Remove "Average Power User Assessed Rank" from kata editor

[FArekkusu]

It looks like this info was added at some point to show how PUs' opinions are different from the general public's ones, but as of now it serves no good purpose:

• It's not clear whose votes are counted towards this metric (we generally consider PUs to be users with 6k+ honor, but it's not clear what criteria is used in this particular case).
• It's not clear how this metric is tracked (is it updated every time you view the kata editor, or is there some running total that gets incremented when a user votes, so it's possible for a "non-PU" to cast a vote, become a "PU", and their choice will not be accounted for because it was made before the "promotion").
• Assuming the criteria for being a PU in this context is having 6k+ honor too, almost all the votes are counted both towards the "average rank" and the "average PU rank", but they always end up differing by 1 kyu (which could be explained by different rounding if this is all true), so this metric is practically meaningless nowadays, and only leads to confusion that there's a separate group of people who underrank everything compared to the general public, when, in fact, there is none.

┆Issue is synchronized with this Clickup by Unito

[Blind4Basics]

thing is, the usual average has its troubles too. Afaik:

• regular avg is a ceiled floored value (I tend to think about the difficulty rather than the number... :/ )
• PU avg is rounded, (but I believe it's in "bank fashion", meaning that x.5 is sometimes rounded up, and sometimes down, depending on the parity of the result)

@kazk: if a change needs to be done here, imo the best (quick) option would be to remove the integer values and just drop the float result instead.

related: #1939

when, in fact, there is none.

errrr.... actually there is... x)

[kazk]

Yeah, that number is pretty much pointless. See:

And I just found out that "power user" in this context is users with at least 500 (!) honor points. Almost everyone assessing rank is "power user" (only [100, 500) are not) and avg is the average of all assessments, so it's pointless. The number is hardcoded and looks like it wasn't updated over time.

#2030 (comment)

CW has many WTFs like this from over the years.

It's not clear how this metric is tracked (is it updated every time you view the kata editor, or is there some running total that gets incremented when a user votes, so it's possible for a "non-PU" to cast a vote, become a "PU", and their choice will not be accounted for because it was made before the "promotion").

The number is updated whenever a rank assessment is done. So in your example, their choice will not be accounted for until the recalculation is triggered.

which could be explained by different rounding if this is all true

Good point. I just checked and the power users' average is rounded to the nearest. The average of all is not rounded and is truncated implicitly when stored in the database.

if a change needs to be done here, imo the best (quick) option would be to remove the integer values and just drop the float result instead.

The field is an integer field so it's not quick.

How does this relate to #1999?

---

I don't think showing an average is helpful to begin with. If we're going to keep the current beta process largely unchanged, I'd show a sparkline:

The root cause of all the conflicts between users when assessing ranks is that everyone have their own scale. Like I've been saying, this is Codewars' fault for not providing any guidance over the years.

[Blind4Basics]

How does this relate to #1999?

it doesn't, typo... XD it's #1939

[FArekkusu]

I don't think showing an average is helpful to begin with. If we're going to keep the current beta process largely unchanged, I'd show a sparkline

I wanted to open a feature request to show a breakdown of all the votes (a bar graph would be great) exactly for this reason, but kinda forgot about it :(

[Blind4Basics]

that's #1939 actually. But the thread got crowded about other topics... (my fault... x) )