One issue I see with this is that a Glicko RD is an extended stat, and therefore premium only. So free members wouldn't see it (not even their own RD), and it might be confusing why you can't accept a seek where you fulfill the rating criterion.
More visible Glicko RD
Especially on websites. The calculations are kind of "in the dark". But if it was up to me the Glicko would be removed for the time tested, simple and elegant Elo system developed by Dr. Elo.
More complicated systems can be made, but one of the first rules in statistics is to make it as simple as possible without losing significant accuracy.
Dr. Elo followed that principle.
Actually, I believe that elo loses significant accuracy in comparison to glicko on the web-the patricipation is quite different from OTB tourneys, ratings are gained in shorter intervals.
However, that's not really the topic at hand.
Actually, I believe that elo loses significant accuracy in comparison to glicko on the web-the patricipation is quite different from OTB tourneys, ratings are gained in shorter intervals.
However, that's not really the topic at hand.
They are quite different, but as I have posted, does minutes, or seconds make you a worse player? Dr. Glickman thinks so. If THAT was true, Glicko would be more accurate in OTB games than (longer time intervals) than internet.
Back to the original statement. Posting the RD in someways is already done. In the details section of a game it tells you how much you will go up or down with a win, loss or draw. No calculations required. And anyone can tell movement for the first game of a 1200 rated player is going to be much more that the 1500th game of a 1300 rated player. No RD necessary.
I did notice in the extended statistics on a player's home page, it lists the Glicko RD.
Take a look. It is there.
Only for premium members. So making it a seek option is very awkward-free members won't know why they can't accept seeks where they meet the rating criterion.
Puchiko, I think the Glicko RD should be visible to all users and then it could be used as a criteria in starting new games.
JollyPlayer, I find the Glicko RD useful and I do check other player's Glicko RD to determine how accurate their Chess.com rating is.
Premium members should be able to search by top and bottom ratings of the 95% confidence interval, rather than just by the ELO rating. The statistics page should also display a players own 95% confidence interval based on the Glicko + ELO.
As for Glicko vs ELO, once you get rid of the idea of ratings being zero sum by having provisional ratings that move quickly initially, its much better to do a lot to try and make ratings accurate.
Glicko makes a lot of sense. Instead of just having an ELO rating, you have an ELO rating that measures the center of the confidence interval of your rating, and a Glicko number that measures how accurate your rating should be (under certain statistical assumptions that don't handle a lot of things).
Take my CFC example. I'm in the 96th percentile of players on here for live chess (generally play 30/0 time control), but in real life I have a CFC rating of 1420ish from 9 years ago. If I play in real life and beat an 1800 player should they lose rating as if they just lost to an actual below 1450 player? Or is the reasonable assumption that my rating isn't all that accurate because I haven't played tournament chess in 9 years and I should go up faster, and they shouldn't lose as much?
Now that I know what Glicko RD is, I think it should be visible next to a person's stats on their profiles. Also, when starting a new game, it would be nice to select your opponent's rating and Glicko RD.