The result, I'm finding, is that so many signals create so much noise as to be next to useless for generating actionable information.
So I'm modifying my methods to create greater structure.
This year I've used two short-term signals, the Heikin-Ashi trend and the fast stochastic. I'm ditching the fast stochastic -- it is just too volatile for my slower trading style. I'm retaining the Heikin-Ashi trend, because it provides a corrective against entering a position with a contrarian short-term trend.
I've used three mid-term signals, Person's Proprietary Signal (pps), the parabolic sar (psar) and the macd (moving average convergence-divergence). These three pretty much move in lockstep. The pps is generally (but not always) first out of the starting gate, followed by the psar and then then macd.
I'll continue to track all three, but with a difference. A trading system can work nicely with only one of those three signals. An alternative approach is to treat them as a consensus-seeking mechanism, where two votes counts as a signal, or where unanimity is required. My initial method will be to use the pps only as my signal.
I've also used two moving averages as confirmations that a proposed position is on the right side of history: The 20-day moving average (essentially, a month), and the 200-day moving average (essentially, a year). I'll retain those two moving averages, and add the 50-day moving average, essentially, a quarter. All are simple moving averages rather than exponential.
However, they won't be used to determine whether or a not a trade should be taken. They are, rather, components in stock selection.
One missing link in my method has been volume. I've always looked at it, giving preference to volume spikes for bullish entry, but it hasn't been a systematic part of my method. I want to remedy that.
My choice is the venerable on-balance volume (obv) technical tool. It is a running total that adds the volume of a period when the price rises and subtracts the volume for a period when the price declines. The rise and decline is determined by the current period's close compared to the prior period's close. To read more, click here for the Stockcharts.com explainer.
The obv, like many line indicators, bounces around a lot. So by a rising obv, I mean that I can draw an ascending trendline on the chart. An obv making higher highs and higher lows is uptrending; lower highs and lower lows, downtrending; and all over the map, neutral (and non-confirming).
Putting it all together, my new method will be this:
1) Upon a signal from the pps, look for confirmation from the obv and the Heikin-Ashi trend.
2) If the Heikin-Ashi confirms the signal, then look at the obv.
3) If the obv trendline is up (higher highs, higher lows), then enter the trade. If down (lower highs, lower lows), then pass on the trade or defer entry until the obv trend confirms the signal.
4) Exit upon reversal of the signal.
A note on stock selection:
I look for growth stocks as vehicles for bull plays. There are a number of factors that I look at. Two of the most important are a high return on equity, 20% or greater, and a low debt to equity ratio, generally below 0.1.
I also look at where the stock stands in relation to the 20-, 50- and 200-day moving averages (ma20, ma50, ma200). Positioning above the ma20 is a minimum condition for adding a stock to my bullish watchlist. Positioning above the ma50 and the ma200 strengthens the bullish case.
So, the new analysis bar will have these components, divided into sections:
The Signal Section: The pps, psar and macd.
The Confirmation Section: The obv.
The environment section: The Heikin-Ashi trend, the ma20, the ma50 and the ma200.
Here's an example, using the S&P 500 exchange traded fund SPY, showing the old style and the new style.
Here's are the old style and new style tables for the 20-year-plus Treasury bonds exchange-traded fund TLT:
The new style will mean fewer signals for the headlines that lead each Morningline, but that's OK. Fewer signals means less confusion.
Also, the on-balance volume trend is a bit subjective. I mean, on TLT at present there's a higher high, a lower low, a lower high, and a downward move that is not yet a lower low. How do you code that? I resolved it as a downtrend by going back to the late September lower high, but it arguably could be coded as neutral (yellow on the charts).
My Forex analysis, using only the Person's signals, will remain unchanged for now.