New short selection

Explanation of the model -- the long sector selection is (and always will be) the primary sector for the rotation model. It is not designed with market timing as much as simple asset allocation, and the best results are geared for a person to hold a long position and rotate off with two quick market orders when a new sector replaces the old one.

If one were holding XLF (for instance) and the model gave a new call for XLV, on a typical day market gaps will average out. Whether the market gapped up or down, these should (on average) gap in a similar fashion. On any given day there will always be differences, but on average an investor should be able to market order out of one and turn around ten seconds later to market order into the next. Limit orders are always better if you can hang around to monitor them, but the rotation model is geared for a buy and hold retail investor who wants to make a quick order and then forget about the market for the rest of the day (or week).

The actual selections are NOT made based on any sector in isolation, but instead on the RELATIONSHIP of sectors. It is not designed to show which sector is hot NOW, but which should be hot NEXT. In other words, if Johnny is eating Thanksgiving turkey NOW, he'll likely eat grandma's poundcake NEXT. XLE, for instance, often bleeds into XLP.

These sectors also conform to those identified by John Murphy on stockcharts.com as indicative of the market cycle, and I will add an occasional comment about what the price and volume movements seem to indicate about our position in the cycle. I will also add comments from two other models and whether they confirm Murphy's sector theory:

1) Elder Force: This is a very primitive cousin of Pascal's effective volume and is geared toward a bigger picture of where we might be. The limitations of this model are the fact that it is so over represented in black box algorithms that it's usefulness is compromised for a retail investor. At this point I like to think of it as a useful comparison of what USED TO work (and what retail investors are probably thinking). If Pascal's model gives a picture of institutions vs. retail investors, this model gives a picture of the herd's expectations.

2) Yield Ratio: This is a model based on the relationship of Fed manipulation of interest rates and actual market action. The model doesn't show what SHOULD work as a result of the interest rates changes, but instead what the Fed THINKS. That is, if the Fed is getting panicked, the short term rate will be driven down (as in the last few months from .15 to .02). If the market prices rise immediately, no problem. But if the market prices do not rise enough in response to this stimulus, or if they in fact fall, the model views this as a failure that's potentially stronger than the Fed's ability to handle. As early as the summer of 2007 this model showed a looming bear market, but the model was late in 2000 and missed the top by about 10%. In other words, it's not perfect (because the people in the Fed aren't perfect). And the limitations of this model are compounded by the fact that we are in extraordinary levels of manipulation that I have no historical measurements to compare. This model worked in the past, but present conditions are so outside of normal limits that I have to take the model with a grain of salt.

Most of my models are based on long term views, and are not nearly as nimble as the Robot. But I've primarily designed them for the types of investor who can't watch the market all the time (which isn't really this group).

So, then, please take what I share as long term views -- but take your market calls from the Robot. I'll be doing the same myself.

Sector selection:

The model is primarily geared for a long only investor. A hedged investor will underperform a long only investor but SIGNIFICANTLY eliminate volatility. Most people on this board will be interested in the Robot timing option (number 3).

Personally, I will keep a portion of my portfolio reserved for option 1, but have another portion of my portfolio devoted to Robot timing.

1) Always long: XLU
2) Always hedged: long XLU, short XLP
3) Robot timing: short XLF at 16.09, with stop at 16.46

Notes on business cycle:

Elder Force -- the Elder Force model is stopped out but has not yet reversed into a long position. Current settings are...

1340.68 current S&P

Short term:
1351.56 high
1337.70 pivot (restart short)
1319.77 low

Long term:
1342.31 (reverse to long)
1256.79 (previous target)
1171.28

We continue to be in a whipsaw formation as some sectors shift up and some shift down. Typical investors in this type of pattern end up losing money on both long and short trades as each get stopped out and reversed.

Yield Ratio -- the yield ratio model is giving a rather wicked bearish signal. I have no idea if this is a side effect of QE2 or something real. We are WAY outside of the parameters this model was designed for, but the signal is giving me a great deal of caution. The model's reward to risk ratio is 0/20 (no reward compared to risk).

Sector formation -- per John Murphy and stockcharts.com, the following sector ETFs outperform at different stages of the market cycle:

8 XLY Bottom
5 XLK Bull 1
9 XLI Bull 2
4 XLB Bull 3
2 XLE Top
1 XLP Bear 1
3 XLV Bear 2
6 XLU Bear 3
7 XLF Bear 4

I prefaced these with the rank of probability my model gives for our placement in the business cycle. Although sometimes indicative of minor tops (like a correction), the current formation shows the most likely position to be the beginning of a bear, and the second most likely to be a top.

Again, this could merely show some weakness and fear among investors. Nothing is a crystal ball.

And NO ONE can predict what rabbit Bernanke might try to pull out next. If anyone says they know exactly, run.

Tim

For 5/20/2011:

If the Robot is long, XLU.
If the Robot is short, XLF.

Also, after reviewing the Robot and my own sector history in more detail, it would be a mistake attempting to set the same price targets. I would recommend either placing a limit order at yesterday's closing price on a selected sector, or trying to enter the trade when IWM goes into position (some kind of conditional order).

But the volume disparities seem to constrain the sectors from making the leap into the sweet spot that the Robot is targeting for IWM.

I was making it more complicated than necessary. The sectors are ALREADY primed to spring when the Robot generates a signal.

Tim