Percival and the Holy GrailTo know whether the market is currently trending, you have to accumulate a database of intraday progression over at least 30 trading days; then continue to add to your database and trade concurrently. Let's first discuss "intraday progression".

For each stock, currency or futures contract, you need to define the day's "opening range". For a typical stock or commodity: it may be daytime opening 15min "hi/lo box". For internationally-flavored futures contract, the day may well be starting overnight. For FX currency-pairs: I recommend first 30-60min range of domestic time-zone (e.g. Far East open for AUD and JPY crosses, but N. Am. open for $/CAD). Exercise discretion and make provisions for markets that would commonly tread water in advance of crucial regularly-scheduled news releases, and then significantly gap. Award a score to each day, ranging between -4 and +4, grading intraday price progression out of optimized (consistent) "opening box". A day (will occur less than once a week) that never significantly traded out of the opening-range all the way through the close will get zero. -4, -3, -2 and -1 are bearish mirror progressions of intraday events described below as bullish +1, +2, +3, +4.

The most typical day (about twice a week) will be a 2-pointer, where market will break its opening box to the upside and end anywhere up there. +1 is awarded to a day (about once a week) that didn't close that high; but it did manage to overcome an earlier short-lived break-down and ended back within the box. +3 is given following same weak start but stronger close well above the box (about once a week). +4 is the case of prolonged trading well below the box, surprisingly reversed by a blistering second-half rally ending well above the box (less than once a week). An extremely rare (once-twice monthly) case would be very late failure of +4 attempt rolling-back into the box and thus getting a zero. You must be very cognizant of the fact that automatic grading of each instrument under the sun, each day, has shown to be somewhat deficient (understandably, those funds who have succeeded and coded at that level of proficiency will never disclose their secrets). Thus, it remains partly discretionary domain (albeit using consistent set of rules and optimized boxes) to assign those -4 through +4 daily grades to each instrument on the trading board of hundreds! Yes, it requires a lot of market-savvy manual labor.

For each instrument, a black box then accumulates running totals as summation of the latest 30 trading days. Although theoretical maximum reading would be 120 either way, we hardly ever see liquid widely-traded markets conquer the seemingly magic 30 level. The key discovery is that a market that just advanced from lower-sum area to over 10 area (positive or negative) is a market beginning to reliably trend! Once beyond 23, such a run will be getting frothy, and we would commence faze out of profitable position. Having rolled back 10 points off the high (say back to 17 from 27) would cause us to abandon trend bias altogether. It should be understood, however, that a new, down-trending bias would only be assumed way later, once the 30-day summation crosses below -10! So it is obvious, that our summation trend-indicator is a laggard. Well, trend-following is a lagging trading concept!

In my over 20 year real-time market experience, I've observed traders disciplined enough to put on biggest exposure exclusively at points of acceleration from 11 to 23 end up as undisputed winners! With decades of steady gains, they would handily outperform a value-investor or a contrarian. These trading programs were hugely favored by macro hedge fund managers and their clients, as lowest drawdown systems and quarterly results that maximized popular yardstick ratios. I will note that large futures positions taken promptly in accordance with above-described signals were roughly corresponding to riding the right side of wave three of three in a typical Elliott Wave impulse, and thus were yielding upwards to 100% return on margin deposit in just one day of trading! That was because commodity exchange margin requirements under SPAN would still be relatively low at that point, as a newly-recognized gapping trend would suddenly begin accelerating.

A necessary postscript is that only a part of the total program was described above, and quite briefly at that. At its best implementation, the program is multi-dimensional and will not (and arguably should not) be fully described in open forum.

Misan Thrope objects:

Holy Grails are mythical. It is philosophically impossible to determine whether tomorrow you will be trending or range-bound.

Rocky Humbert replies:

On your first sentence, arguing that The Holy Grail is mythical is a religious argument – I'll leave it to greater minds.

On your second sentence, if one uses certain physical and/or information models as a guide to financial market behavior, one can reach different conclusions. I respect many people who both agree and disagree with this view. For some, this is a religious debate too!

Lastly, while the Holy Grail may be a myth, the Hole-ly Grail is not. I keep one on my desk. My secretary filled it with coffee once and it made a terrible mess.

Rocky Humbert, quantitative analyst, speculator and master chef, blogs as OneHonestMan.

Don Chu writes:

A very sensible 'hand-count'. Very much like Mark Fisher's work — his opening range AC/BD classifications and rolling multiple day pivot count, a la his The Logical Trader.

Anatoly Veltman replies:

Correct, Don! Mark practiced (and taught) this methodology for decades. I was at first a sceptical participant in his late 80s after-hour weekly group; but listening to the finer points over the next few years (and markets were always discussed in real-time), I ended up incorporating dozens of his ideas into an overall approach. He deserves tremendous credit!





Speak your mind


Resources & Links