Our guest this week is one of the world’s leading experts on chess ratings. Mark Glickman is a senior lecturer on Statistics at Harvard University, from which he completed a Ph.D. in statistics. Mark has been researching and sharing ideas about ratings for decades, and has been the chairman of the US Chess Ratings committee since 1992. He is the inventor of the Glicko and Glicko2 rating systems, versions of which are used on liChess and Chess.com. He is also a USCF Master, and has written and co-written several papers on ratings. In our conversation, we review the history of chess ratings, the efficacy of the “Elo rating system,” rating deflation and other potential issues with the current online and OTB systems. This was a fascinating conversation during which I learned a great deal! More info and shownotes below.
Click here to download the episode
0:00- We begin by discussing how Mark became interested in ratings, and a bit about the history of chess ratings.
Mentioned: Arpad Elo, Kenneth Harkness Bradley-Terry Model, Ernst Zermelo
15:15- Mark tells the funny story of some work they did developing a “rating” model for Tinder to improve their dating app.
17:00- Why have ratings been such a boon for the popularity of chess?
21:30- Patreon mailbag question: “What does Mark think of IM Ken Regan’s concept of Intrinsic Rating?”
Mentioned: Dr. Ken Regan’s paper on Intrinsic Rating, Episode 136 with IM Kenneth Regan
28:00- Does Mark think the FIDE and USCF rating systems are working?
34:00- What is the history of Mark’s “Glicko system” and where does the name come from?
Mentioned: Dr. Christopher Chabris
37:30- Perpetual Chess is also brought to you in part by AImchess.com. Check out the site, and if you decide to subscribe use the code Perpetual30 to save 30%.
38:00- Perpetual Chess is brought to you in part by Chessable.com. Check out what's new from them here: New Chess Courses Online - For All Levels- Chessable.com Don’t forget to subscribe to the How to Chess podcast as well!
41:00- Patreon mailbag question: How do we know if rating gains reflect a gain in strength rather than just that they were underrated to begin with?
47:30- Patreon mailbag question: What could FIDE and US Chess learn from the advances in rating systems that have been made by online video game sites?
Mentioned: Trueskill Ranking System
52:00- Has Mark had any conversations with FIDE about implementing changes to the rating system?
54:00- Perpetual Chess is also brought you in part by Chess.com!
Among their many resources, you can check out their Drills here and their vision trainer here.
54:30- Perpetual Chess is brought to you in part by Chessmood.com
Chessmood offers a huge library of instructive videos. You can hear GM Avetik Grigoryan on Episode 192 of Perpetual Chess. Be sure to subscribe to their YouTube channel here.
56:00- Mark has written that the ELO rating system may have a deflationary tendency. Mark explains why this is, and suggests some potential fixes.
Mentioned: Pandemic Lag by Dr. Kenneth Regan
1:06:00- Mark reflects on his own competitive chess career and shares a few stories and study tips from his journey to USCF Master.
Mentioned: Alexey Troitsky
1:14:00- Are there any lessons to be derived for chess from the sports analytics revolution?
1:18:00- Could Mark envision a universal rating system across live and online sites?
Mentioned: Rating conversion site here
1:20:00- Thanks so much to Mark for sharing his knowledge. You can keep up with him via his website:
http://www.glicko.net/index.html
You can read his papers here:
http://www.glicko.net/research.html