment gauges, when there is a confluence between sentiment and cycles it reinforces my conviction.
Here is the finalized DJ cycle chart, which does show a meaningful cyclic decline starting now:


  1. Steve,
    We finally have our new highs that my counts required in April or May. I am trimming longs and looking for an exit tomorrow. Everything is in sync: S&P, small/mid caps, international, metals. There is really only one thing that I could see that could cause this: The Fed ends QE2 or raises interest rates. It's that or something crazy unexpected that rocks the markets. We need to remember this is one year from the flash crash set up. The thing about my counts is that they don't complete until end of May requiring a May close on S&P above 1286, if i remember correctly (I can check tomorrow to be sure.) So, it may not necessarily be a waterfall yet. We did break out of the neckline of the inverse H&S. Typically after the breakout, you retest the neckline.

  2. Hi Charles,
    Do we need a news catalyst? The stock market hates cost push inflation, especially with no demand pull on the other side. Which H&S do you refer to?

  3. Steve, I think Charles is referring to the Inverse H&S on the S&P (pattern starting around 2/18) with left/right shoulders at 1300 and recently broken neckline around 1340. I think it will end up being a "fakeout" and instead reslove down, but we'll see.
    Dan B

  4. Head and Shoulders form after extended runs, so a H&S bottom should only be valid after an extended downtrend.

  5. Charles, do you have any details on your methods?