poasting in a rep record thread :D

Okay, located 6 old 2010 posts again.

Starting rep: 3193.

First post took 5 downrates to become low-rated. Rep changed to: 3192.
Second post took 5 downrates to become low-rated. Rep changed to: 3191
Next 2 posts (3 downrates on one, 2 downrates on the other). Rep stayed at: 3191
Next post took 5 downrates and the rep dropped to 3190 at the same time that post became downrated.
Last post was downrated 4 times without any effect.

I then went and uprated one again..

First uprate on a low-rated post removed the lowrating but rep stayed at 3190.
Post then took 4 more uprates and became high rated! Rep shot back up to 3192.

I then went back to another low-rated post and uprated it, then downrated it:
Low-rating was removed by the uprate and not added back by the downrate! Rep stayed at 3192.

I then screwed up and accidently increased rep to 3193 :) NOTE: Solution posts are not reliable posts to test on!

I then found another post and began uprating it. After 5 uprates it became high-rated and.. rep stayed at 3193.

With another post, I uprated it to high-rated, downrated it 3 times and then uprated it again and it didn't become high-rated (and rep remained at 3193 throughout).

This reveals:

A post must be downrated 5 times before it becomes low-rated.
A low-rated post reduces your rep by -1.
Posts that are downrated but do not become low-rated do not have an immediate effect on rep.
Uprating to remove a low-rating does not restore rep unless it is uprated significantly.
Getting a high-rated post does not increase rep immediately.
Uprating a post removes 1 existing downrating and adds an uprate (this explains why it takes 5 uprates to go from both low-rated to high-rated and neutral to high-rated).

These revealings imply that 2 of my posts in this thread became low-rated at some point, reducing my rep by 2 and then someone uprated them, possibly just to remove the low-rating.. which didn't restore my rep.

Basically, the post rating system seems to not only be latching, but biased towards reducing rep.

If a post annoys 5 people, it becomes downrated and then if someone thinks the low-rating is unfair, they remove it by uprating which removes the indication that it was low-rated.. but the rep has already been lost. The only way to gain it back would be for the post to become high-rated, which actually gives me 2 rep (one for fully undoing the low-rating, one for becoming high-rated?).

These findings are somewhat concerning and have a few unfortunate implications. For example, if I built a script to reduce someone's rep, I could do so stealthily by downrating 5 times to reduce rep by one, then uprate to remove the low-rating, hiding its tracks.

I will need to do the inverse of this test, to see if it latches the other way...

I currently have 9 rep testing accounts (that took less than 5 minutes to create, manually, in case you are wondering) along with my own so if anyone has any test suggestions that can be done with those resources, please share them!

I may have missed it, so apologies in advance if I did. But, was there any effect on rep if a post was down rated to the point of "low rated" (and thus incurring the rep penalty), up rated to eliminate "low rated" status, and then down rated again to "low rated"? If so, was there a rep effect when the post reached "low rated" a second time?

Or you could add the impostor's rep to Alfie's regular rep so we your script provides his true rep.

I hadn't tested it but I just did.

Rep: 3193
Post downrated to low-rated.
Rep: 3192
Post uprated (removes low-rating).
Rep: 3192.
Post downrated (became low-rating again).
Rep: 3191

... may I amend my previous statement to "extremely biased towards reducing rep"?

EDIT: Namur's rep just jumped up to 3196 without me doing anything. Either he just did some stuff to get some significant upreppage or we just rolled over into a new day and his rep was recalculated or something.

Very interesting results, thank you.

Why in this thread ?

"As I am not particularly active on the forums I can pretty easily check my posts and despite none of my recent posts that I've made in the last week or so that I can find being "low rated"..."

If your findings are correct ^^^^ this is an unreliable method of verification, it could have been any post including the ones in the last week.

Everthing seems to work as i thought it did, with the exception of


"Uprating to remove a low-rating does not restore rep unless it is uprated significantly."

a +1 rep is apparently not being applied to a post tha gets 5 downvotes (thus -1 rep) and subsequenyty gets 1 to 5 upvotes, as, imo, it should.

if 5 downvotes is the shifting point that determines a -1 rep subsequent upvotes, between 1 and 5, should restore rep to its orginal value.

A post fluctuating between -5 and -4 (in terms of downvotes and countering upvotes) a couple of times ends up being more punishing rep wise than a post that gets 10 flat downvotes and stays put after that. I can't believe this is intended behaviour.

In any case it seems clear that the system can't be messed with unless someone cheats.

Oh, I should be at 3198 actually, the 3193 i started out with plus 5 from a question topic that got closed just now on the BL board :)

Damn, I always thought that was Mars.

Mind = blown

Repeated the test..

Rep: 3196
Post downrated to low-rated.
Rep: 3195
Post uprated (removes low-rating).
Rep: 3195
Post downrated (became low-rating again).
Rep: 3195

I then alternatively uprated and downrated through the rest of the testing account and the rep stayed at 3195.

I then low-rated another post with 5 downrates and it still stayed at 3195. Downrated another 4 times and still 3195.

Looking back at my test results, I think I've removed 5 rep points from Namur in total so I may have hit the maximum rep loss limit for the day. If so, it means the theory that the maximum loss is based on how many your started with (i.e. the lowest rep you can end a day on is $starting_rep - 5) is incorrect.

EDIT: There is also the possibility that someone else is up/down-rating Namur's posts to skew the results on purpose.

Sure, test away xyem, mi rep es su rep ;)

If you could confirm that 5 downvotes and 1 upvote does not undo the -1 rep brought about by the orginal 5 downvotes it would be great - if there's a design flaw it seems to me it could be there.

If you need another test subject you can feel free to screw around with my rep as well, I don't care.

I've been here for years and still don't get this rep thing. I guess i just don't post enough on this forum.

No he didnt. He uprepped himself, and probably thought might as well downrep some people. Thats a huge difference from a guy making 100 accounts just to annoy someone. Which is a crazy theory.
You are not lying, I just think you can't admit you were wrong. Well known people with thousands of posts on a forum often have this problem when they argue with unknown people. They would rather start huge multiquote wars instead of just saying "ok maybe that was a bit overboard,sorry yyahoo".

I don't have 1000 rep but I do wonder what you're going on about.
Is it about the exact number of accounts GR had? So if he had 99 or 101 you want it to be said and established that xyem is wrong? Or what?

And downrepped others. Quite agressively. And the assumption that someone makes an effort just to annoy others isn't a crazy theory but daily reality in the internet.


...and now you are talking to yourself, right? You know that you have are wrong and just can't admit it. Instead you are accusing xyem of being wrong, which he wasn't.
(apart from the exact '100' number. But you know that you can use large(ish) numbers exemplary for 'many', right? It's a normal figure of speech. Like "I've been telling you a 100 times to tidy up your room!" or some such.)