sorting by popularity

When you sort items by popularity (or size, or ...), the most popular ones (or biggest, or ...) always come first, followed by the less popular ones.

Is that all that "the long tail" is about ?

Or is it about the following counter-intuitive fact: ? Sometimes (not always), the small rocks outweigh all the big rocks put together, even though any one big rock outweighs any one small rock.

For example, How Long Is the Coast of Britain? Statistical Self-Similarity and Fractional Dimension.

The article mentions Lots of energy was released by the earthquake of December 26 2004, but there are tiny earthquakes all the time; most earthquakes are part of the long tail.

Yes, but can we go even further and claim that all the tiny earthquakes, put together, release more energy than the big earthquake? Is that true ?

--DavidCary 19:24, 11 Jun 2005 (UTC)

This statement in the article expresses a myth rather than a fact about the Long Tail: "In many cases the infrequent or low-amplitude events—the long tail, represented here by the yellow portion of the graph—can cumulatively outnumber or outweigh the initial portion of the graph, such that in aggregate they comprise the majority." It is almost never true in any real world distribution. In these situations an 80-20 or Pareto distribution applies: 80% of whatever comes from 20% of the users. Likewise, 80% of the users cumulatively account for only 20% of whatever it is you are measuring. The Long Tail represents opportunity in some sense, but it is a total misrepresentation to suggest that the Long Tail represents more opportunity than going for the masses.

--underalms 03:15, 23 Feb 2006 (UTC)

The more I think about it, the more I am convinced that this long tail effect is nothing but the next "paradigm shift" and has no meaning independent of the book being written. We seem to be suggesting that there are situation, as underalms noted, where the tail has more probability mass than the area around the center. But that's pretty much nonsense in most satistical distributions created from real world data. In the business sense, the long tail argument boils down to nothing more than economics of scale. If your example companies were truly boutique retailors who refused to serve the center of the market and were successful, you might have a point. But Amazon or Netflix aggresively pursues the entire market, the only reason they can offer more selection is because of their cost structure. Netflix doesn't offer you the chance to rent "Ernest Goes To Camp" because the collective demand for all the Ernest movies excedes that for "Batman Begins" but because they store many times more titles due to their lower cost and that allows them to offer titles that have very low demand. -BennyAbelard Aug 25 2006.

I dont know why longtail woulndt be right in many cases. I mean 80% is 4 times 20%. Lets imagine some store with 100 items, the least sold 80 items would just need to sell at a average at least 4 times less than the top 20 ones and long tail theory would be true. So you have 4x more itens being sold and so can sell until 4x less itens at average and still have an higher amount of itens sold. PS:One place that the long tail is not true is the music industries. In fact the amount sold by the 90% least sold releases (albums, singles, eps....) is problably smaller than 9 times the amount sold by the most sold 10%.187.115.234.40 (talk) 12:40, 31 January 2013 (UTC)

What is the exact technical meaning of long tail?

The article tries to explain it by saying: A probability distribution is said to have a long tail, if a larger share of population rests within its tail than would under a normal distribution.

What is that supposed to mean exactly? I am a bit confused because a normal distribution is after all exponential distribution. So technically, until you reach infinity, there would be some value corresponding to each x-value. So wouldn't it have same share of population as any other distribution? — Preceding unsigned comment added by Methosoldest (talkcontribs) 13:47, 29 May 2013 (UTC)

The normal distribution indeed goes all the way to infinity. The share of the population in the tail is the area under the curve from some x-value all the way to infinity; this area is finite. In some distributions the finite area in the tail is greater than for the normal. Duoduoduo (talk) 15:10, 29 May 2013 (UTC)