A Physicist Looks at Chess Ratings
Vlad is a FIDE Arbiter, a FIDE National Instructor and a player rated 2020, and he has published a book called 'The Rating Revolution'. ChessBase India's Abhyudaya Ram met him during the Grand Chess Tour, where Vlad was attending as a spectator. In an interview, Vlad spoke about why the Elo system treats Europe and Asia differently, why favourites keep underperforming, and which countries are best for picking up rating points.
Interview
Abhyudaya: Welcome, Vlad! First of all, how does it feel to be at the Grand Chess Tour in Romania?
Vlad Ghita: This is my fifth consecutive year attending as a spectator. This time, the event is being held in the National Bank of Romania's palace, which is also a museum, and the playing hall is spectacular. I am very happy to be here watching the best players in the world.
Abhyudaya: Are you originally from Romania?
Vlad: I was born in a different town, but I have lived in Bucharest since the pandemic, so about six years now.
Abhyudaya: You have been around chess for a long time. How did you get into it, and what do you do now?
Vlad: I like to say I am a jack of all trades in the chess world: I do many things, but none of them at a particularly high level. I play actively. I am also a coach and a FIDE National Instructor, working mostly with adult improvers from all over the world. On top of that, I am a FIDE Arbiter and I am gaining experience to go for the International Arbiter title.
Abhyudaya: I saw you arbitrating at the Grand Swiss. What was the experience like, with so many players?
Vlad: It was a big surprise for us, honestly. A week before the event, we had roughly 300 players, but on the first day nearly 600 showed up. It was a bit nerve-wracking! We had 13 arbiters on the job, and thankfully there were no big issues in my sector, just the occasional illegal move and other small incidents.
Abhyudaya: Now to your book, The Rating Revolution. Few people write about rating systems. How did this start?
Vlad: My background is in physics. I did my bachelor's and master's in the United States, so numbers were a big part of my work, and I love statistics and testing hypotheses. My original hypothesis was that FIDE's 2024 rating reform was very good at lifting the floor but not so good at fixing the ceiling. It's like fixing the floor in your house while the roof is still leaking.
At the top level, we are seeing deflation. The 2700 club, for example, loses players every year. I didn't want to just look at the list and say "I feel like this is happening," so I built a framework to quantify it. My conclusion is that the 2024 reform was mostly useful, but FIDE should intervene more strongly.
Abhyudaya: So does the whole system need to be replaced, or just updated?
Vlad: Definitely an update. The Elo formula is beautifully simple. You can calculate your rating change by hand without any mathematical background.
One thing I found is that rating favourites consistently underperform against lower-rated opponents. With a 200-point gap, the favourite is expected to score about 75%, but in reality they score around 72 to 73%. That has been going on for a while, and the 2024 reforms helped but didn't eliminate it. So I'm not proposing a rating compression, just a slight tweak to the Elo formula to bring expectations closer to what we observe in practice.
The biggest problem, though, is the discrepancy between federations.
Abhyudaya: Tell us more about that.
Vlad: Many Asian federations have extremely underrated players. Take India: a lot of Indian players come to Europe with low ratings but perform at a far higher level. There's an economic side to it too. Those with sponsorship or the means to play abroad can gain rating points and move towards FIDE Master or International Master titles. Players who have to stay within India may not get those opportunities, and their rating doesn't reflect their true strength.
A big part of what I am doing is trying to make the rating system fair for everyone. With my 2000 rating, if I moved to Vietnam, Uzbekistan or India, I'd lose a lot of points because the players there are incredibly strong.
Abhyudaya: So European ratings are inflated?
Vlad: Overall, European federations are more inflated than Asian ones. You can test this with World Youth Championship results: European federations tend to underperform while Asian federations take the medals. Kazakhstan has been sweeping these events and winning lots of medals. There's a big discrepancy between the continents.
Abhyudaya: How can your book push federations to change this?
Vlad: I wanted the analysis to be accessible to everyone, whatever their level in mathematics. I wanted to show there's a problem, but also suggest a solution, because I don't want to just complain. I've forwarded my findings to people at FIDE who make these decisions. I haven't heard back yet, but I think the solution is out there.
National federations should also monitor the health of their rating pools. For example, India has more than 37,000 players with a K-factor of 40. That usually means lots of juniors, plus some adults who haven't yet completed 30 rated games.
Abhyudaya: Let me stop you there. For readers who don't know how the rating system works, can you explain the basics?
Vlad: Of course. In the Elo system, only the relative difference between players matters. Your number only has meaning in relation to your opponent's. If you're rated the same as your opponent, you're expected to score 50%. If you're 100 points higher, you're expected to score roughly 64 to 65%, and at 200 points higher, about 75%.
It's a probabilistic model. It doesn't matter that Magnus Carlsen is 2800; he could be 3000. If tomorrow we added 200 points to everyone in the world, the expectations wouldn't change. The system was invented by Professor Arpad Elo, and we owe him a lot.
There are other systems, of course. Mark Glickman in the US created the Glicko system that Chess.com and Lichess use, and there's plenty of debate about Glicko versus Elo. My suggestion is that FIDE tweak the existing system rather than throw it away, since I think it can still be repaired.
Abhyudaya: Many people play only online these days. How do online ratings compare with FIDE ratings?
Vlad: I've been working on this recently with the people at the Chess Dojo. Online systems have one big advantage: wherever you are, your opponents come from everywhere. On Chess.com I can face someone from Uzbekistan or the US on the same day. The pool is well connected, so if you play a lot online, your rating should reflect your true strength.
But we do notice that, say, an 1800 from Uzbekistan is probably rated much higher online than an 1800 from Romania. So online ratings are fairly accurate, but the discrepancies between federations in FIDE ratings are very large. My dream is to see that fixed within the next few years.
Abhyudaya: Beyond federations, why should chess players read your book?
Vlad: Because it contains a global map of inflation and deflation zones. If you're travelling to Europe to chase norms or rating points, there are federations to avoid and federations to target. Some might argue that's not entirely fair, but it's baked into the system. Germany and Austria are good places to play, and Serbia is also a place where you can gain a lot of rating.
Abhyudaya: Is that because of inflated ratings, or actual strength?
Vlad: Take a benchmark like a Candidate Master. Based on the data, a CM in Serbia is weaker than an average CM in Poland or Greece. So Poland and Greece are difficult places to gain rating, while Austria, Serbia and the Netherlands are good targets. There's a kind of gamification in it. But for players who want maximum value for their budget, it matters, because their chess careers are at stake.
Abhyudaya: So it helps players make more conscious decisions about where to play?
Vlad: Exactly, a data-driven decision rather than only a gut feeling. Spain is a great destination and good for gaining points too, but there may be other places where the data tells you that you'll get more out of it.
Abhyudaya: What research went into the book?
Vlad: Most of the data came from FIDE's official monthly rating lists, and I focused only on classical ratings since they're the most relevant. I downloaded every monthly list since January 2021. The pandemic hit the chess world so badly that I treated 2020 as a pause, and 2021 as the restart of the chess boom.
It began as a hobby project and grew into something much bigger: an analysis of the periods before and after the rating compression. I wanted to see from many angles whether the compression worked and what else could be done. Focusing only on the top players feels a bit elitist to me. Of course Grandmasters need to feel the system is fair, but there are lots of club players who feel it's biased against them. Making it fair for everyone will help the game grow, reach more young players and bring everyone into one big chess family.
Abhyudaya: Have any top Grandmasters read it?
Vlad: Levon Aronian was kind enough to write the foreword. Anish Giri recently received a complimentary copy, though I'm not sure he managed to read it on his flight to Bucharest! Your national team coach, Srinath Narayanan, actually bought a copy. His feedback was along the lines of "wow, good work," but he found it a bit advanced and technical. I tried hard not to make it so, though it depends on the reader's comfort with maths. It's not an academic paper. You can read it in an evening if you focus.
He is a person of immense knowledge and a brilliant mathematical mind, but more importantly, he is someone 'possessed' by chess. When Vlad looks at the FIDE rating list, he doesn't just see a list of names. He sees a complex, breathing system governed by laws and subject to invisible shocks.
- Levon Aronian
Abhyudaya: Do casual readers need any prerequisites?
Vlad: Not really, just very basic statistics, like averages, medians and distributions. I've also included lots of graphs, so it isn't dry prose. You can see what's happening from the pictures.
Abhyudaya: What does success look like for this book?
Vlad: I want it to be a conversation starter. I want the decision-makers at FIDE to acknowledge there's a problem and that more can be done. I'm not claiming my solution is the best or that I have the absolute truth, only that it's one angle to tackle the problem from. I'd love to see a joint task force working on it. I'd also like players to understand that there's a bias in the system and that we can make it better.
It's not about selling copies. It's about spreading awareness, and showing that this kind of data analysis can be done by someone passionate about chess who spent many months collecting the data. It was a passion project, since my day job is coaching, and I wanted to do something besides coaching and arbiting.
Abhyudaya: Will there be more books?
Vlad: Think of it like a movie: if the first one is successful, there's a second and maybe a third. It depends on how people respond. FIDE publishes new data every month, so I could easily extend the analysis over a longer period. For now I'm taking a short break and letting people give feedback.
Abhyudaya: Where can readers get the book?
Vlad: On my website, vladchess.com. There's a prominent button that takes you to the book's page. I self-published because I wanted independence in how I present the work. I did approach some publishers, but they didn't see the project as large-scale enough, so I decided to get the word out myself.
Abhyudaya: Is it available as an ebook?
Vlad: Yes, it's a PDF ebook download. I don't have physical copies yet, but I'd like to make some in the future.
Many players tell me the system feels unfair to them, especially youngsters who arrive with low ratings but play far better because they train on Chessable and other platforms.
Yes, I've heard the same from many players, and distilling that feedback into a written work was very valuable. So far I'm happy with the response, and I'd love more people to read it and get in touch.
Abhyudaya: And where can people contact you?
Vlad: The easiest way is X, formerly Twitter, where my handle is @vladchess. It's been my username since the pandemic. At first people thought I was Vladimir Kramnik, which is funny, because the difference in our ratings is huge! I also focus on data analysis of ratings, and I don't really talk about the cheating problem or things like that.
Abhyudaya: Thank you so much, Vlad, and all the best. I'm sure readers will find this fascinating.
Vlad: Thank you!
