But the day is too small a unit to be a useful time frame for most speculators. The smallest useful time frame, I think, is one week –or 5 trading days –being the unit of the smallest normal correction to the minor multi-week move. And the multi-week move is the basic unit of interest to the speculator.

It is a long-observed phenomenon of the currencies that turning-points most often occur at weekends –that is on Fridays or Mondays. In the old days – in the early 1980s when the financial markets were much preoccupied by the release of the US money supply figures on Fridays (nobody pays any attention to them now), this phenomenon was so evident that Currency Bulletin dubbed it “the weekend reversal rule”. But the weekend reversal rule has continued to operate usefully ever since.

What it means is that we measure off the multi-week advances and corrections in the underlying trend in whole weeks. Over the years one has noticed that some numbers occur more often than others. Common numbers for duration of the prevailing trend are 3,6 and 12 weeks –sometimes 9 weeks: for corrections 1,2,3 and 6 weeks. Price cycles being “fractal”, as noted, we often see major movements progressing in similar monthly units –3,6 and 12-month segments being more common than other numbers.

You wouldn’t expect perfection in such numbers games. The issue is whether they are useful. And I have found they are –not just because the numbers have recurred with statistically significant regularity. Perhaps the divisions are useful just as much as a discipline for monitoring the progress of trend movements. When we’ve had a 4-week move, we are not expecting it to reverse, because it is historically improbable. Lt will not reverse if it turns out to be a 5,6, 9-week move or whatever: it will only reverse if it turns out to be a 4-week move! Anyway, the point is that this weekly monitoring can help you stay in a “ good” trend –which means to stay with your underlying rationale when it is proving correct:

There is another strange frequency in the currency markets. Sceptical readers who are used to prejudging events rather than deciding them on the basis of the empirical evidence can skip here. The currencies –and no other markets as far as I can see –are unusually susceptible to the lunar cycle, or more specifically to full moons. Perhaps the reason why currency traders should be susceptible to an outside influence of this sort is that they are peculiarly short of certainties about what makes currencies move.

The effect of the full moon on humans and other animals is well documented. Your local police station knows with great certainty it will be more than usually busy on full moon days and nights –not just with homicidal lunatics (full moon killers), but with a whole gamut of deviant behaviour having in common only that it occurs regularly at full moon season. People don’t howl, but they do strange things at full moon –lunatic things. Read Moon Madness by Paul Katzeff, if you want more background: also Krafft-Ebing’s The Psychopathology of Sex .

The amazing story of the march of science through history has not been about the acquisition of knowledge so much as about the realisation of the extent of man’s ignorance and of the near infinity of what there is to know. Each new major scientific ‘truth’ has displaced an earlier ‘truth’ –only to be displaced again in its turn, or sometimes not displaced but reduced to insignificance. As Robert Pirsig* put it in his masterpiece Zen & the Art of Motorcycle Maintenance* , published in 1974: “there it was, the whole history of science, a clear story of continuously new and changing explanations of

old facts… some scientific truths seemed to last for centuries, others for less than a year .”

The trouble with the intricate and astonishingly accurate system put in place by Isaac Newton to explain the physical world is that it persuaded subsequent generations to look for the truth in things that were capable of proof (or rather being incapable of disproof, as Karl Popper suggested). But there are more things in heaven and earth: the parts the Newtonian mentality can reach are a small, and perhaps not very important, part of the whole. It doesn’t reach to human behaviour patterns that are geared to the lunar cycle.

The point of departure is the fact that for the past 10 years, short-term, multi– week extremes in the currency markets have coincided with full moons in a way that could not possibly be explained by chance. More significantly, this coincidence was noted in Currency Bulletin as long as 7 years ago. In other words this is not an interpretation that has been fitted to the past: one has been observing the phenomenon being confirmed in ‘real time’ over many years. (The chart below is not specially selected: it is simply the latest period available to me as I write. In it, the full moons are marked with a hollow blob; and the new moons with a solid blob. When they occur at weekends, which is obviously quite often, I have tried to place them between Friday and Monday).

Translating this observation into a trading rule is not easy. Full moons aren’t always accompanied by turning points: and turning points by no

means always come at full moons, of course. Moreover, if the rule has been working well over several lunar cycles, it has a way of suddenly falling apart: possibly it self-destructs because many traders have noticed the occurrence and are trying to anticipate it. You have been warned. Study the chart at leisure: one of the conclusions you may draw is that once a trend is manifestly in full swing, the moon has little impact.

Active traders have little to lose and possibly much to gain by observing the following maxim: distrust price action ahead of a full moon, trust the action after it. On full moon days, cut back trading positions to core positions if prices have shown a clear-cut trend ahead of the full moon: other things equal, re-instate only if the trend is seen to continue on the second day after full moon. If prices have been trending sideways before the full moon take no action, but watch the direction of the next two days for guidance.

The influence of the full moon may only be felt for about 3 days beforehand and 5 days after; and it has been most visible in the D-Mark and SF dollar parities. But as you can see from the chart, there has often been a two week cycle running from full moon through new moon and back –a cycle which has sometimes been uncannily reliable for months on end. Rationalise it as you please: the impression is that market action tends to be primitive, dim and emotional before full moons, and more collected and rational after them; and that there is sometimes a periodicity in currency fluctuations which can be almost as reliable as the tide! If this continues to be so, we will have an edge if we exploit the phenomenon. And successful trading is not about being a genius, but about constantly exploiting ‘the little edge’ .As the legendary futures trader Richard Dennis* put it: “If you take something which has a 53% chance of working each time, over the long term there is a 100% chance of it working.”