If it’s your first game of Go, even over the board, you might also not be “participating” in the scoring.
I’ve seen people that have only played online ask how to score the game over the board, and it’s more of a demonstration, same way the auto scoring withe the bots is in a way 
I suppose. It’s hard to remember not understanding things like that
Sorry to bring up an old post, but I want to add something:
when a player sets up a custom game that had rank restrictions, I don’t think a new player with “?” rank should be allowed to join as if “?” is a wild card, like this game
Black to move (online-go.com)
If you don’t want “?” players you can cancel the game at move 1 or 2.
It’s not a wildcard. It’s a measure of uncertainty.
At the time that person joined your game, their rank matched your criteria. It just happens to be highly uncertain, as well.
Wow, this is so big it’s too much work to read it all. Not sure why there’s a need to reinvent a solution that’s been solved before and in my opinion better. I think introducing players with such limited options results in the feeling one often gets that the opponent is perhaps underhanded. Knowing how this works that appears to often be the case and the culprit is not someone playing an a dishonest rank, it’s that OGS “forced” an inaccurate rank on them. My server experience is limited but I think DGS does a good job of accommodating users with all kinds of backgrounds. It lets a new user select a starting rating. Provides a rating converter (think OGS has one too) that lets user input ratings from other servers or Go associations. One can also refer to the ratings comparisons at Sensei’s Library.
I don’t see what there’d be any objection to new users being able to start playing at OGS with a rating that corresponds as well to their skill level as possible as determined by previous experience. I think it would contribute to decreasing perturbations to ratings and the feed-forward effects of that.
Welcome on the forum.
The changes that occured some months ago were in the sense you are advocating and seems great as no one came to complain at least.
There are some big differences between OGS and DGS which make a direct comparaison difficult. The biggest is that OGS manages live games but DGS not. So the user profile is a bit different.
I don’t know how it works because I can’t get into the new account creation process without entering an e-mail address. I really don’t see what the difference between OGS and DGS has to do with it. For one thing OGS does provide “correspondence” play which is pretty much the same as DGS. Should there be a different initial rating setup for OGS users playing correspondence instead of realtime? I doubt it. Unless OGS is going to apply different ratings for different types of games: my 9x9, 13x13, 19x19, realtime and correspondence,…At least 6 different ones?
As far as I can tell, I have one rating that is used when games are created, accepted, scored and my rating is updated. If not, it’s a very opaque process. I don’t even get to see what the actual handicap is that will be applied in a game when I click Accept although at that point the info should be available to tell what color I’ll get and the number of handicap stones - but it isn’t.
But please explain what is special about OGS that makes such coarse settings of initial rating desirable. It’s been so long I don’t remember how KGS assigns initial rating nor IGS where it’s been a long time since I set up an account I haven’t used in ages.
P.S. One might think the basic user interface would be settled but it’s changed a number of times in the past week. So why not the rating system?
Well you can play correspondance and live games on OGS. Live games means that you may have different people coming in like people who know very few on the game, who just want to try quickly for example. IMHO people coming to play on DGS may be different in this they will think twice before engaging in a correspondance game, check footnotes on how to enter and evaluate a level etc… OGS is more prone to some deviant attitude like trolling, sandbagging and has to ensure the QOL of its users
Registering on OGS is quite easy, just need an email.
There is only one rating likewise on OGS. There are some rating given by statistics in your profile but they are pure statistics, they are not taken in account by the pairing and your global rating which is the only one effective.
The changes are still being adjusted. I’m not sure that it’s the right time for adding more to the package. But yes we can already discuss on your suggestions.
I would be the last to argue with what you said about sandbagging etc.; those sorts of abuses are geared towards a quick payoff 
I came back to clarify - just in case - that I wasn’t suggesting changing the rating system (Glicko is a lot more sophisticated than DGS’s approach which adjusts only the players involved and otherwise has no further effects) but the process for assigning initial rating.
The other thing that came to mind was an experience playing in an AGA rated event after many years since my previous one. My current rating is a bit different. I compared my ratings with a couple friends who have more current AGA ratings, checked the rating calculator at DGS and the tables at Sensei’s, picked a rating to enter at and conveyed this info to the T.D. and it was done. That was “realtime” play and much better than entering at an outdated but official rating or at one that was too low or too high.
I’m not sure what can be done about people who want to play other games than Go. It seems they’ll find a way around anything to achieve their ends. So why not give a benefit of the doubt hoping the majority of new users are coming to be reasonably serious and fair? Helping them find the right spot to fit into is the other end of the balance.
What is QOL?
An acronym for Quality Of Life
I dunno if you went into the hassle to read all the debates about the changes that OGS finally implemented for letting people chose between 3 levels of play. OGS can’t just not care about problems due to deviance so found some compromise between let none fix his level on its own (older system defended by people who fought that glicko was quickly enough responsive but which was very stressful at a beginner level) and let everyone give its own rating (which was bit of an utopia considering that OGS has concerns with a bunch of deviant attitudes or even simply not understanding a system)
You can’t just generalize your positive experience you mentioned to anyone coming on OGS. They don’t always come in the spirit of an experienced go player going to play an AGA tournament. There is much more to consider, starting by how will you make people read footnotes that they usually just don’t.
As it stands now, complaints about difficulties for beginners to get fair games and such have disappeared and on the other side there is still a considerable amount of reports to take care, but luckily not so much about inconsistency in the pairings so it seems that it’s pretty well working.
Coarse initial settings are enough because
If you want the short summary, the glicko2 rating system adjusts quite fast, so having everyone start at the same rating should work fine in principle:
However the second case was the one that was problematic. It was indeed only a few games, if all goes to plan. However sometimes you get players resigning against new players because they don’t want to play out a whole 19x19 game against a beginner when they just want a competitive game.
Effects like that slow down the process of ranking players quickly.
Equally the process of having to lose a few games to be ranked was apparently quite problematic for new players. I have an anecdote from a strong player in other abstract games that told me learning to play Go was very harsh here. It surprised me but also changed my opinion.
There are always some drawbacks to letting people pick starting ratings - some players will pick a lower rating to get easier games to start with. Some people will sandbag etc.
Still I think the general new player experience was the main drive for the change.
There’s always different approaches in different places, in chess chesscom will let players pick their starting rating with preset options, lichess will let glicko2 do its own magic. It just depends on the direction of the site, who it aims to appeal to, and what its stance on how it affects new players is.
I apologize if I missed any previous suggestions, but how about determining the initial rank by having the first one or two matches against a bot? OGS offers a unique feature that allows players to face bots.
I can also understand not wanting to play against opponents whose rank is unknown.
You definitely missed a raft of suggestions - all those leading up to this proposal thread
Also: this proposal has now been implemented.
Probably we should close the thread, and let other discussions be had in the context of what we have now