Seeing as how it is of interest, if I may trouble the crowd here with yet another summary, "The Wisdom of Crowds" (which is actually sitting on my desk, next to "Good to Great") says that to have a "smart" crowd, these conditions must apply:

- Independence - people don't look at one another's answers prior to submitting their own.

- Diversity of opinion

- Decentralization

- A way to aggregate the results

The independence thing is the problem with digg/reddit/yc.news etc... - instead of a bunch of people voting on their own, and then seeing what comes out, you see ahead of time how everyone else has voted, and reddit groupthink ensues. News.yc is better because of the high quality of the people, not the algorithm.

Oh, the book is at:

http://www.squeezedbooks.com/book/show/23/the-wisdom-of-crow...

But I haven't finished the summary yet... contributions welcome:-)

I think it's ridiculous to claim that this book represents the ultimate truth in the subject.

This subject is in its infancy indeed.

No, but those conditions make a lot of sense to me in terms of differentiating a "wise crowd" from an "angry mob".

I was referring more to the article's statements than your comment.

I totally agree. Digg, Reddit, and "Hacker News" for that matter, don't actually implement the "wisdom of the crowd". The true hypothesis of a WotC is based on a random survey or sorts. How many people actually look beyond the first few pages of these social news sites, or maybe their friends (or other reputed users) links. Social news is still in a stage of infancy, much like how web search was (maybe is).

Throw 10 random stories to the user on the homepage, let them vote, and the best stories will emerge. What's point of actually showing them the "popular" stories as the first go.

The reason Reddit and Digg are implemented the way they are is that they rely on people's need for attention to generate a constant stream of incoming links. Making it to the "front page" of whatever site you're dealing with is a badge of honor and indicates some sort of social significance. The incentive's just not there for people otherwise. The only people who'll submit links without any guarantee that anyone will see their submission should be using del.icio.us.

There's no "guarantee" that links submitted to reddit or digg will be seen by a significant fraction of people, either.

As for the submitter's need for popularity, I think the basic problem is that you need to find the right way to balance voting on stories (which should be more or less independent of a story's popularity, as the article argues) with viewing stories (which must be inherently based on some notion of popularity or story quality, or the whole point is lost). That seems like it should be solvable, though: for example, you could present both kinds of items in the UI.

There are some good ideas here, but bing a facebook app Wildfire undoes the analysis on this blog, because then you still create a closed loop of readers.

It's basically like automated email forwarding on facebook, and email forwarding already operates like wild fire, when things are really interesting.

Yeah, the fact that it's an FB app was a major turn-off to me as well. What does building it on Facebook give you, other than significantly limiting the set of potential users? Surely, building a social news site from scratch can't be that hard...

"We tend to associate with people who have similar tastes and views to us"

There has to be a better word for this than 'homophily.'

You have a major flaw in your logic for Wildfire: the fact that "we associate with people who have similar tastes" does not imply that "we associate ONLY with people who have similar tastes". I'm interested in startups, but a few of my Facebook friends are. If I bombard them with startup news, most of them will be annoyed.

I'm amused by the implication that The Wisdom of Crowds is a Web 2.0 textbook. It's like "obviously this is not the wisdom of crowds, have they not even read the book?!" I'm not sure that's what they're even aiming at. For my part, I just want a few interesting news stories.

I remember in the early days of reddit some people claiming reddit/digg were a good idea because they used the WoC idea, and thinking that wasn't really correct. I've always thought a lot of the problems with social news could be solved by following the WoC model more closely (ie: it would be a better way of getting "a few interesting news stories"), but apparently I'm one of the 20 or so people who isn't hip to facebook, so I don't know if I'll be trying wildfire.

That's a fair comment. I must admit that it's specifically Kevin Rose who uses that term a lot.

John.