It's a tradition that started back around the time SciFi Channel canceled "Farscape." Many fans were confused why SciFi Channel pulled the plug on the show, when it was one of the network's highest-performing programs.

I got a lot of e-mail during that time, many of them asking why SciFi would do such a thing, and what is the magic ratings number a show needs to survive.

In answering that e-mail I wrote my first of what seems now like an almost annual tradition of discussing how the television ratings system works. It's gotten to the point where I probably just need to do this at the beginning fo every television season, because I see time after time where people question or simply misunderstand what television ratings are about, and how to tell if the numbers are good or bad.

Television ratings, quite simply, is a qualified estimate of the number of people tuned in to a particular program. The company best known for providing such information is Nielsen Media Research, a part of what used to be a Dutch company known as VNU which recently renamed itself The Nielsen Co. Nielsen uses demographic sampling -- similar to what is used in Presidential polls, but on a much wider scale -- to get an idea of who is watching what and when. "Nielsen Families," either jot their viewing habits down in a paper diary, or more and more, Nielsen is introducing electronic means of chronicling individual viewing habits including the use of a device known as a "People Meter."

The information is aggregated and analyzed at a rather large facility in Oldsmar, Fla., literally several miles from where I'm writing this column. It's not a perfect system by any stretch of the imagination, but Nielsen has been working hard to try and make it as accurate as possible in determing numbers.

For most television shows, Nielsen issues two viewing figures. The first is known as a rating, the second is known as a share. The rating is nothing more than a percentage of the number of households in the United States where there is a television. Each year, Nielsen estimates the number of households with a television set, and this past year determined that number to be 111.4 million. Each rating point is literally 1 percent of that household market, which means that 1 rating point essentially represents 1.114 million households.

A typical episode of "Heroes" receives a rating of about 8.4. To get an idea of what exactly that number means, simply turn it into a percentage, remembering that the base number is 111.4 million. So how many households were tuned in to "Heroes" on this particular night? We take 8.4 percent of 111.4 million, and find out that 9.4 million households watched Claire Bennet connect with a long-lost family member.

Does that necessarily equate to 9.4 million viewers? No. Obviously, some households may have more than one person watching, and how many usually depends on the type of demographics that it has. If this were an airing of a Harry Potter movie on NBC beginning at 8 p.m., chances are you may have a man, woman, and a child or two watching. If it's David Letterman on CBS, it may just be one person.

Many times, publications offering viewership will provide it by number of actual viewers, not because Nielsen didn't compile a rating, but because it's easier for the general public to digest 12 million viewers, rather than reading and having to be explained what a rating of 8.2 means.

The second number that normally appears with a rating (usually when dealing with a broadcast network) is what is known as a share. "Heroes" could be listed with an 8.4/12, with "8.4" being the rating and "12" being the share. While the rating gives a percentage of the number of households overall in the United States being tuned into a show, the share gives a percentage of the number of households whose television sets were turned on who were tuned in to a particular show.

Friday nights at 8, you may find my house empty because I'm at religious services. Fridays was a re-airing of "Smallville" on The CW. The rating, which is a percentage of all households, will include me (hypothetically) in its final number, even though I was not even home. However, I will be completely left out of the share, because in order to be there, I would've had to have been home with my television on.

For "Heroes," 8.4 percent of the nation's households may have been tuned in to see the show, but of all the homes where people actually had their television sets turned on (and not reading a book or playing video games), it was actually popular with 12 percent of that audience.

I know what you're saying, "I get the basics, but what the heck does that have to do with my favorite show being canceled?" Ratings are compiled not as a bragging match between networks, but as a way to show potential advertisers how many people are viewing a show, and to help network calculate how much a 30-second airtime of commercial space is worth. Obviously, a show that draws 12 million viewers will be more valuable for advertisers to reach consumers than one that pulls in 2 million viewers.

The numbers more or less give advertisers an idea of how many eyeballs will be on their ad, and what kind of reach they can experience by buying time during that program.

Nielsen in the past year has stretched to include "time-shifted" audiences, those who may not have been home during the show's live broadcast, but taped it to view later. Time shifting has typically been ignored by ratings in the past, simply because those who might DVR a program typically will fast-forward through commercials. So while they might be yet another viewer of the show, they are not necessarily viewing the ad content.

That has changed with more and more product placement in shows, and networks demanding to know the actual popularity of shows, and not just the number of people viewing ads.

There is no magic number for cancelation, at least not a single number that can be applied universally. Obviously, the more viewers who tune in to a show, the more a network can charge for ad space, and the more money it can make. If a network isn't profitting enough from a show, and the powers that be think another program can do better, they will cancel the show and put the new one in its place.

In this day and age, there is much more to determining the fate of shows than just ratings, so even if a show has poor ratings, it doesn't necessarily mean it is going to be pulled off the schedule. "Studio 60," for instance, has not done well in overall viewers for NBC, but it has attracted audience demographics with high amounts of disposable income, an advertiser's dream. Other shows, like "Star Trek: Enterprise," may get a renewal for another season desite not meeting ratings expectations, mainly because of other purposes, like in "Enterprise's" case on a fourth season, to make it a more viable syndication package that the production company can sell.

Ratings can be rather complex and easy to confuse. But remembering that the numbers are nothing more than an estimated percentage of households, whether it has televisions on or not, seeing figures should help give you a better idea of understanding the kind of broad audience a show is getting, and maybe where it might be lacking.

For more information about different variables involved in deciding the fates of shows, see my recent SciFriday column, Why 'Battlestar Galactica' Won't Get Canceled.

I am by no means an expert in ratings, and it has taken me a considerable amount of time to understand them myself. But I hope by imparting some of my knowledge, you can have a glimpse of one of the main variables in the survival or death of your favorite television shows.

Michael Hinman is the owner and site coordinator for Airlock Alpha, writing out of Tampa, Fla. He can be reached at mhinman@airlockalpha.com.

Catch his regular column, SciFriday, each week right here at Airlock Alpha.