I think there is a bug in the rank comparison which is used to figure out if a challenge is “ineligible” or not to another player. When a player has for instance a rank of 5k on 9x9 but a rank of 1k on 19x19 and he is choosing an opponent or he is assigned to an opponent then the overall rank is used to figure out if the ranks are matching. And the overall rank depends on what he played mostly. So it can happen that the overall rank of this player is either 1k or 5k for instance. And that he then plays opponents which are too weak or too strong. I think the comparison should be made according to the specific setting including boardsize and speed. It is probably no problem to use the overall rank if there are no historic data about a game setting. But in the other case the concrete setting rank should be compared.
On one hand, changing matchmaking to use specific game type ranks might make matches more balanced, but on the other hand, the players might also see a large difference in the overall ranks and be confused. For example, imagine an even 9x9 game between a player who is 2 dan overall and 2 kyu in 9x9 vs. a 5 kyu overall 2 kyu in 9x9. The 5 kyu would be at a psychological disadvantage at the very least. Perception and expectation is important. A practical disadvantage is also likely, because the 2 dan could have stopped playing 9x9 for a few years yet improved to 3 dan playing 19x19, whereas the 5 kyu is unlikely to have any experience above 2 kyu in any game mode.
I agree with the status quo for this issue, unless you have more specific examples where your suggestion would be better.
I don’t care too much, because I think for 98% of the cases the ranks are too close together to do so.
Got involved yesterday because I’m a nerd and was interested what the status quo was.
As a gut feeling I agree with Prokyu what I expect. If there is a significant skill difference between time settings or board sizes I would expect it to be reflected that the creator or the game get’s a game he wants.
On the other hand, it’s a bit more confusing for people not looking too closely. This confusion also can lead to frustration and support effort.
All in all I agree with the existing solution because it would lead to the second confusing way, way more often than people getting irritated or getting a game with an opponent they did not want.
Given the other ranks don’t do anything, but their existence can reasonably lead people to assume they do, how about deleting them all and only having the single overall rank?
All your “confusion” arguments can be contradicted by a single thought experiment. Lets assume I register 3 accounts, one for 9x9, one for 13x13 and one for 19x19. That would mean I use this accounts only for the specific boardsize. Mathematically it is the same then matchmaking via bordsize ranks. And where should then the confusion come from? Only from a poor implementation. And this can be avoided.
You’re putting on the saddle backwards .The majority of people will not create multiple accounts. But they are the ones that get confused playing each other when this would get changed. Every account with “mixed” rank might do that by accepting a game with an overall ranking, prominently displayed, the creator of the game does not expect.
Your hypothetical accounts might not do that on the accepting side of the equation, but still got the same problem when they create a game. Overall this does not matter, since they are infrequent.
I get why you want this, this would make the site behave more like we expect, but over the whole population of users I think we are not in the majority and only hindered minimaly in a few, scarce cases.
I recently created a second account for a similar reason: significant difference in strength between ‘live-’ and ‘correspondence-games’. I can now play each account until my ranks are settled and choose between the open invitations with enough confidence my opponent and I are either of similar strength or at least I know what challenge I’m facing upfront instead of getting unpleasantly surprised.
I remain convinced that the way things are organized on OGS, however perfectly explainable, is silly but by just getting a second account I can manage. Switching between two accounts is a bit laborious and I have lost the ability to use AI to review finished games with the new account. But that’s OK, I have SmartGo on my laptop
I like how you think, but I found the other ranks useful personally to see where my overall rank was coming from. I played mostly 13x13 ultra blitz (it’s how I reached 4 dan briefly a couple years ago), and I sucked at 9x9, losing to kyu players on it. Neither of those helped me improve much at 19x19. So maybe @prokyu ‘s suggestion isn’t so bad compared to outright deleting all the ranks.
Though I agree with @Folivoro that this discussion is not really all that important for OGS to continue functioning, we’re just all nerds here trying to solve a philosophical problem.
It’s indeed not all that important for OGS to continue functioning as is, but the issues are very practical, whether we approach them philosophical or not
.The majority of people will not create multiple accounts. But they are the ones that get confused playing each other when this would get changed. Every account with “mixed” rank might do that by accepting a game with an overall ranking, prominently displayed, the creator of the game does not expect.