Gold: Pattern Recognition and Pattern Repetition
July 31, 2009
The monthly gold continuation chart above shows our long-term working view of price action since the 1980 cyclical peak. Excluding that momentary high at $873, we consider that the market traded sideways between twin 1980 highs at $729 and the 1982 low at $295 (range A - A'). The move above $729 in 2007 signaled completion of the long-term range as well as implied an initial upside target at $1163, distance B - B' (derivation of those levels and our interpretation of the range were originally discussed in an October 2007 special report).
The exit thrust from the broad range reached a Comex official session high at $1,017.50 in March 2008. That level from $729 was 66% of the distance of the broad range A - A'. The aggressive reversal from $1,017.50 accompanied by a peak in the long-term stochastic momentum indicator implied that further consolidation or correction would follow before prices moved above $1,017.50. As you can see, that has happened. Corrective weakness during 2008 reached an official session low at $696.60 in October. While that level was marginally below the $729 area of probable support, selling pressure was not sufficient to force values decidedly back into the 1980 - 2007 range - an event that would have nullified the 2007 upside exit. The long-term trend in gold thus remains up with the broad price swings of 2008 and 2009 appearing to be a classic pullback to and consolidation on top of a previous trading range. When we consider that the trading range spanned a 27-year period and that the uptrend phase to the 2008 high lasted either 8 years or nine and a half years, the ongoing consolidation which has thus far lasted only 15 months could be further prolonged. We would not be surprised at that scenario which could include another probe into the $729 - 696 support. That would be constructive price action in that it would both support another lengthy price advance as well as dull much of the bullish enthusiasm that currently dominates the market. However, factors considered below raise the prospect for an early end to the 2008 - 2009 consolidation.
In the monthly gold charts above we show 3 Elliott wave interpretations of the cyclical advance that has developed since the 1999 - 2001 base. Alternate 1 is that of the renowned Elliott wave analyst Alf Field. We do not know Mr. Field but consider his wave analysis of the gold market since 2001 to be brilliant and hope that we have depicted it accurately here in this bald illustration. Using London p.m. fixings instead of Comex daily ranges, Field's view is that the ongoing major up sequence began at the 2001 low point and that it completed its first leg (of 5) at the 2008 high - point I. While he may be correct in his analysis, we have at least two objections to it that question its validity: First is that Field identifies the 2001 low (point 0) as the start of his upward 5-wave count instead of the 1999 low. The February 2001 low area featured daily range and daily London spot fixing prices that were higher than those that occurred during the July-August 1999 low area (points 0 on the inset charts Alternate 2 & 3). Second, not clearly visible on the above Alternate charts but which we will show in detail later, the deep selloff from the 2008 high to below $700 appears to be an overlapping 5-swing sequence. In our experience similar structured moves are either X-waves or triangles. In this instance an X-wave immediately after an all-time high does not make sense. Neither does a triangle formation since wave theory does not accommodate a triangle in the wave II position. Therefore, while we consider Mr. Field's interpretation unique we think that another interpretation would be more coherent.
Inset chart Alternate 2 above shows our October 2008 attempt at reconciling the 5-swing decline from the March 2008 high with preceding waves. Our conclusion in this interpretation was that an annotated 5-wave advance started at point 0, a 1999 low at $252.50, and terminated in May 2006 (point I at $728). Trading following that high featured an A-wave decline, a B-wave advance to the 2008 high and then the 5-swing C-wave decline to $696.60 (point II). The entire pattern from point I to II is a 3-3-5 sequence that wave theorists would identify as a "flat", specifically an "irregular failure", in which the B-wave tops above the high point of the preceding 5-wave advance and the C-wave tops above the low point of the A-wave. We consider that interpretation to be appealing for 2 reasons. First it accounts for the overlapping 5-swing decline of 2008. Second, an "irregular failure" is generally a powerful corrective pattern that often supports uptrend acceleration (which could be typical of a major wave III that could occur above $1,017.50). This may be a valid scenario but we want to consider yet another, Alternate 3, shown on the above chart.
As in inset chart Alternate 2, inset Alternate 3 features the 1999 low at point 0 as being the inception of the current up sequence. Here the 1999 - 2001 rally and retracement comprise Waves 1 & 2 (circled) of a larger sequence that has yet to top. Following the 2001 low at point 2, uptrend development traced out a 5-swing sequence, annotated, to form Wave 3 (circled) that topped at $1,017.50 in March 2008. Ongoing large swing volatility since then represents a probable Wave 4 that is not yet complete. Prospects for Wave 4 are considered in the weekly charts immediately below and in a concluding long-term historic pattern comparison.
The above series of weekly gold charts considers a non-Elliott wave comparison of a pattern that appears to be repeating. Pattern repetition is common in major moves. For example, a cursory examination of the charts from the broad commodities bull market of the early 1970s will show one that that period was replete with pattern repetition both within markets and across the intermarket spectrum. That we are seeing a repetitive pattern in gold's nine year advance isn't surprising but its occurrence may take some of the guesswork out of what the market is currently doing or may do in the future -- regardless of what Elliott wave scenario may be operative. High level consolidations, i.e.., patterns that followed new price highs, can be seen during 2004 - 2005, 2006 - 2007, and we think presently after the 2008 top. As annotated on the weekly inset charts, we have used H,L,A and B to identify that pattern. The sequence features an initial high at H followed by an aggressive selloff to a low at L. From there a recovery move, A, brings prices back to the area of H. Following A, price action degenerates into a choppy lateral consolidation, B, that forms the springboard for renewed uptrend development.
In the left inset chart on pattern development lasted from April 1, 2004 to September 2, 2005. Point A scored a new price high and the choppy consolidation, B, oscillated around H. Pattern duration was 74 weeks. In the center inset chart the 2006 - 2007 pattern lasted from May 12, 2006 to September 7, 2007. Here the recovery move A, fell short of H but underlying trend dynamics were not compromised. The pattern completed after 69 weeks. Note the "seasonal" time similarity: April - September and May - September, respectively.
Our third example of this pattern began at the March 2008 high and is ongoing. At this point we think that swings H, L and A have been completed while B is underway. The up move to point A was dynamic, scoring a new all-time high weekly close. That leads us to the conclusion that the remainder of the pattern should be similar to that of the 2004 - 2005 experience where the initial selloff from point A marked the lower boundary of the choppy B segment. From that we infer that the April 2008 weekly lows (dotted horizontal line) just above $866 basis the nearby Comex futures should be key support to any potential weakness that could develop during August - September. Beginning on March 17, 2008 the current pattern is 71 weeks old as we write. For now, the duration of the pattern falls between that of the previous two patterns. Also worthy of note as regards similarity is the timeframe in which all 3 patterns have evolved. As noted above, the first two began in April and May, respectively, and concluded in the first week of September. The current pattern began in late March so it is, in our view, in the ballpark as regards the start/conclusion points of the previous two.
The single most difficult problem that we currently confront in our gold market analysis is the appearance of the 2008 decline. Our interpretation of that move as consisting of 5 overlapping swings may be erroneous but if wave structure is instead a 3-swing structure then we still do not see wave evidence that the recovery from below $700 to date is the start of a new uptrend phase. Thus we are currently stuck with the view that market action is corrective. An examination of historical charts comes up with very few instances where a 5-swing decline develops after an all-time high. We have come across one instance where pattern similarity lacks consistency with current gold action in terms of the duration of the pattern. However, we do see consistent relationships in terms of where the move is situated (i.e.. a probable fourth wave) and relative retracement vis a vis the preceding advance.
The weekly charts above present a comparison of silver action in the 1970s with gold action from 2001 to present. Unfortunately, our price database for silver only goes back to 1970 so a significant price sequence just before our chart's inception is missing. In 1968 the U.S.. government stopped redeeming silver certificates with silver dollars. Anticipating that event, the silver futures market spiked from the then fixed silver price at $1.293 in 1966 to the $2.50 area in 1967. Subsequently prices retraced to a 1971 low near $1.29. The spike and full retracement may have been waves 1 & 2 in a five wave sequence that ultimately topped above $50. Our top inset chart above picks up at the 1971 low, shown as 2 circled, and continues through 1977. Our pattern interpretation is that the market made a wave 3 advance to its 1974 peak. From there an apparent 5-swing decline (annotated a,b,c,d,e) developed. At the c-swing low, prices bounced off a 50% retracement of wave 3. Renewed weakness saw the e-swing terminate marginally below that retracement level. From there we interpret a choppy 12-month recovery as an X-wave that led into a lengthy triangle, shown as another a,b,c,d,e sequence that terminated wave 4 (circled) in the second half of 1977.
The weekly gold chart interpretation (bottom inset chart above) starts at the 2001 low which we consider the end of Wave 2 that was nearly a full retracement of the Wave 1 spike from the 1999 low area. The 2008 decline following the wave 3 high is annotated a,b,c,d,e. At the c-swing low, prices bounced off the 38.2% retracement of wave 3 ($726 area) that was coincident with the $728 high made in 2006 (an instance of old resistance once overcome subsequently acting as support). Renewed weakness saw the e-swing terminate at $697 marginally below the retracement level. In our interpretation the rally from point e represents an X-wave that, at this time we view as being completed. What happens from here on out remains to be seen but we think that prices will remain range bound between the April and June extremes at $866 and $991 through most of August.
In summary, price action in the gold market since its March 2008 top just above $1,000 appears to be an ongoing consolidation. Monthly stochastic indicator values (14-period) shown on the page 1 chart are rising and not yet to levels associated with major price highs. Thus the long-term momentum backdrop is positive. Weekly stochastic values (14-period), not pictured, are in a sideways pattern but are currently slightly above the neutral 50 level and rising so the intermediate-term backdrop is neutral. Pattern repetition identified and discussed above implies the prospects for an impending end to the consolidation. The ideal time for an upside resolution would be in the first week of September. The long-term picture presented on page 1 and a long-term pattern comparison with historical silver action, seen above, suggest that the market may not be in a hurry to leave the broad range that is determined by the 2008 extremes at $1,017 and $696. Our view is that trading activity so far this year has established two key underlying support levels at approximately $904 and $866 basis the nearby Comex futures. If values, currently around $950, trade below $904 we would expect another test of buyers at $866. On a scale of 1 - 10 we are at a 6 as regards long exposure and would need to see prices below $904 to go flat. We would be more comfortable about major exposure on the long side of the market if prices break to the $866 area. Depending on market conditions in the latter part of August or early September we may revise our overall viewpoint.
Note: Some price points in the inactive front month Comex
futures may have been extrapolated from prices in the first active month and
may reflect either the official exchange session or electronic data that occurs
after normal exchange hours.
Charts constructed with Omega Research SuperCharts and Equis MetaStock
Data Source: Commodity Systems Inc.