This article is best viewed by clicking on the “Print Friendly” icon at the bottom of the post.

In this post, I’m going to take a look at a MIDAS application in the foreign exchange futures and gold futures in relation to current uncertainty over a top that is likely to be fairly near in EUR/USD.

Continued US dollar weakness

The US dollar further weakened against other major currencies last week, resulting in more consternation to many Elliotticians, who in the past two months have seen EUR/USD break above the 61.8% retracement level and significantly above the 50% level which statistically would be a more typical retracement zone for Elliotticians looking at the rally in the Euro from October 2008 as a probable ABC correction. Besides the breaking of Fibonacci levels, bullish sentiment in the dollar remains low in a recent poll (17%) while headline news has been far more alarmist.

Contrary sentiment readings

Other indicators, however, are beginning to show extremes in current Euro strength. In the futures markets, the Commitment of Traders report for the CME-traded Euro futures shows the Net Open Interest Positioning of Large Speculators at 53,597. This is not as extreme as it has been over the past two years, though it is getting pretty close. More significantly, the Open Interest positioning of the Commercials at approximately -70,000 is now rapidly moving towards the extreme levels last seen in mid-2007 in the long anticipation of EUR/USD weakness. The FX options markets too are indicating that daily, monthly, and yearly implied volatility levels are also very close to the extremes when Euro strength was at its peak between March and July 2008.

The problem of timing the top and the role of MIDAS

With continued Euro strength, however, the problem lies in timing this top, as a number of FX commentators have repeatedly emphasized recently. Forecasting precise reversals in implied volatility levels is always a challenge and the same goes for COT data. For example, net Large Speculator positioning kept diverging from a rallying Euro for at least 7 months before the first sign of a top appeared in May 2008.

I believe the MIDAS approach has a significant role to play in this current dilemma and the purpose of this blog entry is to highlight what that role is.

Figure 1: continuous CME Euro FX futures with a critical five MIDAS support curves

Figure 1 is a continuous chart of the CME Euro FX futures. It shows five MIDAS support curves launched from significant portions of the trend beginning in early March 2009.

As pointed out in the MIDAS essay on this site, a significant feature of MIDAS curves is that a trend is capable of only supporting four to five of them before it is due for a reversal or a significant breather. This has been shown statistically many times over. In Figure 1, with S5 being the small dark grey curve launched at the beginning of October, we see this rally now supporting five curves and very likely close to its end.

The role of the MIDAS Topfinder (TB-F) curve

Again however the devil of the detail lies in precise timing. Can the MIDAS system allow us more precision? Yes, it can.

Attention should now focus in Figure 1 on S2 (the red MIDAS support curve), which supports the correction in mid-June but then diverges from the trend. The situation is not perfect for the launch of a TB-F curve, but the fact that there is actual divergence indicates that the trend has started to accelerate and that a TB-F curve has a role to play.

Figure 2: A Topfinder curve on continuous CME Euro FX futures

With this acceleration of the latter part of the trend in evidence, a TopFinder curve was launched from the late April low and fitted to a number of pullbacks in the accelerated portion of the trend (mid-June, mid-July, and mid-August). The cumulative volume (duration) required for this fitting was 29,999,000.

What the chart shows is that the TopFinder curve has a remaining cumulative volume of 3,453,919, meaning that the trend is 88.5% done in cumulative volume terms with 11.5% of the trend still to run. Each day of trading will run the remaining cumulative volume down to zero. When zero is reached, this portion of the trend will be over.

In Equivolume charting, cumulative volume and not time runs along the horizontal x axis, meaning that we can actually see on the chart where price will terminate by the intersection of a vertical line up from the x axis and an extended linear regression line through price. The current chart is using conventional candlestick charting, though we can still get a reasonably accurate projection. This projection is around the 1.52 level at the upper side of the regression channel shown on the chart intersecting the vertical line. 1.52 however is an extrapolation — only the cumulative volume is a truly accurate figure.

What the MIDAS system does not do is tell us what price is going to do next, only that there will be a price response when the duration of this accelerated trend expires (= cumulative volume at zero). With this in mind, Figure 2 also shows a significant negative divergence in the MACD, which so far has been in existence for four months — exactly the time of the last positive divergence also highlighted on the chart. A rule of thumb I have found in MACD projections is that the MACD-measured length of the move is usually the same to double the length of the divergence. This can be seen in the case of the positive divergence, where the market has rallied for 8 months since the 4 month positive divergence. The negative divergence before that signalled a move of roughly the same duration. While therefore the TB-F curve cannot supply a forecast as to what will happen next, other indicators — and certain Elliott Wave counts — indicate a likely end of the trend.

We can also gain support for this view using certain market timing techniques — in this case, the Delta system.

Further timing considerations and the Delta system

Leaving aside other sophisticated market timing techniques such as Christopher Carolan’s Spiral Calendar, one system that is central to my timing work is Welles Wilder’s Delta system. This is illustrated in the chart below.

.

Figure 3: Long Term Delta on continuous CME Euro FX futures

Figure 3 shows a weekly continuous futures chart with the Long Term Delta for the Euro FX futures. Hidden is a multi-week negative divergence in the MACD.

One significant feature of Delta time series is that markets tend to be extremely volatile, often producing significant trends, at the end/beginning of new Delta cycles. In Figure 3, this can be seen in what is likely an inverted (1) in my current solution to the Long Term Delta in the Euro FX futures. More to the point, as an be seen from years 2001 and 2005, pivot 6 is due some time between the end of 2009 and early 2010. There is a degree of standard deviation in Delta pivots, which is why it’s helpful to use a more precise indicator such as the MIDAS Topfinder/Bottomfinder. However, when pivot 6 does print, years 2002 and 2006 indicate that pivot 7 isn’t due until midway through 2010, which would be consistent with the MACD readings and the extremes of the recent COT reports.

The variance issue in Delta timeframes such as the Long Term Delta can be ameliorated to some degree by using a Fibonacci timing technique such as Fibonacci ratios or the Fibonacci number sequence. Using the latter, a date of November 9^th for a projected top occurs three times: first, from the April 21^st low at 144 days; second, from the July 8^th low at 89 days; and third from the September 29^th swing high at 21 days. In my opinion, November 9^th is an important time period within the variance allowed by the Long Term Delta. At the same time, we also have a recent vertical count on a Point and Figure chart (3 box reversal x 0.0025 box size) of the same continuous futures indicating a price target of 1.50.

Price and time projections, then, are supportive of the Topfinder cumulative volume readings.

Gold futures and the MIDAS Topfinder curve

A full report on the Euro isn’t possible without a corresponding look at gold futures, since a multi-month turnaround in the US dollar should be consistent with bearish gold projections. In Elliott Wave terms, it’s possible that the move above the February 2009 high is a wave B of an expanded flat, since a possible wave A declined in three (a three wave A decline = flat, expanded flat, or triangle). This putative wave B is now in an acceptable Fibonacci expansion range for the completion of a wave B of an expanded flat.

Figure 4: Topfinder applied to continuous gold futures

As can be seen from Figure 4, we have also been fortunate enough to launch a Topfinder curve from an accelerated portion of the current trend in gold futures beginning in early July. This curve has been fitted twice (in mid-August and late-September) at a duration (= cumulative volume) of 9,000,000. It is now revealing that this accelerated portion of the trend is 83% done. This figure for the remainder of the accelerated portion of the trend is therefore pretty close to the result obtained from analysing the accelerated portion of the Euro FX trend with Topfinder. It should also be borne in mind that the MACD has been diverging negatively for a significant period of time. While this negative divergence doesn’t carry the same weight as it would in a self-contained trend, it should be borne in mind.

Any turnabout in the Euro’s fortunes will face a wall of support, not only from weekly and daily moving averages but also from five MIDAS support curves.

Appendix: the US dollar, short term rates, and US stocks

Arthur Hill has recently drawn attention to the fact that while the US dollar has been negatively correlated with stocks this year, it is positively correlated with short-term interest rates.

Figure 5: 1 Year Treasury Yield and the S&P 500

Gold was recently amenable to MIDAS analysis and if I have time this week I’ll look at the MIDAS implications in the S&P 500 and US short term interest rates.

Summary

The presence of five MIDAS curves in Figure 1 supports the findings of other indicators that current strength in EUR/USD is reaching its limit. At the same time, the possibility of launching a Topfinder curve to the accelerated phase of this current trend indicates a firm cumulative volume prediction.

While the Topfinder component of the MIDAS system doesn’t provide a longer term view, other price and time indicators — including Elliott Wave counts — indicate the likelihood of a trend reversal.

– Andrew Coles, 19 Oct 2009

Tweet This Post