Contrarians were not surprised by the recent weakness, as it has been apparent to them for several weeks now that gold timers were too optimistic, providing a textbook illustration of the notion that bear markets like to decline a slope of hope.  See July 29 column  See Aug. 6 column for Barron's Online

Unfortunately for the gold bulls, it's not clear that enough gold timers have thrown in the towel to convince contrarians that a bottom is at hand.

Consider where the Hulbert Gold Newsletter Sentiment Index (HGNSI) stands. It reflects the average recommended gold market exposure among a subset of short-term gold timing newsletters tracked by the Hulbert Financial Digest. As of Monday night, the HGNSI stood at 5.4%, meaning that the average gold timing newsletter is recommending that clients invest 5.4% of their gold portfolios in gold and gold-related investments.

The good news, from a contrarian point of view: This 5.4% is a lot lower than the 64.3% level to which the HGNSI soared in early and mid July. The bad news is that it is not even lower.

For example, the HGNSI dropped to minus 10.7% in the correction that ended in early May, or 16.1 percentage points lower than where it stands today. Even so, gold then bottomed out at above the $850/ounce level, some $25 per ounce higher than where it stands today. That difference betrays now much more inclined the average gold timer is today to see the glass has half full rather than half empty.

Another comparison that leads to the same conclusion: During gold-market weakness in early April, the HGNSI dropped to as low as minus 17.9%, or more than 23 percentage points lower than where it stands today. Even so, bullion at that time was trading above $900 per ounce, or some $80 per ounce higher than where it closed Monday.

All of this suggests to contrarians that more gold weakness is in store.

What would lead contrarians to conclude that a bottom was finally at hand? Coming up with a single answer is difficult, since different contrarians have different thresholds that would trigger a buy signal. But, at a minimum, the HGNSI will probably need to drop into negative territory. That's because virtually every intermediate bottom in the gold market in recent years was accompanied by HGNSI levels below zero.

It's impossible to predict when that might happen. It might be, for example, that a couple more days of weakness will lead to capitulation on the part of the typical gold timing newsletter editor. But it also could be that a dead-cat bounce in the gold market leads to renewed hope on the part of the average gold timer, thereby postponing that bottom.

In any case, we're not there yet.