Option Investor
Trader's Corner

Predetermined: The Logic of Corrections

Printer friendly version

Corrections in the market tend to be of a certain amount often enough to think that market cycles spin out into the future and tend to follow certain 'predetermined' patterns. I know that the choice of such a sort of loaded word as "predetermined" raises considerable doubt with many if not most investors and traders. How can market trends be predicted: we don't know what the Fed is going to do, we don't know what OPEC is going to do, we don't know the future period!

We keep our minds busy with events, government policy, economic and earnings trends and the like and mostly, if we think of the bigger picture, tend to consider CHANGES in market direction as captive to unpredictable events and perhaps 'random' events. Heard of the 'random walk' theory of the market movement? Basically, one can't predict the market, so there's no point in trying. The gist of the random theory is that the HOLDERS of this idea CAN'T predict overall future market trends; e.g., economists, most investors and so on. Very wealthy traders, successful over many years and market cycles, seem to fly above and beyond that rule.

I'm not going to go much further in discussing this topic except to show and remind you of how corrections tend to occur in certain predetermined ways or are associated with certain conditions. There are some key ideas that are useful for making profitable TRADING decisions; otherwise, I wouldn't bother writing about how certain patterns repeat over and over.


(Any one of these key ideas should not be taken or used as the sole source of buying or selling decisions. Rather, it is often the interplay of these key factors that are telling.)

1.) Markets don't top out or bottom until traders and investors get quite or 'extremely' bullish or bearish. HOW you could or should measure those bullish or bearish extremes is unknowable to most market participants; how can you objectively measure this?

2.) At significant tops or bottoms the major stock indexes tend to measure as 'overbought' or 'oversold' on technical indicators called 'oscillators'; e.g., stochastics, MACD or RSI (Relative Strength Index). Knowing which of these indicators to use and what 'length' setting to use can be confusing; interpretation of an indicator, can be more confusing even. Sometimes, it's not the overbought (a high reading) or oversold (a low reading) EXTREMES that suggest that the market is topping out or bottoming, but failure of an underlying indicator to match price action.

3.) Retracements or 'give backs' of prior price swings tends to fall into certain pre-set patterns that can be measured and tend to adhere to certain 'fibonacci' price relationships.

There is also a tendency for the major market indexes to fall just once, or 1-2 to a few times, to a certain percentage below their 21-day moving averages and then rally. This tendency tends to be more true for DECLINES, rather than rallies, as advancing trends tend to occur from steady and REPEATED buying on the way up, whereas declines often stem from many participants dumping a lot of stock quickly which tends to get it over with faster and all at once.

I've covered points number 1 and 2 quite often in my weekend "Index Trader" columns or in this Trader's Corner column written mid-week on Wednesdays (occasionally Thursdays). Most readers of my columns I think know a fair amount about my thinking on these topics. You need only go back and read my most recent (Sat, 11/12) Index Trader column and see what I have to say about 'sentiment' (point #1) or overbought/oversold considerations (point #2) pointing to this recent top and sharp reversal. You can link to the aforementioned article by clicking here.

As to point #3 above, I got thinking about this question the other day, when Jim (Brown) was writing about the sharp rebound on Tuesday, in his Market Recap in the Option Investor Daily and said that: "You can only stretch that rubber band so far before it snaps back with lightning speed." How true, but the question I was wondering about from my own technical perspective is HOW far is 'so far'? What rules can we look at for HOW FAR down could this recent correction could have been expected to go? When was the time to cover puts and perhaps go into calls or sell puts, if we wanted to participate in an anticipated 'snap back' rally.

My experience with the following principles, trading 'rules' and technical concepts stem from trading experiences from the late-1970's on that included managing stock index funds using derivatives, some years as a financial futures and stock broker, 7 years as a UBS (then PaineWebber) senior technical analyst specializing in stock indexes, years spent as Dow Jones' global and European manager of technical analysis software services and a couple of final corporate years (on the 105th floor of the North Trade Tower) as a Cantor Fitzgerald manager and CNBC commentator.

The fibonacci number series are numbers that result when you add the prior two numbers; i.e., 1, 2, 3, 5, 8, 13, 21, 36, 55, etc. Fibonacci retracements are certain pre-set amounts of a prior move expressed in percentage terms. For example if the S&P 500 (SPX) runs up to 1500, then falls back to 1400 and takes off on a run to 1600, the most recent upswing was from 1400 to 1600 or 200 points. If SPX starts to correct or retrace some part of that 200 point move, how far might it fall in terms of the fibonacci retracement levels of .382, .50 and .618; i.e., what is the 38, 50 and 62 percent retracement levels of that 200 points? Easy enough to figure: .38 X 200 points subtracted from the SPX peak price of 1600 on that last upswing equals 160076 or 1524; .5 X 200=100 and subtracted from 1600=1500; .62 X 200 = 124 subtracted from 1600=1476.

Two things: there are applications or marking tools in most charting applications that will figure out the common fibonacci retracements and put a level line on those retracement amounts; e.g., click on a 1400 low, then on the imagined 1600 high and the charting application will show the fibonacci retracement levels. HOW to use these retracements or what are the RULES of retracements is another thing.

A shallow correction in a strong trend, up or down, will not typically retrace more than 38% in the major stock indexes. If a retracement goes beyond this, and still within the realms of a 'normal' retracement', will be one-half (50%) of the prior move; this is more common in individual stocks. If a retracement of a major stock index exceeds 50% of the prior move, then we can anticipate the possibility or even likelihood that there will be a 62 percent retracement or a BIT more; how MUCH MORE is often a retracement that equals 66% or 2/3rds of the prior price swing.

While once in a while a retracement will go as far as 75% or 3/4ths of a prior move and still RESUME the prior trend, it is more common to see a full 100 percent retracement back to the starting point of the prior move. Bottoms that are made after a 100 percent retracement have a name: double bottoms. Tops that equal a 100% retracement of a prior decline and don't go further are called double tops.

Are retracements self-fulfilling prophesies so to speak? Do traders start buying in anticipation that a decline will stop at the 50% retracement; or, at 62-66 percent? Indeed, some buying/selling will tend to come in at the fibonacci retracement levels, but can a limited number of traders willing to step up to the plate at these pre-set levels, determine the outcome and ultimate turning points of the huge US stock market? Unlikely! For reasons we won't go into, having to do with the nature of the unfolding of market cycles, 'natural' price progression, etc., the fibonacci sequences seem to reflect a natural order of things, including the unfolding of market cycles.

When I say 'cycles', I don't mean in the limited sense of repeating 'time' cycles of a 1-month, 6-day or 6-year duration. Rather, I refer to how market movements unfold based on what has come before.

I'll start with a more extreme or far-out example just to open this topic and then move to the more mundane and practical aspects of how you can use fibonacci retracements to assess where a correction may end (and in tandem, the possible use of moving average envelope lines) in helping you make better (i.e., profitable) trading decisions.

W.D. Gann, a somewhat legendary trader, active in the commodity and stock markets from the early-1900's to his death in 1955 and who is known to have made fortunes in his trading, held that the future unfolding of a market trend could be projected into the future based on a prior major high, a prior major low or, best, from the price range established by a major high and major low. Moreover, it was his belief that the two price axis of a chart, measuring price (horizontal axis) and time (vertical axis) could be EQUATED in measuring how far a future move might carry before a trend reversal set in, as well as suggesting where major resistance and support might be found in the future unfolding of a trend.

I took the weekly chart of the S&P 100 (OEX) back in 2005 and laid out the grids of the type used by Gann and even in my 'amateurish' way, without the special software that could create 'price and time squares', found much of interest in the unfolding of the OEX trend in the following 2 years. This chart is seen below:

I began drawing in the lines of this future 'grid' based on using the price range of 202 points established by the July '02 bottom at 385 and the '05 top at 587. The notations I've made on the chart give you some idea of how the unfolding trend AFTER early-2005, could have been expected to develop. The really interesting thing I found was how major trends began at the 'end' of the squares determined from the price range of 2002 to 2005. Go figure! This most recent trend reversal could be anticipated, give or take 1-3 weeks, in this current time frame.

Remember with the above chart that I started constructing this overlaying grid above on the OEX weekly chart 2 years ago and had of course no idea of how the strong continuation of the prior bull market would go or where any big corrections might develop. The above chart construction is by way of an extreme example of 'predetermined' cyclical patterns. Now on to the more practical and mundane examples of pre-determined fibonacci retracements, the use of which helped define where at least initial strong support would likely be found.


Deeper corrections, that exceed 50 percent of the prior move, will tend to go to a fibonacci ('fib' for short, no pun intended!) 62 percent or a 'bit' more, which is often 66%, as seen with the recent low in the S&P 500 (SPX) in the 1440 area. SPX is chopping around now, but has a good chance to 'base' around recent lows and rally over time, especially in December.

Note that I use the 13 'length' setting, a fibonacci number, for the Relative Strength Index (RSI) indicator. The fact that an 'oversold' (30) reading occurred in conjunction with the Index 66-66% retracements, was a tip off to collect most of your profits on S&P and DJX puts.

Whether you wished to buy calls at the 2/3rds retracement level, anticipating a bounce was another story, but it was a reasonable speculation if you bought at the 66% retracement, with a stop just under that level, anticipating a rebound back up the 'breakdown' point at 1490; a level of prior support, which, once broken, 'became' resistance later resistance; and a well-known technical analysis concept.

If there is a decisive downside penetration of the 62-66 percent retracement zone, there is potential for an SPX 'round-trip' of a 100% retracement or back to the area of the SPX August lows around 1380 or a bit lower.

The Nasdaq was the stronger index in recent weeks and resisted the decline the longest. The final sharp price break in the Nasdaq 100 (NDX), when it came, could be anticipated from the bearish 'rising wedge' pattern outlined below, the break of the steep up trendline and the 21-day moving average. However, what had been the strongest market segment would not be expected to have the DEEPEST correction, suggesting that the correction would be at most a fibonacci 62 percent; and NDX of course stopped just shy of this level but also got 'fully' oversold on the RSI.


'21', another fibonacci number, is the single most useful length setting for the moving average to use for the major stock indexes in assessing how the intermediate-term trend (e.g. 2-3 weeks and more) is faring. This average can define both areas of support and resistance. In the S&P and Dow, prices tend to trade in a range that is around 3 percent above or below the key 21-day average in a non-volatile period, and ranges up to 4 percent in the S&P and Dow when the market starts showing more volatility, especially on the downside. Recent lows in the Dow 30 (INDU) have touched the 4% lower envelope line.

The August low cut seen in INDU (above) cut through the lower envelope line (at 4% below the centered moving average) but the Dow quickly snapped back. A decline to the lower envelope line, coupled with oversold extremes, tends to be a tip off in downside corrections to at least cover shorts and exit puts. Chances are that when prices reach the area of the lower envelope line especially given an oversold 13-day RSI extreme, you have garnered 90 percent of the profit you're going to make in DJX puts; e.g., assuming purchase around the time the trend reversed in early-October, or if entry was after the mid-October rebound rally failed to extend its gains above the 21-day average.


The Nasdaq will tend to see highs and lows that hit extremes of 4-5 percent above or below the 21-day average, with the greatest extremes occurring on the reactions occurring after a prolonged (what me worry?!) and strong advance as was seen in August-October. If the rally was an 'extreme' so to speak (no direction but UP), the REACTION will also tend to extreme.

If the prior intraday low (mid-August) occurred at 6 percent under the 21-day average, as in the example provided by the Nas 100 (NDX) index chart below, a reasonable expectation or guess for the next low will be to expand the lower envelope to what it was at the last extreme. And, sure enough, the recent intraday and closing low was very close to the same 6 percent lower envelope line.

Envelope lines give a valuable idea of where, on a PRICE basis, a correction has reached an extreme. Do I use them alone, without assessing other patterns and tools? No, but it gives an excellent read on where a correction has a strong probability of having run it's course. This is not to say that there will be an immediate strong rally and the prior UPtrend will resume. Deep corrections of around 62-66 percent of the prior advance are showing that a serious underlying concern about the viability of the trend has set in.


Trader's Corner Archives