Fifteen months ago we wrote about the "Darkness of 2006," our forecast of the impact of a pronounced economic downturn in the United States, caused by trade imbalances, government deficits and rising consumer debt, as well as high oil prices and an expected slump in the real estate market.

Were we wrong, or just early? That's open for debate, as housing prices have started to decline in some parts of the U.S. and economic weakness did encourage the Federal Reserve to stop raising interest rates.

But overall, the U.S. economy has chugged along merrily: unemployment remains very low, foreign lenders continue to happily lend to cover the deficits, and the stock market has forged ahead, taking the Dow Jones Industrial Average to a new all-time high.

One prediction that we did get right (darn it!) was that our own portfolio would likely suffer. Back in January 2005 it contained 25 stocks, every single one of which was up in value. Of the current 18 companies, 12 are up, or 66 percent.

The response to this state of affairs, as gleaned from our mailbag, has been interesting. Some of our newer subscribers are feeling somewhat shell-shocked as they survey their results.

But many who have been with us for a while are apparently blithely unconcerned, as they have harvested some nice gains along with us this year with the sales of such oldies as Claude Resources, Hudson's Bay, Stride Rite and Xanser.

So, are the pessimists a bunch of nervous Nellies who are unable to put a short-term gyration into perspective? Or are the optimists a crew of Polyannas who just don't get that line on the front page that says, "Past returns are not necessarily indicative of future returns"?

When we consider our historical record, the examination always starts with our 10-year annualized return, which best encapsulates whether the Contra methodology is working or not. When people complain that the portfolio has not performed well since January or April, we are unperturbed, though we do acknowledge that it is always better to be up than down in any time frame.

Many of our readers ask questions about more specific details on our performance, so the archive was opened and loaded into the mother of all Excel spreadsheets to try to come up with some answers.

All stock sales since January 1996 were examined. For many companies, there were multiple sale points; in these cases, only the initial sale was considered -- with the exception of firms sold and then bought back after the mandatory 30-day waiting period for taking capital losses. In these cases, each holding period was counted distinctly.

For the record, this approach increased the number of losing transactions, and included companies that were eventually nicely profitable the second time around; we think this policy provides a more balanced picture of our success rate.

The screening process left 93 sales for us to study. Without further ado, then, here are some of your questions and the results.

What percentage of your picks are winners?
Forgive us if we seem Clintonian, but that depends on how you define a "winner." By the broadest measure -- simply those we made at least some money on -- 71 percent qualify.

More meaningful, however, is the percentage of stocks sold after appreciating by at least 50 percent -- a key element of the Contra methodology; 59 percent cleared that hurdle. And 42 percent at least doubled in price.

What percentage are sold for a loss?
Twenty-nine percent went out the door on the negative side of the ledger -- a higher ratio than we expected to find.

What percentage reach the initial sell target?
At 41 percent, this was lower than we anticipated. Paradoxically, one factor that knocked this ratio down was our success in acquiring stocks that later became takeover targets.

Buyouts have been highly profitable for us: Dominion Textile, Noma, Royal LePage, Minolta-QMS, CompUSA, OfficeMax, Leitch, Sodisco-Howden, FCA, Gulf Canada, General Host, and Hudson's Bay were all lovely plays that offered up capital gains of between 75 and 334 percent.

Yet all of the bids came before we had held the stocks long enough for them to get within striking distance of our targets. If we add these gobbled-up stocks into the mix, the success rate increases to 57 percent.

What percentage are wipeouts?
Overall, 9 percent have this nefarious distinction. Losing most of our deployed capital on an investment is something we strive to avoid, and progress has been made in this area: all cases in which we lost more than 90 percent of our investment occurred in the period from 1996 to 2001.

In the last few years, the only really big loser has been Stelco, cast off in 2003 after a loss of 67 percent. However, Fonar is looking mighty ugly right now, to be sure.

What percentage gain does the average sell target represent?
Our initial sell targets vary widely, depending on our assessment of the potential appreciation. On average, targets were 245 percent higher than our buy point; the lowest was Dynex in 2001, at 48 percent, while the highest was Markborough in 1996, at 1,700 percent.

Neither made it while we owned them, but Dynex has done outstandingly well since our sale. More patience would have helped on that one.

Are lower initial sell targets achieved more frequently?
To determine this, all of the companies were divided into quartiles, from those with the most modest targets to the loftiest. The results were surprising and somewhat mysterious.

The first quartile, ranging from a hoped-for gain of between 48 and 112 percent, had a success rate of 46 percent. One would expect that these targets, being the lowest, would be easiest to achieve, and though the rate was better than the overall average, it was still less than a coin flip.

The second group, ranging from 113 to 186 percent, had a much poorer achievement rate of only 31 percent.

But it was the third quartile that really shone: with optimistic targets ranging from 187 to 298 percent, the success rate was 57 percent.

For the fourth bunch, in which we were shooting for between 310 and 1,700 percent, only 30 percent cleared the bar.

Perhaps this is just a statistical fluke; if not, it is difficult to reckon what this says about the methodology. There seems to be a sweet spot of expectations in the triple-to-quadruple range that, for some reason, the methodology is most efficient at identifying.

What is the average holding period for a Contra stock?
This is a little trickier to ascertain, as some stocks are bought multiple times, as well as sold off in chunks.

To simplify the calculations, only the time of the first purchases and first sales were used. The average holding period for the 93 positions was 3.5 years. For stocks that achieved the initial sell target, the holding period was longer, but not by much, at 3.9 years.

Some were with us seemingly forever -- Mitel was in the stable from 1987 to 2000, while High Liner, a holding since 1993, is the current "iron man."

What's the longest period a stock spent as a loser before eventually hitting the target?
Looking over the data, it's quite startling how many companies spent long periods as dogs before becoming swans. The most noteworthy transformation was probably Intertan.

Purchased in December 1996 at $5.75, it sat forlorn on the Contra Buy list for the next eight issues, at $4.12, $3.88, $3.75, $5.63, $5.38, $5.50, $5.25 and $3.56. In 1999, it not only hit its target of $9.63, but blew through it, resulting in sales at $11.25, $16.19, and finally $21.19.

What's the biggest percentage of losers that the portfolio has had?
The worst point for the portfolio was reached in October 2001, when 19 of 26 stocks, or 63 percent, were lower than what we had paid for them. However, in Contra fashion, while that sounds gruesome, the portfolio had actually done very well, with numerous stocks popping in value and being sold.

By July 2002, only 17 percent of our stocks were below their purchase points. In terms of our overall returns, the period from January 1, 1998, through December 31, 2000, was downright ugly, as the portfolio dwelled in negative territory. That was three long years of crud, folks!

The techies had a good laugh at our expense during much of that time. It was also a period when our subscriber base declined.

During the dark days of January 2001, we wrote that our portfolio was better than it had been in years. And it was: the next three years saw spectacular returns. And when we later wrote that our portfolio was not nearly as good anymore, people continued to sign on, and have been surprised that returns were not so stellar.

Conclusions
The key to all of this is not to simply assemble the data, but to analyze it. During this "information age," time, effort and money are wasted on conducting research and compiling reports -- which are filed neatly away and forgotten. And then the cycle repeats. Even individuals are now so overloaded with information that it becomes overwhelming to synthesize and utilize.

The question is: what can we do with the information that came out of this exercise? One move is to pay additional attention to that third quartile -- which isn't to say that other levels of returns will be ignored.

A second resolution is to try to choose our buys more effectively. The 29 percent that depart the portfolio at a loss seems too high, and makes us question whether we might be even choosier than we are. And, given that several of the losers we did not repurchase went on to do very well after our sale, we may need to be even more rigorous about deciding when to give up on a company.

It should be noted that, while the ins and outs and ups and downs and variations on themes are interesting to contemplate, at the end of the day the numbers that still speak the loudest and clearest are the overall financial returns.

Fortunately, over the long term, our returns stack up with the best in North America. That makes us proud. However, years such as this one make it clear that we cannot afford to spend too much time polishing our trophies; instead, we would do well to ensure that our thinking caps are well fastened.