An underlying theme of much of Gann's work is in determining trends. Whether a Gann fan line of 1 x 1 works or not is due to other considerations. It is generally accepted that Gann was open to any ideas, some of which still seem esoteric to me. However, I am not in a position to discount them since whenever I ask people whether they trade them, the answer is "yes." Putting aside the reasons why Gann seems to work at times, the other question is can you trade Gann?

For this you need to consider that there are generally three types of markets to trade: trending, choppy, and volatile. Trading a trending market means going long on uptrends and short on downtrends. A choppy market moves sideways and you attempt to pick highs and lows as best you can, and buy on the bottom and sell at the top - support and resistance play a major role as well as indicators. Trading a volatile market means trading prices only and not indicators. The best example is looking for patterns and then for breakouts, gaps, dead-cat bounces, etc. (a volatility trader might even boast that they don't use indicators).

Trending market traders loathe chop. The reason is that to trade trending markets you usually use moving averages. Try as you might, moving averages always lag. Which means in a choppy market -- one that moves up and down -- you are late getting in and when the market reverses you are late getting out. To be successful as a trending market trader you need to find a market that trends - some stocks and especially some commodities tend to trend. There are also helpful money management schemes that can be employed in a trending market but would have problems in other markets. So Gann didn't just come up with Gann fans, he also had to be concerned about determining if the trend changed and if prices are moving up or down while in trend. Gann is best used in conjunction with trend-trading systems. Gann's rules of trading also contain advice that fits a trend trader.

An approach attributed to Gann for determining whether a trend changes is the use of swing points: peaks or valleys. To visualize a swing peak or valley think of price action in a trend channel. Swing peaks occur when prices touch the upper channel line and valleys are the points where prices touch the lower channel line. A swing point marks the point at which a trend could change direction. More generally, swing peaks form intermediate peaks (valleys) when subsequent price changes fail to rise above a previous peak or fall below a previous valley. If a subsequent price move rises above a previous peak, the trend is up. If price falls below a previous valley the trend is down.

Arnold (see "Gann" by Curtis Arnold, Stocks & Commodities Magazine, March 1983) defines swing points by saying that in a downtrend a swing occurs when there are three days of higher highs, and for an uptrend when there are three days of lower lows, except for reversal days. A reversal is a temporary retracement and has not created a swing point. The giveaway to a reversal is the reversal day. In the context of a downtrend Curtis states, "a reversal day top occurs when prices move higher but then close near the lows of the day, usually below their opening and below the midpoint of the day's range. An even stronger reversal is indicated if the close is below yesterday's close." (See Figure 1 for an example). Just the opposite is true for uptrends.

Figure 1: Trends and reversal days.
Graphic provided by: Technical Analysis Inc..

Arnold trades Gann by going "short on a close below the 1 x 1 from a previous swing bottom." If you go back to Figure 1 in Part II of this series, and examine grid line G1, you will see a low on October 30, 2000. This is a swing valley. The grid line G1 forms resistance/support for the upswing which ends on Novmber 3, 2000. Using Arnold's rule, since you are still in a downtrend, you would short when the close drops below G1 (a 1 x 1 line since I made the grid using a 32 x 32 rise over run) which is around November 3, 2000.

Robert Krausz in his article "The New Gann Swing Chartist" (Stocks & Commodities Magazine, March 1998) writes about this subject as well. He was fortunate to have bought a copy of Gann's course from someone who had purchased the course from Gann. There is a notation on one of Gann's charts that says, "Use 2 day charts and rules better than 3 day." What is the implication? Krausz explains the use of a HiLo activator, which is a three-day average of highs along with a three-day average of lows. Using the HiLo activator looks a little like envelopes that move along with the price. It is used as a filter to test the validity of the swing. Krausz makes an important distinction between swing direction and trend direction. Currently the Nasdaq is in an overall downtrend, but has swings that temporarily carry prices higher.To be technically correct when I discussed manually creating G4, the Gann fan line whose origin is O3 (see Figure 1 in "Gann and Fibonacci Part II of III"), went through the upswing not the uptrend, of prices. It also suggests that Arnold's article may want to change, except Arnold still would give you a strong signal, although maybe not quite as fast.

Arthur A. Merrill in the article "Swing Expectations", (Stocks & Commodities Magazine, December 1988) makes an extremely valuable observation. He asks: if the price has already gone up by X%, how likely is it that it will go up Y% more? The idea is to make a distinction between noise and something significant happening. To determine trend change, Krausz uses swing points and asks is the price higher than the nearest swing peak or vice versa. Merrill has a different slant that I like even more, because using ZigZag (an indicator that filters out changes less than a chosen percentage) allows me to see the volatility it took to create a trend with a bunch of swing points defined by percentage change. This also gives me a chance to change the percentage when a downtrend is established so I can take advantage of the fact that stocks come down faster than they go up. For example using a 5% change on the Nasdaq Zig Zag shows five swing points in the uptrend going from October 1999 to the peak in March 2000, while the downtrend from September 2000 to the present shows 19 swing points (so shorting on a downtrend requires more margin - cash flow).

J.R. Davis ("Swing Charts", Stocks & Commodities, August 1998) writes that Gann did not use daily charts and suggests the creation of swing charts by counting up days and down days - a bit of a simplification of Arnolds' approach.

So how do the pieces fit together? David Lammar ("Trading with Elliott Wave and Gann", Stocks & Commodities, June 1988) writes that Elliott waves provide price targets based on the ratios of .618, .5, .382, and 23.6, which you will recognize as Fibonacci numbers. Lammar writes that he was able to significantly improve his performance using Gann in connection with Elliott wave determination. So there appears to be a connection between Gann and Fibonacci numbers. Gann would have been very familiar with point and figure charts, which require that you write a new X or O on your chart only if the price change is significant enough, so for Gann it looks like the determining factor was also two or three days in time.

Warning and Disclaimer:
Trading involves risk of loss and may not be suitable for you.
Past performance is no guarantee or reliable indication of future results.
This message is of the nature of general information only and must not in
any way be construed or relied upon as legal, financial or professional advice.
No consideration has been given or will be given to the individual investment
objectives, financial situation or needs of any particular person,

The decision to invest or trade and the method selected is a personal decision
and involves an inherent level of risk, and you must undertake your own investigations.