Daily games causing a lot of rank variance?

[Hades12]

I have noticed in my Daily games, there is a large variance of strength between ranks. People 3-4 stones weaker than me play surprisingly even, and if they don’t make a huge jodeki or fighting error in the first half of the game, I have to rely of beating them in endgame. On the flip side, I also play opponents 3-4 stones stronger evenly, without a lot of struggle.

Having more time to think and read out variations lessons the gap of strength, since stronger players can often play the right shape or right direction by instinct.

I have also noticed ranked graphs on OGS showing people fluctuation 4 stones ranks or more, from 3D down to 1K and back up to 3D. Has anyone else seen odd rank fluctuations from daily games on OGS?

[gennan]

I think this is just a quirk/feature of the OGS rating system.

In any rating system, there is a choice by the owner on how much the ratings can change whith one added game result. You might call this a rating volatility parameter. In chess Elo systems, that parameter is called the K-factor. In the EGF rating system it is called "con". If it has a high value, ratings will be volatile (large steps). If it has a low value, ratings will change very slowly (small steps).

It seems OGS chose a fairly high value. The advantage is that new players need relatively few games to get an estimate of their rank and players can "rank up" fairly easily. The disadvantage is that ratings can fluctuate a lot.

KGS's rating system is more the opposite and players tend to complain it's too conservative for regular players.

[Ferran]

It seems OGS chose a fairly high value. The advantage is that new players need relatively few games to get an estimate of their rank and players can "rank up" fairly easily. The disadvantage is that ratings can fluctuate a lot.

Maybe it was a good idea when it was correspondence only?

Take care