I recently ran across this link to acceptance rates for papers in graphics conferences. The SIGGRAPH chart has some missing years (including the first four), presumably because data was not available. Graphing the trends yields some interesting information:
Excluding years before 1985 (when the conference was still “finding its legs” and acceptance rates were very high), the acceptance rate has hovered between 14.9% (1998) and 23.7% (2007). The long-term trend appears to be that the acceptance rate is flat, and the number of submitted and accepted papers steadily increase. In the shorter term, submitted papers appear to be flat or even declining after 2003, with accepted papers following suit (2009 has the lowest number of accepted papers since 2002). I’m not sure why that is; a 2003 flattening seems too late to be attributable to the dot-com collapse and too early to be related to the big graphics conference restructuring of 2008 (where Eurographics was moved to spring and SIGGRAPH Asia was introduced). If anyone has a good guess, please leave a comment.
I didn’t bother graphing the other conferences. The Eurographics table only has information from 1998 (the conference has existed since 1979, only five years less than SIGGRAPH). From 2002 on the acceptance rate has been similar to SIGGRAPH (before that it was significantly higher). The I3D table is pretty complete; it shows consistently high acceptance rates, between 25% (1999) and 42% (2008). Graphics Interface and EGSR (EGWR in earlier years) have similarly high acceptance rates.
Notice that post-1999 there was the first multi-year downward trend in submissions since pre-85. This generally follows the peak in attendance of the conference, which I think peaked in 1999. I think that trend has lots of reasons, a combination of the slow fading of “hotness” in the field combined with the value of a conference in the internet age.
So the real question is no so much why the downward trend the last few years, but where on earth did the huge rise come from from 2002-2004? 2002 the first post-boom years for SIGGRAPH. Post-9/11 lots of people didn’t fly and 2002 was a small conference. So why the record submissions? Perhaps in times when it’s harder to get there (fewer companies paying, etc), more people submit hoping to get a free trip? And now that they are all submitting every year, we’ll go back to the post-1999 slide but from a bigger number?
I think the reason why 2009 saw the lowest number of accepted papers in a while is the desire to spread the high quality papers more evenly across the big three conferences. This particularly applies to Siggraph Asia which still needs to establish itself as being on a par with Siggraph and Eurographics.
There is also a growing trend to include papers from Transactions on Graphics in the Siggraph schedule, so that the number of presented papers in 2009 will probably be similar to 2008, although fewer of them come from the Siggraph submission process.