An idea I'm putting forward again because I've just read someone who wrote about it: it would be nice to take into account unique corrections received, not all of them, to calculate a user's correction ratio
In some languages, it's not unusual to get 5, 6, even 7 corrections to a same entry. It's enjoyable but it also means your correction ratio will jump through the roof, forcing you to correct massively before being authorized to write another entry
I realize 5 correctors often mean 5 different corrections even when they correct the same sentences, but you don't choose how many folks will correct you, and in some languages it's impossible to get as many corrections. Therefore, I do think it's somewhat unfair that people who get corrected a lot are forced to correct a lot too.
I think it would be a good idea to explain the rules about correction ratio somewhere on the site. Before looking at the codebase I also had no idea that the correction ratio is only enforced when there're uncorrected entries lol. That alone would clear up some confusion I think
what's the reason for it not just being (number of sentences you've corrected)/(number of sentences you've written that have been corrected)?