Recommending research articles

Challenge for AGT researchers: come up with a good mechanism to help the most noteworthy papers get the most notice!

As a start, Noam Nisan is trying an experiment to get people to recommend interesting papers in Algorithmic Game Theory/Economics. Journals no longer serve well in disseminating results quickly – is it possible to dispense with them altogether by using another way to give a signal of quality and noteworthiness?

I have doubts that this particular proposal will succeed, because of incentive compatibility problems. When we cite a paper when writing our own, it is usually taken as a recommendation (negative citations do occur, but relatively infrequently I guess). If we don’t cite it, a referee or the author may notice and our reputation will suffer. If we do cite it, our paper is probably more likely to get published. So there are some incentives to participate in that form of “recommendation”. However, recommending papers via Noam’s proposed scheme doesn’t seem to me to have as strong an incentive to participate. Who has the time to do it, if the recognition obtained for good recommendations is so low? Still, it’s good to have a first step, and I am sure with enough time and money a nice system could be devised that cuts out the for-profit journal publishers.