Now, I'm aware of this question:

Too many review audits

But it was essentially a bug report, and it was beyond debate that the system was broken for a while.

I still think I'm getting review audits too often, even when the system works by the rule. They're blatantly obvious and they're getting annoying.

It's not just a matter of technical inconvenience, but user experience in a more psychological sense.

Reviewing is a community, voluntary work and being audited all the time is like having your boss come around to see "how you're doing" twice a day. No big deal, you could say, but with time it evolves into an atmosphere of distrust where one is inclined to shout out: okay, what is it that I am doing wrong??

I never failed to recognize a review audit, not once.

My proposition would be that if a user successfully passes several of them in a row, their frequency should be lowered. I'd use a Fibonacci number based algorithm or something similar; you get the idea.

If one gets caught not paying attention, it would be resetted, perhaps decreasing at slower pace from then on. Similarly to how causing car accidents can affect your insurance rate.

What is your view on the matter?


In fact, since review audits are plain random gibberish, they're so easy to spot that after a while one begins to reject them at a glance. One gets trained to recognize them at once.

In other words, if they come up too often, they kind of wear out after a while and thus no longer serve their purpose of forcing to pay more attention.


1 Answer 1


This is close to my views, recently posted on Meta Math Stack Exchange, but I won't delete my post, for the time being at least, because (1) the details are different and (2) the topic needs to be refreshed since it seems that no improvement has been made since your post of 2014.

In my experience, some of the test items were succinct correct answers, which might have been cast out by a robotic reviewer as being "too short". Arguably, there is a better case for testing reviewers with such good, albeit superficially deficient-looking, answers than for using the gibberish tests that we have both encountered. In the case of mathematics, a case could also be made for using as test items answers that are mathematically meaningful but quite wrong (i.e. not able to be patched up with a reasonable amount of editing).

In addition to gibberish, I have been asked, as a test, to review a foul-mouthed rant comprising just two words repeated many times.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .