Smoothed CI-NCI Ratio and the Coppock Curve
In my previous communication I pointed out a weakness in our using Peter Eliades' daily CI-NCI Ratio as a leading indicator of deep and very deep Monthly DJIA Coppock Curve buy signals; because that Ratio often dips to 0.9400 multiple times prior to a buy signal, we can never know at the time if any particular dip is the last one.
To overcome this weakness I have substituted, into a revised and updated version of my previous communication, a smoothed CI-NCI Ratio obtained by applying six ten-day moving averages. Here is the result:
The Monthly DJIA Coppock Curve has dropped below -125 into what I call "deep territory" 13 times since 1926, with the 13th occurrence coming on 10/31/02. All 13 drops were necessarily followed sooner or later by a deep or very deep Coppock buy signal, with the latest signal, a deep one, arriving on 5/30/03. I define a deep signal as a turn up of at least 10 points from -125 or lower, while a very deep signal is a turn up of at least 10 points from -250 or lower.
It so happens that 10 of those 13 Coppock drops were associated with a solitary instance of the smoothed CI-NCI Ratio dipping to 0.9500 within about two years (or 600 trading days) before the eventual deep or very deep Coppock buy signal. In the 1978 and 2003 cases there were no dips at all; while in 1973-75 case there were two dips. In the 11 cases where there was at least one dip, the date the smoothed Ratio last reached 0.9500 to the downside preceded a deep or very deep Coppock buy signal by 557 trading days (8/31/32 very deep signal), 250 trading days (7/30/38 very deep signal), 312 trading days (6/30/41 deep signal), 135 trading days (9/30/42 deep signal), 187 trading days (8/29/47 deep signal), 145 trading days (1/31/63 deep signal), 167 trading days (4/28/67 deep signal), 273 trading days (8/31/70 deep signal), 151 trading days (1/31/75 very deep signal), 164 trading days (9/30/82 deep signal), and 208 trading days (10/31/88 deep signal).
To summarize: Except for the 1978 and 2003 cases when it did not dip to 0.9500 at all, and except for the 1973-75 case when it dipped to 0.9500 twice, the smoothed CI-NCI Ratio dipped to 0.9500 only once during the 600 trading days prior to each of the other 10 deep or very deep Coppock buy signals; and the lags from (last) dip to signal were between 135 and 557 trading days.
What lies ahead? If the Monthly DJIA Coppock Curve rises out of deep territory in the coming months, the 5/30/03 deep buy signal (which in my last communication I considered unlikely to arrive because of the complete lack of any warning dips to 0.9400 by the daily CI-NCI Ratio) will no doubt be a winner to some extent. If, however, the Coppock Curve turns down again before leaving deep territory, the deep buy signal of 5/30/03 will be voided, and I will treat it as if it never happened.