Click here to Email this article
The NYSE McClellan Summation Index ($NYSI) is a derivative indicator from the McClellan Oscillator ($NYMO). Read this article first and this second before continuing. In these two articles, I explained that the MO itself is a derivative of the Advance/Decline (A/D) line, where the A/D line is a derivative of the raw data of daily Advances and Declines that can be found daily in most financial publications. So ultimately The McClellan Summation Index (SI) is derived from the number of advancing and declining issues in a given market. It’s not based on prices as most indicators are but based on advancing and declining issues of an entire exchange.
The McClellan Summation Index (SI) is obtained by summing up the daily values of the McClellan Oscillator (MO) resulting in daily accumulation “cumulative total” or "summation" of the MO daily value. We get it by adding the daily (MO) value (positive or negative) to the prior day's (SI) to get today's (SI) value.
Yesterday's SI Value + Today's MO value = Today's SI value
The Ratio adjusted NYSE McClellan Summation Index - Enlarge
A negative (MO) produces a lower SI. A positive (MO) produces a higher SI. The SI changes direction whenever the raw oscillator (MO) crosses above or below its own zero line. In other words, the SI rises when the MO oscillator is above its zero line and falls when the oscillator is negative below its zero line. The result is plotted as a slow-moving curve that oscillates in relation to a zero line (fluctuates above and below zero). Unlike the MO, its fluctuations are smoother and its zero line crossovers are less frequent.
Whereas the MO is used for short to intermediate trading purposes, the SI provides a longer-range view of market breadth and is used to spot major market turning points as well as projected long-term trends. The SI often gives more reliable long-term clues to market breadth than the traditional advance-decline (A/D) line or the MO. It’s a popular indicator with good track record that can be used on data as back as to 1920s. It can signals market turns and regarded by many* as an excellent indicator of the overall "health" of the market and the market's current trend.
Traditional versus "ratio-adjusted" SI:
The "traditional" calculation of breadth indicators, such as the MO and SI, begins with subtracting declines from advances, then continuing with the Oscillator/Summation calculations. As stocks have been added to the NYSE over the years, the MO and SI have become skewed to higher and higher readings and reaching greater high and low extremes over the years, but these new record highs and lows didn't accurately reflect the true internal condition of the market and making historical comparisons impossible. To remedy this, a new Ratio-Adjusted calculation** was adapted. The resulted chart differ greatly than the traditional one and produce different readings. You should be careful which one you're using or you'll get incorrect interpretation.
From this chart (click here), you can see that the traditional SI is blasting to historical highs. On the other hand, the ratio version, while obviously higher than normal, is still relatively unimpressive in the 76-year historical perspective (until 2002). The highest and lowest historical readings are still associated with the 1929-32 Bear Market.
Another example, let's compare the 1968-70 and 1998 Bear Markets, representing declines of about 36% and 20% respectively. Even though the 1968-70 decline was much more severe, the McClellan SI (traditional calculation) were -2252 for 1970 and -3526 for 1998, giving the false impression that the 1998 Bear Market was worse than 1968-70. Using a ratio-adjusted calculation the SI reading for the 1970 low was -2402 compared to the 1998 low of -1444, clearly reflecting the proportional relationships of the two bear markets, and making it possible to make an accurate historical comparison.1
What are the numbers to watch for?
The traditional SI*** normal range of movement fluctuates between Zero and +2000 (see this chart) although it can and does move outside of this range during extreme or unusual market conditions. Historical extremes are approximately +4000 and –2000 though in 2003 we broke the +6000 level and on March 2009, we broke the floor to hit below –4000. We came back from these extreme low levels to again break UP to above the +5000 level (another historical extreme) last month at an impressive total of 9000 points move from bottom to top. As of Sep. 11, 2009 we stand @ +4751.37 on the traditional SI.
On the other hand, ratio adjusted SI normal range of movement is between –1000 and +1000 (see chart below) although it can and does move outside of this range during extreme or unusual market conditions. Historical extremes are approximately +1500 and –1500 with exception to the great depression era when the extreme was +2000 and -2500.
The Ratio adjusted NYSE McClellan Summation Index - Enlarge
Why the SI is an important long-term market indicator?
The SI offers many different pieces of information in order to interpret the market’s action. One should not just pay attention to the numerical value of the SI, because understanding its chart structure offers much more insight and accurate analysis. In this article I will cover the interpretation of the “Ratio adjusted” SI as followed:
1. The direction of the SI movement, up or down is an indication of whether money is moving in or out of the market. A rising SI is positive for the market and a falling SI is not. A market is termed “neutral” at the zero level. When the SI moves below the Zero Line and stays there, it is a very negative sign and indicates that a possible long-term down trend is starting or underway and likely to become more severe putting the market into bearish status and territory. Movement above +1000 is considered a positive sign that puts the market into bullish status and territory.
Historical "Ratio adjusted" NYSE McClellan Summation Index (SI)
Courtesy of decisionpoint.com - Click chart to Enlarge
2. The true significance of the SI comes into play when the reading is outside normal range of -1000 and +1000 indicating an unusual situation in the stock market. Unlike many articles on the Internet that simplify the SI in numerical value that is not only wrong but also over simplification, I would like you to know that there are truly no clear "oversold" or "overbought" value or levels that can be used to forecast market direction or condition with great accuracy from this indicator. There are instances where the SI "oversold" bottoms have formed and were followed by continued selling in the market and lower SI lows.****
3. Top and bottom signals carry more significance if the index is also diverging from the associated market average or composite.
4. Historically, significant major market bottoms occur and could be anticipated after the index falls below -1000. initial first buy signals occur when a positive divergence forms from bearish territory readings creating a bottom formation and followed with a sharp, swift, and smooth upward from that bearish territory to bullish territory (above +1000). A second confirmed buy will occur with the indicator’s ability to hold or regain bullish grounds during normal market retracement after the initial move from the bottom.
5. Major initial sell signals occur when a negative divergence forms from bullish territory (above zero) creating a top formation that might take many weeks and months, which followed by a decline into bearish territory (below zero). A confirmed sell will occur with the indicator’s inability to recover bullish grounds again when the market turn back up (sometimes even to new highs).
7. The slope of the summation curve is determined by the difference between the actual reading and the zero line as I explained above. So, an "extreme bullish reading" will cause the SI to rise sharply, and vice versa which can results in a lot of whipsaws. It’s better to use an MA crosses over. This is often less timely, but it filters out a significant number of false signals. A suggested time frame for this is either 35-day simple MA, or my favorite the 21-day EMA (plotted on the chart above) and even there, you will find few number of whipsaw signals, indicating that this approach is a mechanism to reduce but not to eliminate whipsaws. This indicator renders a bullish signal when it crosses above a 21-period moving average and a bearish signal when it crosses beneath it. The best bearish “signals” occur if the cross occurs under market overbought conditions and vice versa.
As of today and from the chart above we can see that:
1. SI gave a long-term "bullish call" see above #4 and chart.2. SI is rising indicating that money is going into the market.
3. SI is above +1000 indicating a long term bullish status.
4. SI had no significant divergence from the NYSE Composite.
5. SI 21-day EMA is above the SI rendering a bullish signal.
Until next article and as always, I wish you a profitable trading and a wealthy life.
Please rate this article at the top of the page. Thanks!
GoodVibe
Mr. Lucky
Footnotes:
*Many smart people say they won’t bet against
the SI long-term signals. I don’t know about that! I never
put my faith or trade based on any one indicator. Long or short-term
signal by any indicator do fail. Indicators should never be considered right or wrong. They need to be
analyzed in context with proper interpretation. Regardless of the
indicator, you must use several correct periods to establish
context. Whenever possible, you should not make generalizations about the current behavior of market indicators without first observing their behavior over an extended period of time and in a wide range of market conditions. For example, a sharp rise by the SI will signal that some up
cycle is underway. So the absolute direction you can forecast from an
indicator like the SI directly depends on the cycle that you study and
the direction of the broader trend.
**More about calculation - Reference 1.2.3
Please read more from here in the footnotes. Also check the original work of the SI inventors Sherman and Marian McClellan in their book "Patterns for Profit" found here.
The Ratio Adjusted version adds a step to the calculation process, dividing the A-D difference by Advances plus Declines, and then multiplying the result by 1000 to get it back into the realm of real numbers. In effect, we are mathematically pretending that there are always exactly 1000 stocks trading. As such, it is better for multi-year comparisons.
There are two methods for calculating the Summation Index. The first method (the one originally used by the McClellans) simply maintains a running total of the values of the McClellan Oscillator. The second method uses the following formula:
Summation Index
=
1000 + (10%Trend - 5%Trend) - [(10 x 10%Trend) + (20 x 5%Trend)]
Where:
5%Trend = 39-day EMA of (Advancers-Decliners)
10%Trend = 19-day EMA of (Advancers-Decliners)*McClellan Summation Index (New Method):
Mathematician James Miekka developed the new formula for calculating the daily SI that prevents drift and forces it to maintain a consistent relationship with the Zero Line. Rather than adding the current MO value to the prior day's SI (the traditional method, which causes the undesired drift), the Miekka formula derives the Summation value directly from the daily 5% and 10% index readings. This not only stabilizes the SI, it also allows us to calculate the SI for any day without knowing what the prior day's reading was. Also important, the Miekka method insures that independent calculations will always be within a few points of one another -- differences are often caused by rounding and variances in advance-decline data.
The formula is:
***According to the McClellans, the beginning of a new bull market is signaled if the NYSE-based SI first moves below the -1200 levels and then quickly rises +2500 points from its prior low. In the traditional SI, the McClellans addition, relatively sharp movements, greater than 3600 points from overbought to oversold extremes, indicate the beginning of a major move.
****Overbought and oversold readings may vary among indices and historical precedent. Always gather as many indicators to support your analysis without disregarding others that don’t.
****On May 2009, SI crossed over +1000 and barely touched above +1200 in what so called "overbought condition". Looking back to 1928 to see what happens after SI crosses above 1200, out of all the instances when the SI was above 1200 only two lead to any declines of substance during a bear market.
For Readings above 1200, from 27 separate times the SI went below zero and above 1200, you find only two times that there were anything over 10% declines and they were in 1940 & 1941. A decline in 1940 started one month after the reading and lasted 6 months and was a decline in the 16%-17% range while in 1941 it was coincident with the top and lasted 7 months and was around 28%. That will lead us to believe (if history to repeat itself), any decline that materializes from a trading top has little risk of turning into anything more than a normal retracement. Ref.
Same happened in June-July 2009 when SI pulled back to +400 from +1200 and the market retraced from 956 to 869 to launch a bullish offensive that took S&P back to 956, crossed the S&P 1000 mark, and took the SI to +1393 higher than the previous 1200 levels.
GV, Thank you again for this excellent article! The most bullish case I've heard lately. People poring money in the market, confirmed long term bull... and then a probable cross of the 21 EMA if market turns down for two days. Honestly, this analysis is the one that forces you to see both sides.
Posted by: InterSat | September 15, 2009 at 12:12 AM
InterSat - There's only one thing missing yet to "confirm" the so called new bull market, which I don't subscribe to yet. The next retracement (if and when) MUST hold the SI above zero and if it visits sub zero, it MUST stay there for no more than a cup of coffee and get back up quickly to above zero for lunch and dinner. :) No way around it.
Posted by: GoodVibe | September 15, 2009 at 06:40 AM