Friday, May 30, 2008

Collective intelligence requires more than voting

Giles Bowkett has an insightful point on voting schemes at sites like Digg:
When you build a system where you get points for the number of people who agree with you, you are building a popularity contest for ideas.

However, your popularity contest for ideas will not be dominated by the people with the best ideas, but the people with the most time to spend on your web site.

Votes appear to be free, like contribution is with Wikipedia, but in reality you have to register to vote, and you have to be there frequently for your votes to make much difference. So the votes aren't really free - they cost time.

If your popularity contest for ideas inherently, by its structure, favors people who waste their own time, then ... the most popular ideas will not be the best ideas ... The people who have the best ideas, and the ability to recognize them, also have better things to do and better places to be.
Let's compare these voting schemes to prediction markets. To function properly, prediction markets provide incentives for truth telling and for gathering knowledge. In prediction markets, even ones using play money, you have to put something at risk to have a chance at a gain, and people usually only take that risk if they think they are likely to succeed.

In most of these voting schemes, you have an unlimited number of votes. There is no incentive to do research -- it takes more time to do that than to vote -- and no cost to voting on something and being wrong.

Simple voting schemes do not create the proper incentives to get good outcomes. As Giles points out, those with knowledge have no additional incentive to participate than those without knowledge. In fact, the people with the strongest incentive to participate most likely are those seeking to manipulate the system for some external gain, a problem we see on Digg.

Please see also my previous post, "Summing collective ignorance".

[Giles post found via Dare Obasanjo]

28 comments:

Benjamin Meyer said...

Completely agree. This is the reason why I believe that reddit.com _will_ fail in the long run to another news site that does not have this problem. Not overnight because of the large user base on reddit.com, but it will fail if it does not change.

Giles Bowkett said...

Exactly - this entire genre is a mass of consistently mis-constructed prediction markets, which provide the same set of counter-productive incentives. The incentives are good short-term but bad long-term so the decay always happens.

Anonymous said...

Agree! ..but... does it really matter? Most of us have already come to the same conclusion as you, but still peruse the site out of curiosity and there are many times it leads to something of value.

It's kind of like those trash magazines besides the cash in your local grocery market. If it wasn't for them, I would have never learned that Britney Spears at one point in her career shaved her head!

Thanks for the post

Anonymous said...

The idea that reddit will fail makes the assumption that some day redditors will be intelligent enough to figure out or even understand what you just wrote.

dr. nar said...

It's naive to assume reddit can't branch out into subcultures. The main reddit page will probably suffer as you describe, but reddit is capable of creating subcultures via subreddits AND the software itself can be applied to other communities like we see with http://lipstick.com/.

Anonymous said...

So, taking this a step further, reddit will become a community that will appeal to and be dominated by people with too much time on their hands.
Is this a problem with democracy in general?

emulatard said...

This isn't true because I don't want it to be true. I believe in reddit with the same fervor and absence of critical thinking that you might expect of anyone.

Bill said...

Completely disagree. As with most economically-based theories, it surmises that there must be a fair trade-off. This is patently absurd as shown by the creation of the content in the first place. Many people generate content because they want to see better content out there. As such, many people contribute to Reddit because they want to contribute. It's called the Benevolent Earth theory, and it applies here, I believe.

Good writeup though - well thought out.

Anonymous said...

Totally disagree. Most people on sites like these could care less about obtaining points. There are a select few who make this a priority but I would argue that a vast majority have no desire to chase points. Most go there to read a vast array of news/stories and give an opinion or two on the ones that interest them. If people don't agree, they just don't care.

sparky said...

I'm sure you're right, but so what?

Reddit having a finite lifespan does not affect its usefulness today. Just like you dying sometime in the future does not affect your usefulness today. Or the sun. Or anything.

It's a useful toy, a place to find "new and interesting" things on the web. Its future irrelevance doesn't stop me from using it right now to find fascinating articles like this one.

ohxten said...

Regardless, I always stay away from the reddit main page. It's mostly left wing/right wing conspiracy theories and tazer stories. Yeah, interesting.

I find the programming subreddit to be enough for me. :)

K. Woodward said...

I'll bite, but I think all this argument establishes is that 1) sites like Digg, reddit, etc. will fail or 2) there is room for more than one such site.

If we suppose that people will leave the sites because they don't have enough input -- I think that's what you're getting at, and I think it's a reasonable assumption -- it is possible that they'll abandon such sites entirely, but equally likely that they'll find a replacement good. Note the number of converts from Digg to reddit, "Oh, it's so much nicer here." Things be what they will, a site like reddit with lower traffic than Digg will feel more controllable to a new populace.

There are plenty more voting-based news sites. As a tyranny of those-with-too-much-time arises on one site, those with less will substitute to others. Don't have hard figures, but anecdotally I'd buy this.

Since switching costs are low -- practically non-existent -- there's nothing to prohibit this. I'll apply bill's mention of the Benevolent Earth principle: the users who are spending the most time voting aren't doing it to keep down the little guy, but to maintain the site as they see fit; this in mind, if the lower users leave, the upper users probably won't follow. Even if they did -- they're malicious griefers -- they'd have to allocate time across multiple sites, becoming low-level users themselves.

I think you hit on some good points, but I definitely disagree with the certainty of your conclusion.

chip33550336 said...

I disagree with the notion about subcultures solving the problem. It seems like you will run into the same issue in your subculture. I don't think a topic is enough information to maintain a collective intelligence.

Anonymous said...

I've got to agree with Sparky on this one, but I'll take it from a different angle. Reddit will fail if it stays static and unchanging. If they don't change then yes, they will enter TechCrunch's deadppol. I don't think they are going to let that happen though, as they seem to participate in the sites success quite a bit.

drawk said...

Unless you are a contrarian, in that regard when you are strongly flamed for an idea you know is true. You have a product on your hands...

Rob said...

What you (Greg) wrote is true, but not the whole truth. There are a number of other factors that cause the problems seen on Slashdot, Digg, and other "social" sites.

For one, there's psychology. If you accept that the participants of a social site can be divided into those that would contribute to its overall intelligence, and those that would drive the site towards a more maintstream focus, then as the mainstreamers start to accumulate, the smaller, more astute crowd starts to leave. It becomes a problem that feeds on itself; how many people do you know of that are really educated in some specific field would spend any appreciable effort on Digg, or Reddit? Or, even, Slashdot?

Once such a site goes mainstream, it loses the interest of the people at the farther end of the bell curve, and becomes average, driven by the same audience that the television shows and advertising pander to.

Anonymous said...

If we assume that individual vote weights will be normalized by the number of votes made by a user, is your faith restored?

-- GWF

Joshua Moore said...

Disagree. If we agree that subcultures are a good thing, then they can use tools like subreddits to disseminate information, and provide a set of common morals. In society we have religion and similar institutions to provide a common set of values to the society. On some level, these common values are important for a group of people to make a concerted effort of any kind. While main page reddit may be to broad to be useful for that purpose, which doesn't mean it's not just plain interesting sometimes, subreddits can help a group get something done, because they found a mechanism for agreeing on principles beforehand.

Alex Toronto CANADA said...

I agree about the cost of visiting digg and redditt.
I visit because I am curious about the new topics that come up and I really enjoy reading people`s opinions about them. There are a lot of very intelligent, witty people who comment. I have never bothered securing an account as I don`t care about voting and I don`t think it makes the sites any more interesting to me.
I am near middle age and I realize most of the people who visit are in their 20`s and 30`s.

feedblog.org said...

This is only true if you assume that every vote is equal.

If you apply a transitive trust metric this doesn't necessarily have to be the case.

You can count votes based on the total rank of the user balanced by their number of votes.

This way users who submit a story once but NAIL IT will be able to see more boost for a story than users who take the shotgun approach to voting.

You'd be amazed at what math can do to solve voting problems :)

Ryan Andrews said...

Simply put, you are wrong, sir.

If digg and reddit were judged by the sole "upvote (digg)- downvote (bury)= total score", I would completely agree with you. But then answer this: Why are some stories on digg that get 300+ votes not on the front page? Because the algorithm corrects this! Same with reddit. It is not ordered numerically, but rather with what is "hot" and fits the mold of the the algorithm.

Of course, it is a popularity contest, but as my 6h grade teach told me, "What isn't?" You don't get a job? It's because you weren't popular enough with the hiring staff. Don't get ladies? Yeah, try and tell me that's not a popularity contest. Look, you join reddit and digg because you believe that you have the right, and the ability, to support your ideas. While the community may not agree with them, they remain present in the forum of your choice.

You make the argument that reddit/digg costs time, but in all honesty, that is misleading. If you didn't spend the time on digg or reddit where would you spend it? it is like saying that creating job opporunities that include hard work and heavy lifting would reverse the unemployement numbers It is a paradox. If you didn't spend your time on social sites, you wouldn't be more productive, but rather waste time elsewhere.

How about this: Instead of critizing why these won't work, you come up with a better solution for a social news network. This means you must provide unbiased, uncensored, and unpartisan views on ALL subject matters. When you can do this, please contact me.

Kirk said...

And that, my furry friends, is why you need to apply views to the data. People with high comment karma should see personalized stories based on how insightful people think they are.

Group people into categories based on points per comment made. Then cache those ten different views so the server doesn't 'splode.

AaronJames said...

I can't see why so many people hate on reddit. I'm just the type of person where if I don't like something I won't use it and basically just don't regard it. I like reddit though, mainly for it's randomness I guess. Saying it will fail is absurd though, it's not like it's graved in stone, it's just code so it has the ability to adapt and evolve. That's like saying youtube will fail because so many videos are just unoriginal or plain not entertaining. It's all perspective really, those who like the site will continue to visit and browse through whats been posted regardless.

Anonymous said...

Great post Greg!

I was a Digg user during its infancy, and witnessed the gradual decline in its comment quality. Comments within Digg are now in a sad state of decay.

Reddit is following this same pattern, but it has far less users than Digg, so the comments have not deteriorated to the same levels.... Yet.

The reason I read the comments is not to vote on them, but to learn more about the subject of the article. Reddit still contains much quality in this area, although you have to filter through many junk-comments now to find the intellectual comments.

Intellectual and thoughtful commentary don't get as many votes as funny or sensationalist comments, so I ignore the voting numbers, and determine for myself what the quality comments are.

The voting system also places an emphasis on who is right, and who is wrong, or what the voter believes to be right or wrong. There is too much emphasis on whether someone agrees, or disagrees with the editorial, when fruitful dialogue often requires a neutrality by the participants, in order to evolve.

I still have much faith in many of the "commentors"; every day I am impressed with the thoughtfulness and awareness of many. But it is worrisome to see the pattern of such comments, and how it threatens the pursuit of quality, truthfulness, and objectivity in new media.

This comment would be downvoted on Digg, disagreed with on Reddit, but hopefully it adds something positive to your discussion.

David, Canada.

Rajiv Das said...

The main issue from the point of view of the user is what's in it for me.
Either you make the incentive explicit, like giving users points or you keep it implicit. Conventional wisdom shows that, the more degrees of freedom the more likelihood of noise. But we want noise, we want people to speak, that's the spirit of Digg.

Has there been any study done of effectiveness of various incentive schemes in such open voting systems?

satts said...

Greg,
Can we attribute sustenance of platforms like Digg and Reddit to them being open and without any incentive schemes.

Google Answers wanted to catch the quality , by attributing a monetary tag, where as Yahoo Answers simply removed that and kept it simple.

Answers are better than many other platforms.

george tziralis said...

So, Greg, are you proposing a market-based app as a digg-killer? Hmm, interesting... :-)

Anonymous said...

So, why are you all putting well thought out comments here? What's the payoff? ;-)