@flovo I have some questions about the Glicko2 point table we have in our profile. I tried to find some explanation on the forum but it seems my searching skills dropped with covid confinement.
Looking at this table I have a lot of questions:
- Looking to a column, the last one for example, how can be possible to have a general correspondence rating higher than the 19x19 correspondence rating if there is no other game in smaller go board played? How can the general value increase?
- The combination never played (like for example live game in 9x9 in the example above) that are shown dimmed with a default value of 1150 +/- 350 do contribute to the general rating of the column?
- If the answer at point 2) is yes, Does it make sense to play some game in such combination with the objective to raise a so low value affecting the general value of the column?
- if the answer at point 2) is no, Does it make sense to avoid such a combination if you don’t like at all - for example - playing at 9x9? Otherwise, a single bad performance can affect forever your general value of the column.
- Is it relevant the number of games played in each combination for the corresponding general value? Or it affects only the error bar?
- Why the estimated ranking in terms of kyu/dan is reported only in the general upper left corner plot and not in any other combination? Does make sense to have different kyu/dan levels in different conditions/sizes?
- why there is a mismatch between general rating in decimal kyu/dan values calculated in the plot and the number reported above your avatar? Do these numbers come from the same algorithm?
I don’t want to chase you if you already answered to these questions. Even a well-pointed link is enough for my curiosity. Thank you in advance.