In other popular online games, there’s a relative peak in players at each new rank. Psychologically, this makes sense–you stop playing for a bit after reaching a new rank, satisfied with having ranked up. In OGS’ rank distribution we don’t see this. Is it just visual or is the rank distribution algorithm artificially forcing a Gaussian shape?
3 Likes
looks like it actually plotting 31 columns but in smooth style
previous attempts actually result in almost perfect shape
if this one using much more data, its possible to believe shape would be more perfect
but looks like too many inactive users are included this time so it became less useful/meaningful
by the way,
last_updated":"2025-03-28
What peaks are you expecting to see, one at all the ~40 integer kyu and dan ranks?
1 Like
Ogs doesn’t do rating-resets after reaching a new rank. So when you reach new rank, it might go from 1.8d to 2.1d for example. So no peaks at specific intervals
2 Likes
Still, dips before each new rank can be expected. If you’re close to the next rank, you might play another game to try and get it. Then you either win or lose, so your rating changes away from close to the next rank.
3 Likes
Some ideas:
- More ranks overall
- Many different ways rank is displayed so no universal stopping point
- Hard to stop playing your correspondence games
- Your rank can adjust when not playing if games get annulled
2 Likes
Kinda surprised its so prominent on the chess graph. Is it all attributed to psychology, or is there something else at play?
2 Likes
my few cents:
- Chess players in my experience are even more rating focused than Go-Player
- Chess players look directly at the number. The majority of player here look at the rank that is a result of that number and actually don’t really know how close they are to ranking up or down. If you play on Lichess and got a rating of 1801 you know you’ll be below 1800 if you lose. A player here might have a rating of 1801 which makes him an 1k and will rank down if he loses the next game, but he does not know for sure unless he looks up his rating (don’t know the rating threashholds out of my head, pulled them out of my arse).
- I’m for both graphs not sure if the way they are plotted does not reinforce / obfuscate it. I don’t know how the Lichess graph would look like if they would have chose smaller brackets, I don’t know if the OGS graph get’s smothed im some way
2 Likes
I guess then if it’s just visually smoothing based on the bar graph above it could be that peaks exist at some scale that’s not visible in the bar graphs.
Also, while I agree with you @Folivoro about people being less rank focused than in chess, I cannot imagine there is no peak whatsoever after 1d on OGS haha