Yes, on player pages you can switch between the old and the new database with a link on the page. Try it out on my page if you want: Dave_de Vos | Player card | E.G.D. - European Go Database
Yes, one of the adaptations was to improve the alignment between predicted winning probabilities and observed winning probabilities (between different ratings).
We didn’t see a clear neccessity to change it into a more complicated rating system such as Glicko2 or the AGA rating system. The update even made the EGF rating system slightly simpler (one policy rule became mostly redundant, so we removed it).
The system uses a volatility function (“con”) depending only on the player rating. It doesn’t take into account that some players play a lot on the internet. But how could it possibly know where and how much everyone plays on the internet? The system not clairvoyant. And if it knew somehow, how would it take that information into account?
Everything on the internet tends to be more fleeting and changing more quickly than their real life counterparts, so quicker changes may be more desirable for online ratings and long term rating stability would be less important (people expect more instant gratifcation online).
But I don’t think Glicko is inherently quicker. The EGF rating system could also be made quicker by increasing the volatility, but the drawback would be that EGF ratings would oscillate more. The ratings would become more “noisy” and many players would feel forced to change their declared rank up and down every time they participate in a tournament. Many long time tournament players feel some attachment to their EGF rank and they would prefer it to be fairly stable over time.
So EGF ratings tend to be more official and more focussed on a longer term (potentially spanning multiple decades). Also take into account that a typical EGF player might play only a few 10s of EGF rated games a year, while a typical online player might play 100s or even 1000s of rated internet games a year. So the EGF rating system has to work with much less data than online rating systems and you don’t want single game results to give wild rating swings.
In summary: We made a conscious trade-off between stability and responsiveness of EGF ratings.