Registered User
Joined: 10/7/2004
Posts: 799
Location: Duluth, GA
|
"Swing" traders have typical hold-times of 5-15 days, and commonly use daily bars as the basis for their chart evaluations, and as as the root of their parameter specifications when using indicators. For faster plays of only 2-3 days, many traders prefer to use hourly bars so that they can get better resolution of what is actually going on.
Of course, most traders don't rigidly follow just one particular time window .. they will often take trades that they expect will be longer or shorter than their average. Unfortunately, many charting and analysis packages require them to create entirely separate sets of charts and formulae and system logic for each hold-time variant ... or just "live with" a non-optimal average. Some folks go to the extra trouble to do this, even if they only have daily-bar capabilities, by shortening the parameters used in their indicators, or by revising their visual interpretation of chart patterns, etc.
Unfortunately, a need to add nuances like this just creates confusion, and fosters mistakes and bad decisions. As a result, many traders are operating at a disadvantage, simply because their trading software is too limited! It's MUCH easier just to decide on ONE set of rules and formulae and pattern-searches, then "squish" or "stretch" the bar-lengths that are displayed and used for calculations to fit the anticipated hold time. Let the computer do the work - don't force your brain to re-adapt!
Another very important but often-overlooked reason for adjusting the bar-length to your typical hold-time is in the evaluation of Risk, when the trade is entered. This is often the basis for setting (and moving) stops during the trade, and should be (a least partly) a function of recent volatility in the price. The evaluation of that volatility is done by considering the High-Low range of a typical bar (possibly also using the interbar gaps). Therefore, the bar length used for volatility calc's should be *proportionate* to the length of the hold-time expected.
This is very well illustrated by the needs of traders work on much shorter timeframes, sometimes in and out of the trade in just an hour or two. In those situations, daily bars are useless, so the trader picks and appropriate "faster" barlength for charting and for indicators ... typically 1-min or 5-min. The "risk" of the trade is related to the "wiggliness" of those very short bars, which can't be reliably predicted from the (much larger) daily-bar volatility. The formulae and rules *must* be adapted to the shorter bars. If the trader prefers using hold-times which cover more nearly a complete day, then bars of 10 or 15 min are more commonly used. If the hold time might carry over into a second day (to capture an overnight gap), then a 30-min barlength are more appropriate.
TeleChart provides all of these intraday bar-length choices, both for charting, and for formulaic indicators and pattern-evaluation. That is, they allow the trader to scale the rules to fit their personal hold-time. TC-Net's charts will quickly and automatically adapt to any of these selections. PCF's are limited to daily-bar resolution, but the PCF formulae can easily be plotted as Custom indicators, and the SortBy function provides versatile ways to scan for threshold-conditions on a realtime basis.
But what about the medium-to-long-term trader who is more active than a pure long-term buy-and-hold investor? These folks make up a growing (but mostly silent) majority of the trading community, as our confidence in Social Security and Investment Advisors and Fund Managers continues to dwindle. For these traders, a "fast" trade is at least 20 days (to avoid wash-sale-rule penalties), and often will have hold times of several months to a year or so. Traders who prefer working with Options often use these time frames. Daily bars are *not* the answer for this kind of trading ... it would be like looking at a chart full of 10-min bars for a trade lasting 10 days ... too much information!
In our examples above, it seems that the "comfort range" for visual and parametric analysis of a typical trade will have between 10-20 visual bars within the trade's duration. This allows us to easily identify chart patterns before and during the trade, and gives adequate resolution for formulaic indicators to be evaluated, without introducing a lot of "wiggles". So, for the mid-to-long-term trader, it stands to reason that a similar 10-20 bar optimization would be useful.
Some charting packages offer Weekly and Monthly bars in recognition of these longer-term trading situations. But this is much more limiting than the 1, 5, 10, 15, 30, 60-min & daily choices discussed above, for more active shorter-term traders. Reliance on "calendar" bars derives primarily from historical times when computerized access to and manipulation of data was limited to only the exchanges and institutional traders. The "hoi polloi" depended on newspapers and magazines as their primary source of data ... daily, weekly and monthly bars matched up nicely with the distribution patterns of the publications.
Of course there are some good psychological and fiscal reasons for using calendar-based bars: it's commonly accepted that the prices during a week tend to follow certain patterns (probably due to the weekend-hiatus), and many corporate and industry reports come out on a monthly or quarterly basis. However the specific times of those disclosures often are not on the exact calendar division-dates. Even the "weekly" patterns are confused, by holidays, options-expiration, and significant news events.
Therefore, now that computers are ubiquitous, and have the ability to present data in virtually any way you can imagine, it makes sense to apply the same "scaleability" of formulae and patterns to the mid/long term hold times that we take for granted in the world of faster trading.
The simplest way to implement this is to allow the trader to assign their chart's bars to be whatever multiple of days best fits their hold-times. The final "closing" price of the bar is the Close of the most recent day; the opening price is the Open of the first day that the bar includes; the High and Low are the highest and lowest prices reached during that time. For example, a 3-day-bar's Open would be the Open of the day-before-yesterday (not counting days when the market was closed).
This "N-day bar" approach is exactly what TeleChart provides ... and it is (to my knowledge) the only widely-accepted package that does so today, in an easy-to-use "no-brainer" format. Many people are just using this capability as a convenient way to check out the "long view" - but with the implementation of Custom Indicators and SortBy tools, TC users are given an important additional level of adaptability and control, rather than becoming "slaves" to out-of-date limitations.
To recap, here are the benefits of N-day bars for the trader whose hold times might vary from hours to days to weeks to months:
1) Volatility-based Risk evaluation for position-size and stop-placement automatically adapts
2) Parameters for specification and interpretation of Indicator formulae can remain unchanged
3) Visual recognition of Classic and Japanese chart-patterns remains consistent for different hold times
4) Processing time in the software is optimized for the trader's needs
Once again, I applaud Worden's "out of the box" thinking in the design of TeleChart ... TC does the hard work, and lets our decision-making remain natural, fast and consistent.
|