Weighted Ranking Methodology for Putvote

Well, this is only a suggestion and is left to the creators to implement it if it pleases them.

Disadvantages of the current ranking system:
1. There is no real algorithm involved. It is sorting only on basis of votes received.
2. Those who do not receive votes but put a lot of votes suffer in the system.
3. There are higher ranked voters who have not submitted as many articles.
4. The ranking sometimes does not differentiate between a more prolific user and a less prolific user of the portal.

Putvote essentially is a portal that is enriched by participation or collaboration. A community uses this portal to share bookmarks and people who feel like promoting a link can vote for it. In such a participative model, it is sometimes essential to ensure that ranking of participants is based on a weighted model rather than a simple sorting of one parameter.

I have come up with a weighted methodology for putvote. This method has the same basis as that is prevalent in the ranking of colleges/schools. With this method, the ranking is more rational. A participant should engage in higher participation by putting votes and submitting more articles. The better articles they post, higher the votes that they receive. So, a participant that only receives a lot of votes but does not put as many articles will drop in ranking. A participant who submits a lot of articles and puts a lot of votes will increase their ranking.

For the purpose of calculation, the weightages assigned as as follows:
For putting vote: 0.35 or 35%
For receiving vote: 0.35 or 35%
For submitting article: 0.30 or 30%
Totally, it adds up to 1.00 or 100%. The idea behind assigning 0.35 to putting and receiving votes is that both are equally important in the foundation idea of the portal. The idea of assigning slightly less value to submitting an article was to ensure some participants just do not spam the portal with articles.

So, the formula that I have used is:
Weighted score = weight for votes received x votes received + weight for votes cast x votes cast + weight for articles x articles submitted.

Proof of the pudding lies in the eating, eh? Here's some for you. Only first 40 ranks are illustrated minus "guest" as I have not considered "guest" as a rankable user. It is too wide a population in comparison to dedicated users.

Here it goes...


(click above pic for a better view in a new window)

Some places, the ranking assigned is joint. Not shown here, but shoefiend and krish share the 38th place and there is no 39th place. And for some users the ranking jump is huge, the biggest being ang with 26 places. The biggest slide is for munnacircuit with a drop of 15 places. And if you observe, the top 10 rankings remain the top ten. :-)

So, what do you think?