Oh this thread turned somewhat fighty real fast o.o’
Like i said, i dont personally care one way or another since this doesnt concern me in any way, but i wonder what harm would it do if we had the visual representation go down to 30k? We have ratings that goes below the 25k threshold, and ranks are just more human-friendly way of representing that rating. Why not to have it go down to 30k, its just a number lead from ratings? It doesnt have to affect matchmaking or how handicaps are used (since most games here are played without handicaps anyway)
Tho, i’m not sure would it fix the problem of new players getting beated by others who are in the lowest rank. New players do lose a lot, thats just how it goes and its important part of the learning curve. Actually it would just move the problem from 25k to 30k, not much else would change.
The lowest rank will always be a mix of total beginners who are losing a lot, and little more experienced players who are constantly beating those beginners but losing against bit stronger players. It does not matter how low do the ranks go, it will always happen at the lowest point.
I can remember well when ogs had the 30k slum, since i was one of the many players who lived in that brutal slum without ever reaching 29k. I’m pretty sure it wont get as bad with glicko2 system, but the lowest rank will always be a harsh place with mixed skill-levels, and it will always be big step for beginner to reach the 2nd lowest rank.
“Visual representation” is all that is being proposed. The actual rating already goes down there, it is only the visual representation that doesn’t…
… in fact this thread could be called “Give a visual representation to ranks below 25k” because that is all that’s missing as far as I know: the glicko is used the same in all other respects. AAUI.
I have a technical question (that might sound stupid): Is the glicko rating of let’s say 750 (example, I don’t know the actual lowest possible number) equal to a zero in a way, is there an actual zero, or does glicko start from 750 thinking “any player has at least that rating, meaning they can tell black from white and that there are lines present”? This is not strictly on-topic, just trying to educate myself.
Not really gossip-worthy, but… Few years ago when ogs had still used elo and the lowest rank was 30k, there was huge amount of 30k players (including me) who were mainly playing with eachothers. It was also possible to set your own starting rank when creating an account, so many experienced players (mainly from kgs) made 30k accounts “to see how high they can climb on ogs”
This made 30k rank a real dump with huge variety of different strengths, so “real beginners” like me had no way of ever reaching 29k. It was solved in aug/sep 2014 by having serverwide mass rank adjustment, where everyones rank across the userbase was raised. As a result my rank got from 30k to 20k overnight, so i’ve never been 21-29k myself ^___^
I run a youth go club and IME ranks go a lot lower than 25k in reality. Ranks in my club go down to 42k. But it seems that 6 year old novices are a bit weaker still. They need some time to grow into the 42k rank that I give automatically.
A game between a 25k and a 35k is not a coin flip at all. In fact, you can determine the correct handicap between such players pretty accurately (on 9x9 and 13x13, because at this level I don’t let them play on 19x19 yet).
I use handicap tables that seem to translate quite well between board sizes. I’ve been using those tables for 8 years with only some minor updates along the way. I’m 3d EGF and I try to update their ranks so that I score close to 50% against all players in my clubs with the handicaps from my tables, so that every player is properly ranked within the club community.
These are the handicaps I give on 9x9 (without komi):
8 stones ~ 42k
7 stones ~ 36k
6 stones ~ 30k
5 stones ~ 24k
4 stones ~ 18k
3 stones ~ 12k
2 stones ~ 6k
These are the handicaps I give on 13x13 (without komi):
13 stones ~ 23k
12 stones ~ 21k
11 stones ~ 19k
10 stones ~ 17k
9 stones ~ 15k
8 stones ~ 13k
7 stones ~ 11k
6 stones ~ 9k
5 stones ~ 7k
1000s of games were played in my club over the past 8 years with these handicap tables and it works pretty well IME.
Does it make a difference that in a real club most players climb the ranks fairly quickly due to having a teacher or stronger players to discuss and learn from whereas online I guess most people are just learning by trial and error.
These are young children. They don’t study by themselves. Most of them only play a few games on small boards once a week and that’s it.
Also, there are great differences between players. Some of them are 15k-10k after 1 or 2 years, but many don’t seem to be able to break through 35k. They enjoy playing, but they don’t care much about improving (and I don’t mind either, I just want them to have fun playing go).
The colors in the tables are my way of expressing player ranks in my club (most of them don’t even know about kyu/dan ranks).
The large color bars are diplomas and the small color bars are patches (similar to the system of colored belts and patches used in judo and other martial arts).
I have much less experience with adults, but I taught some parents who were interested (some have more than one child in the go club).
From this (limited) experience, I feel that adults progress much more quickly than children (in terms of number of games played, not in calender time, because the parents play much less frequently).
Most parents reach 30k after only a (few) dozen games on small boards (if they keep playing), but only the fastest children progress that quickly.
One difference is that the parents ask questions and they want to understand how things work. They are better at logical reasoning and learning actively.
Young children have a much shorter attention span, so I only teach them a little bit at a time and they need much more repitition than adults. So it’s a slow process. Adults learn much more efficiently.
With children it seems more like a process of gradual assimilation, learning happens almost by accident. Most children prefer to just play and learn by trial and error.
Many children do seem to like solving tactical problems (if they are easy enough), but after solving a few, they have enough of it and prefer to go back to playing.
With the recognised ranks extended to 35k such a user (30k in your example) would enjoy the same choices as the rest of us. ie. They could still seek a no-handi game if they wished using custom game or quick-match prefs.
There is a definite question mark over how well auto-handi will work below 25k but that seems to be an ongoing issue throughout the whole OGS ranking system anyway and so I count it as a separate issue and as mentioned above, players can choose to seek auto-handi games or not.
That might be part of it in terms of being able to perceive progress but for me the larger issue is allowing 35k-25k users full access to OGS functionality in terms of match-making, viewing their progress, viewing opponents level and even custom tournaments.
Bollocks. If this were the case they would all end up with around the same Glicko2 rating.
This deserves thought. It might be worth adjusting defaults based on rank even though people can ofc change these settings themselves.
Unfortunately 25k and lower players are even less represented on forums than the average. I see an advantage to OGS in effectively offering more of it’s features to more members, in this case those ranked below 25k who are often newer to OGS.
Currently tournaments with NO rank restriction still show a 25k-9d range limit.That would/should change as part of this.
A monster that OGS should slay! (or at least make irrelevant).
Irrelevant. It typically takes longer to get from 9k to 1k than it does to get from 19k to 11k. This is simply diminishing returns. Some people will never reach 1k, some will never reach 11k and some will never reach 21k but that doesn’t mean they can’t enjoy Go. And enjoy OGS fully.
I can’t give exact numbers from this chart but there is a significant % of users below 25k(~10%), a fairly small % below 35k(1.5%) and for comparison less than 1% above 4d.
These figures predate CV-19 which may have changed things significantly, probably in favor of the proposal.
