This sounds very odd to me.
I want ranks because the whole world of go uses them. I know I can’t compare ranks on different servers and irl, but it’s a pity. What I’d love is to be able to compare my level to other players and worldwide rank is used to do it.
Some like Elo, some like glicko, but the only language that ALL players know is kyus an dans.
Remove them from the server and you’ll see how many people will complain. Not just handicap lovers.
Why would they produce different results if handicap is what matters the most?
The 1k would play against the 7p with 6 stones, and it would be an even game. Why is there another discrepancy when it comes to handicap stones?
Why don’t pros play with handicap in official games?
Let me say one more thing onto this point, if OGS only cares about ranks in a handicap purpose, why is sandbagging punished if someone doesn’t playing a handicap matches? I don’t think sandbagging is good, but it is something that happens and should be punished because people don’t want to play people that are multiple skill levels above them while thinking that they are the same level. Essentially that comment tells me that sandbagging is perfectly acceptable by mods if it isn’t in a handicap match, and I know that this is not the truth.
No, it’s “don’t mess with the rating system”. Ranks are for handicap games, rating is for games without handicap. Sandbagging is not only fake rank, but fake rating as well.
I don’t understand how intentionally misunderstanding the other’s argument is used as an device in an argument. It doesn’t help to convince others.
I don’t understand what you mean here.
Where do you you get that handicap from?
Pros played with (small) handicaps for centuries in official games in Japan, even in the 20th century in the Oteai system (from the 1920s to 2003). But they abolished it in 2003, because the Oteai system lead to slow inflation of Japanese pro ranks (they could no longer compete with Korean pros and Chinese pros).
Kageyama.
You mean form his book? Those were teaching games, so he was playing educational games against a student. He goes on to tell the story where he beat the same 1k with 9 stones when challenged to do so.
Also, I don’t know how Japanese ranks from the 70s compare to current EGF ranks.
I know I was beaten most of the time by a strong Korean amateur (pro level, 8d EGF) with 4 handicap 12 years ago when I was a solid 4d EGF.
Right.
Okay, but why have ranks that imply handicap when you’re not allowed to actually use handicap to measure ranks? Ranks would become just fictional numbers IMO.
Pro ranks measure achievement and not strength, we all know it. So referring to professional ranks doesn’t help. Our goal is not to copy professional system.
Our goal is to make a system that would allow completely random people who probably haven’t met each other before find worthy opponents and get the correct handicap if they want to. In order to achieve that we first take Glicko2 that assigns players rating so the players can see their relative strength. But what to do with handicaps? Well, we take Glicko rating and assign ranks on top of it. We can assign ranks however we want but it would be reasonable to assign them so one rank difference roughly corresponds to one stone handicap.
Glicko2 only has rating (and deviation, and volatility), we assign ranks on top of that however we want. It’s nice to have them correspond with handicap stones because that way players of different strengths can have interesting games using handicap.
And what if those handicap stones don’t mean anything to certain rank ranges?
What if playing with handicap stones doesn’t correlate to how well you would play without them?
How useful would they be then?
Nobody forces you to play with handicap if you don’t like it. But it’s hard to tell how well you play in a competitive game without handicap against a much stronger opponent. The weaker player will just be crushed in 99.9% of the games.
You can avoid this problem by only playing against players of similar skill. But then you don’t really need ranks. Just ratings are enough (on whatever scale you like, Glicko2, WRH, Elo, …).
Actually I just give reverse komi. It avoids every issue with handicap stones and still makes it a fair game.
How do you know how much komi is fair for 8 ranks difference, or perhaps for a difference of say 300 Glicko2 points?
That table is actually invalid by a factor of 2. It even says so on that page:
It makes the mistake of equating one handicap stone with komi instead of 2*komi.
Yes I know.
This is an experimental table for figuring out reverse komi on 19x19 games.
Recently, I used KataGo AI to estimate the komi value of handicaps.
https://lifein19x19.com/viewtopic.php?p=253248#p253248
I think that table is much more accurate than the table on senseis.
Yes, individual pro ranks have a fair amount of variability, but still I think it’s safe to say that the overall range of pro ranks has always been 2-3 full stones handicap and top amateurs have always overlapped with the lower end of the pro range (implying 1-2 full stones handicap between top pros and top amateurs).
So even if we cannot claim accuracy of individual pro ranks, we can still say that for handicap purposes, pro ranks in general overlap with amateur ranks from 7d EGF to 9-10d EGF.
But they literally say on the page I linked
The rating system is derived from ELO rating system used by International Chess Federation (FIDE).