For those of you who don’t know, Pixels.camp is the largest geek event in Portugal, the follow up to Codebits. I love the event and I have given talks in previous editions.
The major component of the event is the 48h hacklathon. Participants team up to make any project they want, building up to the 90s pitch. During the pitches, participantes in the audience are voting for the project with a like or dislike.
This is actually a hard task, as authentication is required (to prevent people who are not in the event to vote) and not all the projects get the same number of votes.
So far the organization has used the likes – dislikes as a metric for ranking projects. Marco Amado proposed using the ratio of ( likes – dislikes ) / ( likes + dislikes ). I do not believe this is fair because a project with 5 likes and 1 dislike, ratio of 0.(6), would rank higher than a project with 100 likes and 25 dislikes, with a ratio of 0.6. I would consider the second project to be better, even if only by causing more positive interest. Using just the difference is more interesting in my opinion.
However, I agree that it is a flawed system. I end up dislike every project that, according to me, shouldn’t win a prize. This decision is based on hunches, because voting is real-time and the first projects always end up having higher rankings because there is no comparison at the moment.
My suggestion is to change ranking to have two different votes: like and love. Likes would work like Facebook likes: they would be used to increase the project author’s ego. Loves would be used for the final ranking, unlike likes.
Each user would be given 100 votes, which would be equally divided across their loved projects. Then it is a matter of making the math and ranking the projects.
This approach keeps the simplistic approach of binary voting, it takes into consideration how many votes there are, and it has the advantage of not having dislikes, leading to an happier event!