What should rep gain/loss be on Papers?
The Codidact team has recently informed us that it is now possible to have different rep bumps on different post types. See this new answer by Monica to the old meta discussion that created the Papers category in the first place.
Papers were always intended to be significantly more substantial works than ordinary answers. This seems to have worked out. We have gotten some good papers. In all cases, the authors have put serious effort into writing their papers. That's exactly what we want.
It was also intended all along that the rep bump (both positive and negative) be more substantial to acknowledge the extra work we expect, even demand, of papers. At the time the Paper category was created, this was not possible. Now it is.
So now the question is what exactly should that rep bump be. I figure 3x more gain from upvotes than on ordinary answers makes sense. I think the reduction for a downvote should be the same as the gain from an upvote. Downvotes on papers are intended to be rare, and that's how it's also worked out. A downvote on a paper means it is outright wrong, badly written, or the presentation is quite misleading or confusing.
Unlike with answers that can be more casual, we want people to really think and put some effort into polishing their papers before posting. The reward should be higher, but the consequences of posting something half-baked should be more substantial too. We want people to take papers seriously.
I just checked, and so far there have been 17 upvotes and 1 downvote in the papers category. Currently, upvotes are +10 and downvotes -2. Even the one paper with the downvote would actually yield more rep for the author with the new system than the old.