Autoretirement rules need adjustments

[Blind4Basics]

the autoretirement is now forbidden with one single "very satisfied" vote, even with 11% of satisfaction rate and 8 "not satisfied" votes.

see here: https://www.codewars.com/kata/sort-the-temperature

┆Issue is synchronized with this Clickup by Unito

[kazk]

See Auto-Retire section.

The kata has assessed rank 8 kyu so it needs 12 votes to retire. This rule was always there, but it used to get retired because the satisfaction percent was incorrectly calculated.

A kata can also get retired for getting 4 "Not Satisfied" in a row.

I personally don't like requiring this many votes for white kata (and also more published kata requires less votes to retire makes no sense), but I haven't given much thought to it yet.

We can probably retire earlier if the satisfaction percent is really low like this case.

I can't spend much time on Codewars right now, so I'm not sure when I can do it, but feel free to post suggestions.

[FArekkusu]

The rules have to be adjusted because now the amount of katas "left in the void" will only grow due to the latest fix to the satisfaction rating calcuation:

---

Idk how to handle katas which end up with 30-50% satiscation rating (because they have no chance to climb up, but it's not obvious after how many votes can these katas be proclaimed "dead"). But IMO, at least anything that falls in the 25% percentile should be automatically retired.

[kazk]

Like I wrote above, those should retire when the number of votes reaches the number of required votes. The kata posted orinally is retired now.

Maybe we can make the requirement proportional to the number of votes given to make it retire sooner, but not too soon.

Something like

votes	auto-retire if
4	0
6	<= x/8
8	<= x/4
...	...

where x is the requirement to approve.

[IanTaylorGMSL]

I just had a Kata retired for getting 4 down votes in a row with no feedback.

I have two other Kata in the beta process 1 awaiting approval and the other waiting for more people to attempt it.

The two Kata in the beta process both had issues and suggestions raised that I have resolved.
In my opinion the auto retired kata was the best written of the three when first published as I felt I had learnt from the mistakes of the first two. As far as I can tell I have followed all the advice on publishing katas for the retired kata (I did not for the first two as I had not found bits of the guidelines).

In my opinion auto retirement should be based on unresolved issues in some way. e.g. If an issue has been unresolved for more than a certain amount of time then it should trigger a vote between retiring the kata and marking the issue as resolved. I think the author should not be able to resolve issues themselves but instead mark it as fixed which notifies the raiser to review whether the changes have resolved the issue.

Perhaps there could also be certain special types of issue that trigger an immediate vote. e.g. duplicates where the vote is on whether people agree it is a duplicate of the kata linked in the issue.

Finally it is unclear on what happens with Retired Kata, is it possible to fix any issues and republish it or is it now permanently dead?

[Steffan153]

I just had a Kata retired for getting 4 down votes in a row with no feedback.

Same thing for me. It was one day and --- it was retired.

[Voileexperiments]

Finally it is unclear on what happens with Retired Kata, is it possible to fix any issues and republish it or is it now permanently dead?

People are always free to publish retired katas as a new kata by copying everything from the retired kata to a newly created kata; "retired" is just a "draft" state lock. It's mostly implemented to prevent low-quality beta katas further clogging the beta kata list (yes, there are lots of low-quality beta katas, and beta kata list is very clogged).

[IanTaylorGMSL]

The overall problem seems to be that the auto-retirement rule to filter out low quality katas relies on the subjective view of whether the first four people to complete and vote like the kata.

It sounds like the work around is to copy the retired kata into a new one and hope the first people to see the copy either give feedback so it can be improved or have different criteria for what they like in a kata.