Brighten Godfrey points me to his large collection of conference statistics. His page comes with a nice program for plotting averages across any collection of conferences for which data is available. Brighten also points (via Google) to the official statistics from ACM for all of its conferences, which are unfortunately either very limited (eg, SOCG) or bizarrely sporadic (eg, STOC).
Since several people have asked, here is the raw data I used for my graphs (also in exported (and therefore not guaranteed to work) Excel and plain text formats). Enjoy! Please send me updates and corrections if you have any.
Does anyone have data for European algorithms conferences (ESA, SWAT/WADS, etc.)?
We now return you to your regularly-scheduled free speech, free speech, free speech, and private property rights.
I cannot understand why FOCS sometimes donot give reasons of rejections.
For most of the conferences including stoc, soda, crypto,
you get notification with comments from reviewers.
Why focs is so different?
Just because it is superior?
Posted by: | June 24, 2005 at 02:03 PM
The data from ACM is somewhat slanted, as it lumps together short and long papers.
(Which not how they are interpreted by anyone in the community.)
They just count the number of entries in the TOC, and then take the number of submissions from the preface.
The missing items are from years the PC chair didn't put this information in the preface.
Posted by: | June 25, 2005 at 12:06 AM
I cannot understand why FOCS sometimes donot give reasons of rejections. For most of the conferences including stoc, soda, crypto, you get notification with comments from reviewers. Why focs is so different? Just because it is superior?
What are you *speaking* about? Both FOCS and STOC committees put a lot of effort into giving feedback to the authors. If you did not get comments on your paper, you *should* contact the PC chair and complain loudly - this is inappropriate and everybody knows it.
Posted by: | June 26, 2005 at 11:19 PM
> I cannot understand why FOCS sometimes donot give reasons of rejections.
> For most of the conferences including stoc, soda, crypto,
> you get notification with comments from reviewers.
> Why focs is so different?
> Just because it is superior?
I think that it used to be the case that STOC/FOCS feedback was, to put it mildly, limited. But, in recent years, there has been a very strong effort on the side of PC chairs and members to give the authors as much feedback as possible.
Still, there is probably room for improvement. E.g., I do not see why the numerical scores are not reported to the authors. Outside of TCS, this seem to be a standard practice.
Posted by: Piotr Indyk | June 27, 2005 at 12:33 PM