Researchers studied tens of thousands of chess players and found something that upends how most students prepare for competitive exams: it's not how much you practise. It's what you do between attempts.
There is a version of JEE preparation that looks, from the outside, like everything a serious student should be doing. Six to eight hours of problem-solving daily. Hundreds of questions a week. Mock test after mock test. And at the end of it, a score that barely moves. The student works harder. The score stays put. This is one of the most common and least understood patterns in competitive exam preparation — and researchers studying chess players at scale have recently given us the clearest explanation yet for why it happens.
Chess is one of the few cognitive domains where skill is precisely and objectively measured. Every player carries a rating that reflects their actual ability against thousands of opponents — there's no subjectivity, no grade inflation, no lucky paper. This makes chess uniquely useful for researchers trying to understand how humans actually get better at hard, complex thinking tasks.
Studies examining tens of thousands of chess players — including landmark work by researcher Neil Charness and colleagues — found that the single strongest predictor of chess skill was not the number of games played, but time spent in deliberate study: analysing positions, reviewing past games, and identifying and correcting specific errors. Players who accumulated thousands of games without structured review improved far more slowly than those who played less but analysed more.
A more recent Stanford University study reinforced this with data from 96,000 Chess.com users over six months. Players who mixed in structured game reviews alongside playing improved their ratings significantly faster than those who only played. For intermediate players, game reviews alone produced an average rating gain of 31 points — with no increase in the number of games played.
The pattern that emerged across all skill levels was consistent: playing games gives you experience. Reviewing your games gives you improvement. The two are not the same thing, and conflating them is the central mistake most people make when trying to get better at something difficult.
The players who improved fastest weren't the ones who played the most games. They were the ones who understood why they lost.
Chess and JEE preparation look very different on the surface. One is a board game; the other is a three-subject, three-hour exam that determines admission to the best engineering institutions in the country. But at the level of cognitive skill development, they share a structure that makes chess research directly applicable.
In both domains, performance is precisely measurable and ranked. In both, the gap between average and exceptional performance is not primarily a gap in raw intelligence — it's a gap in the quality of feedback that the practitioner extracts from each attempt. And in both, the instinctive response to a plateau is to do more of the same thing that caused the plateau in the first place.
The chess finding that game reviews produced 31 rating points of improvement without increasing play time has a direct JEE equivalent: a student who reviews 30 mock test errors with genuine diagnostic rigour will improve more than one who solves 300 additional practice problems without understanding what went wrong.
The underlying principle that chess research keeps pointing back to is deliberate practice — a term coined by psychologist Anders Ericsson after studying expert performers across chess, music, medicine, and sports. It has a specific meaning that's easy to misread.
Deliberate practice is not practising hard. It's not practising a lot. It's not even practising consistently, though consistency matters. It is practising with the explicit goal of identifying and correcting specific weaknesses, with immediate feedback on whether the correction worked.
The discomfort point is important. Ericsson found that deliberate practice feels harder than regular practice precisely because it forces you to confront what you don't know, rather than rehearsing what you do. A student who finds mock test review easy is almost certainly not reviewing properly. Real review — sitting with a question you got wrong and tracing back to the exact moment your reasoning diverged from the correct approach — is genuinely effortful in a way that solving another fifty problems is not.
The chess research identified something that maps almost perfectly onto a pattern seen repeatedly in JEE preparation: the confusion between experience and learning.
An experienced chess player who has played 5,000 games but never analysed them is not a skilled player — they're a practiced one. The distinction matters enormously. Experience accumulates automatically; learning requires deliberate extraction. You can spend two years preparing for JEE and accumulate enormous experience — thousands of problems, dozens of mocks, familiarity with question formats — without extracting the specific lessons those experiences contain.
The equivalent in JEE is the student who has done twelve full mocks but can't tell you with any precision what their dominant failure mode is. Are they losing marks primarily to conceptual gaps, to modelling errors, to calculation slips, or to strategic mistakes like misreading questions under time pressure? If the answer is "I'm not sure," the mocks are generating experience without generating learning.
One of the consistent findings across chess expertise research is that grandmasters spend a disproportionate amount of their training time not playing games, but studying positions — particularly positions where they or other strong players made mistakes. The game itself is almost secondary to the analytical work that surrounds it.
Magnus Carlsen, the former world chess champion, has described how pattern recognition at the grandmaster level comes not from volume of games but from depth of analysis of specific critical positions. A position studied deeply for an hour — understood from every angle, with every candidate move considered and the best response found — is worth more than ten quickly played games where the same type of position appeared but was handled on instinct and moved past.
The JEE equivalent is the student who takes one genuinely difficult question they got wrong and spends forty-five minutes understanding it completely — reconstructing their original approach, identifying exactly where it broke down, working through the correct solution independently, and connecting it to the underlying concept that was shaky. That forty-five minutes is worth more preparation than solving twenty more questions at roughly the same difficulty level.
Chess players ask: "At which move did my evaluation of the position go wrong?" Not "what was the right move" — that's what the engine tells you. The question is where your thinking diverged from what the position required, and why.
The JEE equivalent: "At which step did my approach to this problem diverge from what the question required?" Was it in how you set up the problem? In which principle you reached for? In an arithmetic step that cascaded into a wrong answer? Or in how you read the question in the first place? Each of these has a different remedy — and calling all of them "wrong" without distinguishing them is the same as a chess player saying "I lost" without understanding why.
The practical implication of the chess research is not complicated, but it does require a shift in how you think about what preparation actually means. Volume is not the enemy — you do need to solve a large number of problems, and there's no substitute for familiarity with question formats and difficulty levels. But volume without analytical review is a treadmill, not a ladder.
Three things that follow directly from what the research shows:
Your review time should be comparable to your solving time. A three-hour mock deserves at least ninety minutes of structured review. Currently, most students give it ten. The ratio needs to invert.
Every wrong answer needs a category, not just a solution. There are four fundamentally different failure modes — conceptual gap, application error, execution slip, and strategic mistake — and each requires a different response. Reading the correct solution and moving on doesn't tell you which failure mode occurred, which means you can't fix it.
Tracking patterns over time matters more than individual errors. One wrong answer in rotational mechanics might be noise. Five wrong answers in rotational mechanics over three mocks, all involving the same failure to identify whether external torques are present — that's a signal worth planning a week of preparation around. The chess research found that improvement came from recognising and correcting recurring patterns, not from reviewing individual mistakes in isolation.
The shift in one sentence: stop measuring your preparation by how many problems you solve, and start measuring it by how many failure modes you've genuinely understood and corrected.
A chess player who analyses fifty lost games deeply becomes a strong player. One who plays five thousand games without analysis stays mediocre. The same principle, the same research, the same outcome — just with a different board.