Editor's note: This is a special bonus column from Jon Markman, whose commentary usually appears on RealMoney. If you'd like to see all of Jon Markman's

RealMoney

commentary, click here for information about a free trial.

There must be something in the air, because satellite radio seems to turn legions of presumably sane investors into legions of raving fanatics. Under fire from sellers and analysts recently, some

Sirius Satellite Radio

(SIRI) - Get Report

partisans draw strength from their Rasputin, Howard Stern, as they unleash a torrent of invective toward sellers and skeptics. In the past day, since

writing negatively about Sirius shares, my email has been flooded with nasty bile from the radio dial.

This sort of extreme loyalty is typical of the middle stages of a crazy momentum run. The stage can run for some time, so I wouldn't discount the possibility of a move back to a share price of $8, $9 or more in the short term. But the fact remains that all parabolic advances in stocks end with parabolic declines. It is one of the few immutable laws of the investment universe.

The disconnect between Sirius and reality appears to run along two key vectors: The stock has a low absolute price, and it represents a popular consumer product. That's a fantastic combination in some situations, but poisonous in others.

A Bumpy Ride

Let's dial back a bit. Back in the early 1990s, Sirius traded in the low single digits for a couple of years as it worked on getting its radio-beaming bird in the air. The stock ultimately ran up toward $70 at the apex of the

Nasdaq

bubble in March 2000; then, after a number of missteps, it fell all the way back to less than 50 cents in 2003.

At its low last year, debtholders who had been handed stock in the aftermath of a busted convertible bond deal couldn't sell the shares fast enough. Their panic, combined with legitimate concerns about the company's survivability, created the bottom.

As the stock was coming off the bottom in early 2003, it made a great value play -- and I said so in a couple of columns. The company had been recapitalized and had enough cash to get the hardware built and the programming on the air. The service was awesome. Still, the stock mostly went sideways for a year, hanging around $2-$3.

All that time, Sirius kept diluting shareholders by issuing more stock. Management ultimately took the share count up to 1.26 billion, which is way off the charts for a company with only $47 million in trailing 12-month revenue. Management used the company's cash kitty to bring in big sports deals and rebel radio personality Howard Stern, but the revenue and income to justify these payments is still far, far in the future -- even if the company gets 10 million to 15 million subscribers.

What Sirius has done, in effect, is perfectly emblematic for our overleveraged times: It did the corporate equivalent of a home refinancing to do a remodel, only management took money from the gullible public, not from stern-faced bankers.

Many of the Sirius fanatics who were buying the stock at $8 or $9 don't seem to realize that a share of stock represents something much more than a fan's vote for a favorite product. It represents a share of future earnings. The "real" price of a stock is the number by which you multiply a future earnings estimate in order to get the price. In this case, there really are no earnings expected in a reasonable time frame, so investors ratchet back to multiply a revenue number instead.

Through the long history of the stock market, investors know that there is a range within which earnings and sales multiples fit. On the low side, even a good retailer will sell for around 1 times sales. On the high side, a good technology company with strong growth prospects will sell for around 10 times sales.

And there are extreme exceptions. Right now, for instance, cellular-service king

Research In Motion

(RIMM)

, which makes the popular Blackberry devices, sells for 18 times sales. Go all the way up the line and you'll find that

eBay

(EBAY) - Get Report

sells for 25 times sales.

But these really are exceptions, and the companies and stocks have gone through hard times to get there. We have a historical record for the latter two stocks' sales multiples, and eBay's average high price-to-sales ratio in 1999 was 72, while Research In Motion's was 112.

The stocks were ultimately successful. Their shares fell 65% and 85%, respectively, off the highs set at those price/sales multiple levels before value was rediscovered. In time, real sales caught up with the hype and they grew into great companies and great stocks. (Even now, both have far fewer shares outstanding than Sirius -- 15 times fewer in the case of RIMM and half as many in the case of eBay.)

My guess is that similar compression is ahead for Sirius. In time,

Ford

will come to its senses and add Sirius' radios as a factory install option. And eventually the company will get the subscribers that its partisans expect. But in the meantime, there will be inevitable disappointments and the shares will settle back toward $4.50 to $5, or worse.

The great Fidelity portfolio manager Peter Lynch was best known for the notion that you should buy shares of products that you like. But he made most of his money, by far, for the Magellan fund in the stocks of boring old regional banks.

Jon D. Markman is publisher of StockTactics Advisor, an independent weekly investment research service, as well as senior strategist and portfolio manager at Pinnacle Investment Advisors. He also writes a weekly column for CNBC on MSN Money. While Markman cannot provide personalized investment advice or recommendations, he welcomes column critiques and comments at

jon.markman@thestreet.com.

Interested in more writings from Jon Markman? Check out his newsletter, TheStreet.com Value Investor. For more information,

click here.