I keep struggling with my own rank differences between correspondence / live / blitz games in finding a ‘proper match’. I have made comments on the issue whenever it seemed appropriate but I would like to address it again in a separate thread.
The red lines I added to the first image are not exactly corresponding with my own ranking overview (my blitz ranking should be higher) but I guess anyone can see what I’m trying to point out here. If everybody would see their own real ranking in the graph which shows the outstanding invitations the choice whether to accept a game (according to the offered settings) would be made much more easy. Each time when I consider to accept a game I check the ‘profile’ of the opponent to see how our rankings ‘really’ compare. I don’t mind to play against a stronger / weaker opponent but I try to play games within a certain bandwidth of difference in strength. Since it clearly seems that most players on OGS prefer to play without handicap (or simply refuse to play with handicap, for whatever reason) it is hard to asses what you’re about to get yourself into.
At my local club the response to this issue is simply “play more games!”. The response over here is “create a second account”. Both options do not sit well with me. I want to be able to play with one account and playing more games is what I do, especially live games. The drop in my average ranking is a clear sign of that but even after two years or so the ranking gap between the two disciplines (they are two totally different ways of playing) has still not become any smaller. It makes me feel a bit uncomfortable, because I feel a cheat when I accept a correspondence game against a 10 kyu and I know that I’m going to feel miserable when I accept a live game against a 10 kyu, with or without handicap.
Anyway, I’ve said what’s on my mind. To me it seems that if there’s time to implement a ‘thumbs up / down’ because the OGS website needs to improve with a feature taken from a well built chess website, something like this can be fixed too. No disrespect to the developers meant by this remark, by the way. I really do appreciate OGS!
Please feel free to weigh in. Maybe I should not be bothered too much (though that would take some serious arguments) or somebody comes up with a better idea, who knows?
The blue line represents my average ranking, as commonly shown on the page with outstanding game invitations. The red lines, on the other hand, show my separate (average) rankings in blitz / live / correspondence. If everybody is shown their own ranking in the different disciplines in which the game can be played (blitz / live / correspondence) the choice to consider accepting an outstanding invitation would be much easier, whether proper handicap is applied to the ‘real ranking average that is representative of of the type of game or not.
My average ranking now is 10 kyu. My ranking in correspondence games is 7.2 kyu and live games is 12.9. It is hard to find suitable games if just one average is used as reference, especially if the respective averages (live / correspondence) of my opponent are the other way around and the decision to accept an invitation has to be made quickly. And if I create an invitation because I want to practice my live games the reference for that invitation is based on 10 kyu in stead of my real average of 13. Am I making sense here?
I suppose my issue is quite peculiar and personal. The feeling that the way it’s organized now is not well balanced is apparently not shared with many people, given the fact that it has taken a full day for 3 likes and one reply.
Correct me if I’m wrong, but it sounds like your concern would be much less of an issue if you could play with handicap, but you’re also complaining that people tend to search with handicap disabled?
I’m not complaining about people choosing to play without handicap. That’s a choice everybody is free to make, even though one of the remarkable features of the game is just that, being able to play a fair game against any opponent by offering or taking handicap. But that’s a totally different discussion, which I do not want to address here.
If I, as an average 10 kyu (but actually 13 kyu in live games) accept a challenge from a 10 kyu opponent (who is actually 10 kyu live), I have a clear disadvantage even though our average rank is the same and the graph shows we’re a proper match. My issue is that I would like to to be presented my opponents in a way that is more balanced.
I’ve used the same example somewhere else on the forum. Trying to figure out how Chess.com deals with outstanding game invitations I came across this remark on their forum:
Because your point seems to be misunderstood, I’ll summarise it in different words, maybe it helps.
The issue is:
Some individuals, like you, have a very different rank per boardsize.
Therefore, let’s change the horizontal rank-indicator-line to be at your Boardsize Rank, rather than your Overall Rank. This helps you to easily see which game invitations are fair to you.
I think this is a great suggestion. I don’t use that overview, but I can see the value.
It is not clear to me if you also want the game invitation to be plotted at the Boardsize Rank of player who created the game.
Both suggestions do however go against the explicit policy of OGS to use the Overall Rank as the leading strength indicator in matchmaking. But that’s of course not my call to make.
Yes, I agree that (coming from being a longtime chess.com player) I am surprised to learn those graphs are using overall rank instead of per-time-control.
Why is that? It’s not properly explained anywhere as far as I know. I’m a frequent visitor of the forum for 4 years now and I have never read anything about this policy
All matchmaking uses the Overall Rank and nothing else. The Sub-Ranks are calculated separately, but their use is strictly for your information. They don’t do anything behind the scenes in regards to matchmaking.
It used to be different, with matchmaking per Sub-Rank. The change was probably right before those 4 years ago, but my memory is hazy.
Chess has 3 playing speeds, but go has 3 playing speeds and 3 board sizes, so we’d have 9 different ratings, that’s perhaps too much, most of us don’t have 9 reliable ratings.
I don’t know how I feel, recognizing this is a complex topic with no “correct” answer, but I could easily see the response that could easily overrepresent e.g. your skill at blitz 19x19 games if you’re queueing for a blitz 9x9 game. And the same could be said if you sharded the other way, e.g. use the average rating of all time controls across only your current board size.
It’s complex, I know. But like I said, Go is a game suitable to play for all players against all players if circumstances are taken into account (within reason and a certain bandwidth)
This is why I’m also coming back around to my first comment in this thread. Coming up with a perfect solution for ratings display is not possible, and any compromise will leave some people unhappy. But pushing for a culture change of encouraging more acceptance of handicap play is a path that can improve matchmaking quality, in a way that feels more sustainable and achievable than “convince the developers to make a different set of tradeoffs than the ones they’ve already made, in ways that will improve some scenarios but make others worse”.
(I imagine there is also a difference between “changing the matchmaking algorithm” and “changing the visualization for the manual game finder”. If I understand correctly, you’re asking for the latter, which also seems like an easier thing to ask for as an optional setting, so of course do not let my ramblings deter you from whatever advocacy you want )
