I rarely go to paper presentations. I can read the papers, so why bother? OOPSLA is a five-ring circus, and there are more important things to do, like go to demos, go to panels, and talk to people in the hall. I'm happy some people like to go to presentations, but I am not surprised when people don't.
When I give a presentation, I am faced with the fact that there is no way to cover the material in the paper in 20 minutes. I don't try. I instead give the main point of the paper and try to convince people to read it. But since I already intend to read the papers, I don't need to listen to someone try to convince me to read them.
So, my point is that the actual paper presentations are not the main point of OOPSLA. It is an important point, but there are lots of important points.
Peter is right that OOPSLA papers are usually not interesting to the average attender. I claim they are not even interesting to the average person on the program committee. Details of Java compilers are not interesting to anybody except those writing compilers. Of course, we all want better compilers, so we want to encourage people to work on those areas, which means we need to have a place for them to publish, but we shouldn't expect OOPSLA attendees to want to listen to presentations on compilers. As far as I can tell, this has been true for the past 15 years at OOPSLA. So, what happened in the last few years that is different?