Opponent rank settings bug

Yes, I was mulling that option too.

You raise a good point about the experience when your rank changes. If the numbers were “absolute” it might imply that they don’t change when your rank does (which would be wrong).

The last display reflects what was out some discussion, not sure we can get even better.

What interest me is if it’s bugged or not.

I think we have a set of reporters saying they’ve seen that it is bugged.

So it probably is.

This is in the “important but can’t get to it yet” list. I imagine surely we must be nearing the point where we can tackle that one :slight_smile:

(Possibly after the new problem of conditional moves … and >9 rank difference ranking).

However, there is also a possibility that the reporters were confused about how it is supposed to work, because it seems that it is not unambiguous yet.

How is this (my interpretation of Uber’s suggestion):



How’s that?

@Civilian

It looks even more confusing without units.

Maybe you could infer that +3 is three ranks stronger, -3 is three ranks weaker, but just the number 9 or 3 by itself, even if it’s in the weaker column sounds confusing.

Weaker 4, doesn’t tell me if it’s like 4kyu or 4 stones weaker.

At least Weaker -4 makes me think at least theres not a -4kyu so it’s probably 4 ranks weaker.


Maybe it’s something to do with the weaker and stronger being out of line with the Opponent rank wording.

Opponent rank 4 might be what makes it feel like “4kyu” or something, even though the word weaker is floating above

1 Like

I think it needs a few more words and different layout to be clear, like (dots just for spacing):

Opponent’s rank: … From: [4] ranks weaker
. . . . … . … . . . . . . . To: … [3] ranks stronger

However, that is getting a lot of clutter and cognitive load on the user for what’s supposed to be a quick match screen. Were I the OGS product designer, I would say the additional value from offering asymmetric relative ranks ranges is not worth the additional UI needed to clearly communicate it, the addition burdens it places on users parsing the UI, and complications and likelihood of bugs it introduces into the code and would thus simplify the feature to:

Maximum Rank Difference: [3]

The Custom Game setup screen could have the asymmetric options.

5 Likes

I like that but there is some uncertainty that it would fill the expections of the players. Because of some recommendations, there should be a part of them more inclined to play with stronger as weaker (No moral judgement here).

1 Like

Then use custom game not automatch. And if someone expects to be able to do advanced settings like asymmetric rank ranges in an automatch then they have silly expectations and should either be disappointed or update their expectations.

I guess you also wish to totally remove rank setting. So it would always try to get even opponent and if fail always any opponent.

Given the current complexity level of the automatch page, I think a ‘Max rank difference’ field with a single number is reasonable. But yes I would like to go even further with a more radical simplifying redesign, though the parallel Fischer and Byo Yomi time settings and that Exact/Flexible/Multiple dropdown would be in my crosshairs before a max rank diff control.

2 Likes

It’s never, to this point, had units, has it?

The “+” and “-” are appealing, but horrendously deceptive, because they are totally ambiguous.

In my mind, I’ve ruled a firm line through them. No matter what we do, I think we need to get rid of “+” and “-” when refering to rank, because the meaning of the direction is indecipherable.

I’m returning to this as the best thing:

Having mocked it up, personally I love it. It is unambiguous and clean.

The established downside is that the “remembered values” will change when your own rank changes, but really … who cares? The effect is what you wanted anyhow.

3 Likes

Custom or automatic? The difference is not only about simplifying the pairing, it can be if you want to chose who you play with.

I’ve spent the afternoon trying to debug this, with the help of ChatGPT of course.

ChatGPT also can’t get it’s head around the whole “+”/“-” and “lower number is higher rank” thing either :squinting_face_with_tongue:

It’s just a complete mind-zapper to try to think about ranks, rank differences and rank comparisons while holding ambiguous descriptions of direction in your head. Is a variable called “upper_rank_diff” the weaker or stronger rank boundary? Is a rank difference positive if the person is stronger or weaker? Etc. No wonder there are bugs :exploding_head:

In retrospect, “if only it were absolute rank values being used, the UI and the implementation would have been easier” :smiley:

Easy to say in hindsight.

Anyhow, I seem to have found an actual problem (sign of a rank-comparison getting dropped) so I will submit the suggested fix and see how we go.

3 Likes

A fix for the bug that is “sometimes I’m matched wrongly” is in.

The updated selector (as per screenshot above) is also in.

:crossed_fingers:

3 Likes