Professional Documents
Culture Documents
TradingWithProperlyScaledStaticCharts AL LARSON
TradingWithProperlyScaledStaticCharts AL LARSON
Stock, commodity, and currency traders who use most modern trading programs are missing some great opportunities. Why? Because most modern trading programs cannot draw a properly scaled static chart. Many traders do not even know they are missing some great opportunities. Do you even know if your charting program can produce a properly scaled static chart?
Using Angles
Gann used lines that had harmonic slopes to find key turning points. Harmonic slopes meant that the slope of the line was the ratio of two whole numbers. Gann realized that these lines represented the speed of a natural cycle, and that multiple cycles would harmonize near key turning points. On this chart, the 45 degree lines represent a 30 year cycle of 360 points. The 22.5 degree lines are the 60 year cycle, etc. Where these "Gann Angles" converged to an intersection, he looked for a turn and a trade setup. Gann related these angles to astrocycles. His 30 year cycle is the Saturn cycle, which harmonizes with 3 Jupiter-Saturn cycles every 60 years. When his angles converged, he expected to find confirmation in the astrocycles - sort of a "second opinion." Gann angles applied to the badly scaled charts produced by most modern charting programs have little meaning and little chance of helping the trader. Properly scaled charts are required for the proper use of Gann Angles.
Using Patterns
Scaled static charts also permit the use of two dimensional patterns. Current charting programs that use only indicators are using only one dimension. The current focus on developing computerized trading programs that use only indicators is unfortunate, because there is a much greater set of opportunities if one uses two dimensions. One of the early pioneers of the use of patterns for trading was George Bayer [2]. He used precisely scaled elliptical overlays on top of properly scaled charts to find key market turns. This chart shows one of his ellipses on a well scaled wheat chart.
The ellipse is placed at a 45 degree angle. The ellipses are of a precise, predetermined size, and they only work on properly scaled charts. George Bayer produced many books and courses illustrating the use of his ellipses on properly scaled charts. Like Gann, he linked these patterns and angles to cycles.
Can't We Do Better?
But does it make sense? Why should modern charting programs only produce "joy ride" charts? Sure, they are emotionally stimulating, but that may not be a good thing. More importantly, the computer limitations that were the underlying reason for the "joy ride" charts are gone. My original IBM PC had 64 kilobytes of memory, and a 10 megabytes hard drive, and a 480 x 320 x8 display. My current PC has 8 gigabytes of memory, and 250 gigabytes of solid state disk, and a 1920x1080x32 display. So my PC has 125,000 times as much memory, 25,000 times as
much disk, and a display with 54 times as much resolution. It runs 100 times faster. So there is no reason to only produce "joy ride" charts. Such capabilities can easily do outstanding scaled static charts.
Getting It To Work
However, there are some tricks to implementing properly scaled static charts. They are harder to program than "joy ride" charts. I've learned this from building two proprietary charting programs that produce properly scaled static scaled charts. First, one has to have a clear concept of time and how to add and subtract time. The answer to this is to use Julian Equivalent Dates (JED), as used in astronomy. One needs code to convert from year, month, day, hour, minute time to JED, and back. These routines are readily available. Second, one must use time as the basic unit for the horizontal axis of the chart. The chart has a starting JED, and an ending JED. This axis is totally independent of any data. It does not depend on the time stamp of the data, and is established before any data is applied to the chart. The vertical scale can be fixed to a predetermined range, thus setting the price scale of the chart, or the scale can be used to set the vertical price scale. Both options are needed. The time axis is fundamentally continuous. In the basic case, the chart displays time linearly, left to right. This type chart can then display any type of data occurring any time between the starting JED and the ending JED. An example is a natural cycle, which moves along even when exchanges are closed. Accurate use of time cycles requires this sort of CALENDAR time. It has no time gaps. A less accurate time scale may have segments omitted. A common example is to omit weekends. This TRADING DAY time should not be the default, but an option on top of the basic CALENDAR time. The inclusion of space for holidays should also be an option. Another case of segmented time is session times. One may wish not to see the time intervals when a market does not trade, but it should be an option to display this time interval on a chart. Any elimination of time intervals, of course, reduces the utility of two dimensional tools, so the user should have complete control over collapsing or including these "no data" segments. Data must be buffered with its associated time stamp. For computing such things as moving averages and cycle filters, the data storage must begin before the Chart Starting JED, so a Data Start JED is needed. Ideally, data is stored independently of how one wants bars to be drawn. One option is to store one minute data, and then draw any bars desired from that data. Wave59 does this. In time period where there is no data, the data values needs to be filled with the last value or interpolated when a new value after the gap is available, so calculations that are time sensitive are accurate despite the data gap. Obviously, one must establish internal bookkeeping for the mappings between time, data samples, bars, and pixels. The user should be able to compute indicators based on time, not just bars. For example, moving averages and cycle filters should be expressible in time units, such as a 110 minute EXMA. This should be independent of how bars are being displayed. As a chart is adjusted along its horizontal axis, bars may end up being plotted on top of each other. This is OK. The user can adjust the bar width. But the bar width should have no effect on shrinking or expanding the chart's time scale.
Traders who use pattern overlays need the ability to interrogate the blank space on the right side of a chart to learn when a pattern may terminate. Some "joy ride" programs, which uses data bars to set the horizontal time scale, give incorrect answers when asked for the time of a future bar location. For example, if the data is daily bars without weekends, the software may assume that time in the future includes weekends, because internally it does not know whether it is drawing calendar day or trading day charts. This flaw has been in eSignal for years. Some programs will not even report future times. The data cursor needs to always report time accurately. None of this is beyond today's technology or the skills of current programmers. They just need to understand the requirements.
For now, it is the experience of the author that the human, given an easy to adjust pattern overlay, placed over a properly scaled static chart, can easily place it, thus identifying a price-time region of opportunity on a chart. If the user could then simply draw the "window of opportunity" on the chart, he could then tell the strategy to execute only in this window. That would be a quantum leap forward from indicator only trading.
References _________________________________________________ [1] How to Make Profits in Commodities, W. D. Gann, 1951 [2] George Wollsten, Expert Stock and Grain Trader, George Bayer, 1946 [3] Trading MoonTides Tutorial at DayTradingForecasts.com at http://daytradingforecasts.com/ttsec.asp [4] Fractal of Pi course at http://secure.moneytide.com/hans/institute.asp [5] Polarized Fractal Efficiency, Dr. Hans Hannula, Stocks and Commodities, January, 1994 [6] personal conversation with a trader at Fidelity Investments, 1987. _______________________________________________________________________ Dr. Al Larson is a trader, teacher, programmer, researcher, and electrical engineer with over 4 decades of market experience. He formerly published under the pen name Hans Hannula. He can be contacted through his website http://moneytide.com