the el farol bar problem, as james surowiecki explains in his book the wisdom of crowds, goes something like this:imagine that there are a certain number of people and on a particular day, they all want to go to a bar called el farol. the problem is that the bar is rather small and if everyone goes at the same time, it will be crowded and no one will have fun. these are the rules:
if less than 60% of the population go to the bar, they’ll all have a better time than if they stayed at home.if more than 60% of the population go to the bar, they’ll all have a worse time than if they stayed at home.
the other problem is that you have no way of telling how many people are going to go to the bar because all the people have to decide at the same time, whether they are going to go or not. social news sites, it seems, have an incredibly similar problem when it comes to content submission. if everyone submits at the same time and a large volume, some people are always happy and some people are always unsatisfied. however, if everyone submits in turns and in moderation, everyone gets a chance to have their content promoted and everyone is happy at least some of the time. let’s start with the following conditions (for sake of this example, let’s assume all the submitters and the voters are the same people and are all relatively equal power-users):
- there is a fixed maximum number of articles that can move to a site’s front-page on any given day.
- this number is without a doubt smaller than the total amount of interesting/cool/promotable content created on that day (never mind the preexisting content that will be found and submitted).
- there is fixed average number of votes that all users are going to cast on all the stories, and these votes are less than the total number of votes needed to promote all the stories, and are not uniformly distributed among all stories.
- once a story is submitted and not promoted, it loses its chance of ever being submitted and being promoted.
to put this in numbers, let’s assume there are 500 great stories, all of which will be submitted, and each of them needs on average 50 votes to be promoted. however, a maximum of 200 stories can be promoted, and there are only 100 users, each of whom will cast on average 100 votes (meaning that 10,000 votes will be cast, which if distributed uniformly, can promote all those stories). what we see is that because all users over extend (submit 5 stories rather than 2), the votes get divided in such a way that even less than 200 stories get promoted.
- if every user submits 2 stories, we get 200 stories promoted. the content producers are happy, the users are happy (for a perfect ratio), but we create a backlog of good content.
- if every user submits 5 stories, we get 150 stories promoted (because of the vote split the average is less than 2), and the group as a whole is less happier than it could be.
the ideal case
like in the ‘el farol’ bar problem, all parties would be better off if they took turns in participating and did so moderately. the easiest way to institute this would be to have a mechanism where everyone is limited to a certain number of submissions daily. the problem here is that this system would not only go against the basic principles of socially driven sites, but it would be much hated and would elicit an extremely negative response. the simple fact is, even if it hurts us, we love and value freedom of choice above most other things. so the system we want to find is one that comes naturally, from the community, through consensus between the individual users.
what actually happens
what happens is that users fail to coordinate their submissions with each other for several reasons. first of all, some users are simply unable to (networking issue) or don’t find it efficient to (time issue) to coordinate their submissions with other users. second, for many users it’s not in their best interest to throttle their submissions. we have to keep in mind that not all users are equal, even from within the good content, not all content is equal, and everyone thinks that they got it right and therefore the vote split will work in their favor. furthermore, you can never be completely sure that you made the right calls and there for submitting 5 rather than 2 is better for you because you get to hedge your bets.
in fact, i’ve tried to talk to several top users, many times, to try and come up with a system where we all don’t clog the system to the point where no one is having fun, but also get to participate to a point where we are having enough fun. one major problem we faced in trying to come up with the system is getting everyone else to participate. if 5 users participate and 5 others don’t, the 5 who decide to throttle their submissions lose out.
this post is a part of my journey through james surowiecki’s the wisdom of crowds.