Yeah, I don’t like that the privileged coordinate system is the Western one. I would prefer the Japanese one, because numbers are a lot easier to locate mentally on the board than letters
Easier for who ?
This is still a bit tangential. I was just remarking that generally it will cause a lot less confusion, in person, in chats, on the forums if we don’t have 6 different people talking about 6 different displayed rank styles.
The number kanji aren’t hard. I learned to recognize them watching Hidetchi and Nick Sibicky videos years before I ever studied Japanese. But Arabic numerals should be aliased to them for move highlighting in chat
It’s not really the point.
It’s not so easy mentally when you’re looking at symbols that you’re completely unfamiliar with.
What was your point, then? How much prettier than arabic numerals they look? 
They’re still big-endian decimal numbers. And three of the digits are just hash marks (4 if you count 五, which while only having 4 strokes, is sometimes written with 5 in order to use it for hash marks). And in any event, exposure breeds familiarity, so it’s rather circular to argue that we shouldn’t use it because it’s unfamiliar, when it’s only unfamiliar because we don’t have it as default. And it would not be unfamiliar in the countries where Go is most popular, anyway
And my main issue, is that “k” is just as unfamiliar as a symbol for ten as 十 is, if not far worse, because it’s in vigesimal, a number base with which fewer people are familiar than decimal (citation needed). And I’m ironmanning by choosing “k” here. Most letters are far worse: I have no sense of where “m” or “p” are
You asked what my point was only to just go on to discuss my point…
It’s a bit naive to think that you can just introduce something new over something familiar and argue that people should just get used to it. It’s really not circular at all, since we’re debating about using something people are familiar with in their every day lives, the alphabet and Arabic numerals vs kanji numerals.
And this is my point that
It’s fine if you have a preference, but that doesn’t mean that in every context it’s better to either force that preference on others and suggest they get used to the kanji, or in the context of the decimal rankings (this thread), potentially have several concurrent and each confusing in their own way, displays for rankings. It’s just chaos really, my screen says your 3kyu, your screen says your 2kyu and the next person says 3k+ and another 2 ↓. It’s just a usability nightmare really, when we’re all looking at the same page, presumably a game page or tournament or something.
I don’t think anyone is arguing for rank numbers differing between users; aren’t the +- family of suggestions predicated on having the number in all decimal ranks correspond to the integer rank, and using balanced ternary in the mantissa to further subdivide? so your example would never happen, and is thus a strawman
Yes, but the western system is the something new, and the Japanese system is the something familiar. You’re just arbitrarily restricting the consideration to an insignificant portion of the Go world
Noone’s advocating to use the alphabet, strictly speaking, they’re advocating to use glyphs from the alphabet as vigesimal digits to denote columns (or possibly rows, or both (see sgf)). And you’re latching onto surface-level apparent differences in arabic and kanji numerals, while ignoring the fundamental sameness of them as big-endian decimal systems
There’s four different systems under discussion, and then the ± or ↓, ↑ are cosmetics on top of that really. I don’t understand what you’re arguing about, or what you think is a strawman…
This is just being silly.
I think it’s time to move back to the decimal display discussion, as this is just being obtuse.
We can continue debating over this in a pm/dm if we really want to.
No, it’s pointing out your willingness to selectively ignore similarities and differences in the systems as it suits you
It’s a strawman because it describes an example ( 3k+ and 2 ↓ describing the same user), which would not happen with any of the systems under discussion
Would you be happy if I changed it to ( 3k+ and 3k ↑). Is that your main issue? (Some systems will do different binning of the ranks, I didn’t 100% whether it was possible) Are you saying that having a setting to toggle between ± and ↓/↑ wouldn’t happen?
Hardly - but again pms please, if you really insist on this pointless attack on my comments so far.
Yes, this would fix it
It’s only pointless if I’m wrong
Lots of different options to toggle and select from might mean a lack of a coherent product vision. In product management, you don’t give the user lots of options. You almost enforce the product vision you follow personally as a founder. There will always be people not satisfied with the product. That’s not the point. The underlying truth is that people usually don’t really know what’s better for them. They always fight the change - until they adapt, understand the usefulness of the new solution, - or leave the product. Increaaing the number of settings is one of the biggest product design sins, as you want to simplify and streamline user experience, not make it more sophisticated.
2 Likes
technically… OGS has already picked the option for us, and we are just users griping about it 
2 Likes
Here’s an idea, not sure if its a good one, but i like it myself:
Ogs ranks are “the range where the players rank probably is”, shown as ± value. Currently the center of that range is rounded down and shown and shown everywhere on the site. This means that if someone has rank like “4.4k ± 1.1”, then the ogs’s rating system is 90%(?) sure player is somewhere between “5.5k” and “3.3k”. And this “4.4k” is then rounded down to 5k as their rank.
Instead showing the rank as one specific value like “5k”, how about having the entire range written there after the players name? Going from “user [5k]” to “user [6k–3k]” as the new ogs norm?
This way it might be more clear that:
- ogs ranks are just estimates made by some random rating algoritm, its not some highly accurate reflection on players skills
- there might be considerable overlap on what the algoritm thinks the rank of any two players actually is
- a player has not yet have an established rank if their range is huge
- the expected level of amateur players tend to vary quite a bit depending on many factors, mood of the day does affect the performance of all humans
Negatives:
- more clutter, potentially 3-5 more symbols after every rank that reads on the screen
- would it affect handicaps?
- a bit weird ngl
Yes, absolutely. But it shouldn’t prevent us from a more high-level discussion - about product vision. I think what’s lacking in the conversation is the clear goal of ranks. Why do we need to have ranks and what purpose they serve? Next, what user should infer from the rank information, why do users need to know ranks (and also types of ranks - username ones, detailed ranks, etc), what’s the goal? And after that we could figure out the system to present ranks to serve a specific user experience.
1 Like
It seems that some people feel that decimal ranks are the ground truth for displayed ranks on OGS.
So 1.2k would be the “true” rank, while labelling that as strong [1k], weak [1k] or strong [2k] is merely a design choice.
But to my understanding (and OP’s understanding) it’s the other way around. The rating range of [2k] is the ground truth, so the upper and lower ratings thresholds of the [2k] rank should be kept as they are.
In OGS’s current decimal rank display system, 1.2k means strong [2k], i.e. close to the upper rating threshold of the [2k] rating range. The ground truth is strong [2k]. The ground truth is not 1.2k.
OP feels that it is confusing that 1.2k means strong [2k], because the integer digits 1 and 2 don’t match, so he proposed to change the decimal rank display so that 2.2k would mean strong [2k]. And then we discussed some alternatives.
6 Likes
Typo? I think you mean strong [2k], close to upper threshold.
But besides that I agree with you. The truth is weak/strong within a rank, and the decimals are a design choice for how to communicate that to users.
5 Likes
Ah yes. I’ll fix it. Thanks.
2 Likes
A side question – someone above mentioned the option of waiting until the “next update of the rating system”. I don’t have any info on this; my understanding was that the OGS ratings should now work well with AGA/EGF ratings and that no further updates were needed.
Does anyone have information regarding whether another update is planned (or not) ?
I think the essence of rank is it should communicate the handicap stones. 1d should give a 1k one stone. That narrows it to B or C.
Then the decimal system gets you situations like 1.1d gives one stone to a 0.9k. The decimals create a visual incongruity with the leading numbers making the user process the idea of a 0k player before they logic it back to 1k.
With the ± system a 1d+ gives one stone to a 1k+. A 1d- gives one stone to a 1k-. That seems clearest to me. Then we can always just say a + is stronger than a -.
2 Likes