The seven real causes of a plateau, how to recognise which one is yours, and what to do about each. Playing a lot is not training.
It is one of the most frustrating feelings in chess: you play every day, you watch videos, you do puzzles, and your rating has spent six months moving inside the same forty points. The cause is almost never talent. It is the method, and it is usually one of these seven.
This is number one by a distance. Playing a game uses what you already know; training puts something new in. They are different activities and only one of them raises your level.
An afternoon of twenty blitz games is an afternoon of rehearsing familiar patterns. That is fine — it keeps you sharp — but it adds nothing that was not already there. If ninety per cent of your chess time is playing, the plateau is not a mystery.
What to do: make at least half your time study and analysis. If you play five hours a week, two and a half should be something else.
Analysing is not running the engine and looking for the red number. It is reconstructing what you thought at the time and comparing it with what the position was asking for.
Without that step, a lost game becomes an anecdote. With it, it becomes the table of contents for your next month of work.
What to do: after every serious game, write three lines before turning anything on: where it started to feel wrong, which move you hesitated over, and what plan you were following. Only then, the engine.
Everyone studies openings. They are entertaining, they are everywhere, and they give you the feeling of progressing. And yet, in any set of games below 1800, the overwhelming majority of losses are decided by tactics and one-move blunders, not by the opening.
Studying what you enjoy is the most pleasant way of standing still.
What to do: sort your last twenty losses by cause and work on the biggest bucket, even if it is the dullest.
Tactical puzzles work, but only when done properly. Doing them badly means moving fast, trying the move that looks right, and hitting "next" if it fails. That trains pattern recognition, not calculation.
What to do: a short set with rules: do not move until you have seen the whole line to the end, and if you get it wrong, look at the solution until you understand it, not until you recognise it. Ten puzzles done that way beat a hundred rushed ones.
If all your chess is three minutes, you are training reaction speed and nothing else. Long calculation, evaluating a closed position and endgame technique all need clock time in order to exist at all.
What to do: one long game a week, written down if you can. Just one. It will be the one that teaches you most.
Every new opening costs months of games before the typical positions become familiar. Changing every two months means starting from zero forever: you never reach the part where the opening starts paying off, which is when you recognise the structure without thinking about it.
What to do: pick a solid repertoire and stick with it for a year. Not the best one: the one you understand.
The least glamorous and the most real. Chess is calculation, and calculation degrades with bad sleep, with playing at two in the morning, and with stacking games after a loss.
The tilt streak — playing on while angry to win the points back — is responsible for more rating drops than any technical flaw.
What to do: two losses in a row, you close it. This is not monastic discipline: it is that by the third game it is not really you playing.
Put your last twenty serious games on the table and answer:
The answer is usually uncomfortable, and that is the sign it is the right one.
Before hunting for the cause of a plateau, it is worth ruling out that there is no plateau, because it happens more often than it seems and the wrong conclusion leads to changing a method that was working.
The rating lags. What you learn this month shows up in your games one or two months later, because it first has to become automatic. A flat month after a month of work is completely normal.
Your opponents rise with you. Rating-based pairing means staying level costs more every month: to hold the same number against steadily stronger people you have to be playing steadily better. A stable rating over three months, on a site with a growing population, is disguised improvement.
The first thing to improve does not show in the result. Before you win more games you lose fewer to oversights, manage the clock better, understand more positions. All of that happens weeks before the number reflects it.
The way to check is to stop looking at the rating and count: of your last twenty losses, how many were from dropping a piece? If that number is lower than three months ago, you are improving whatever the rating says.
If you genuinely are stuck, this settles it in an afternoon and is far more reliable than any intuition:
Take your last twenty losses. Do not look at them with an engine. For each, note one thing: which move it turned on, and why.
Sort them into five groups: (1) I dropped material; (2) I missed a tactic from my opponent; (3) I did not know what to do and drifted into a worse position; (4) I lost a won position; (5) I lost the ending or ran out of time.
Count. The largest group is your problem, and it will almost always explain half or more of the losses.
And work that group for three months. One only. Each group has a different answer: group 1 is the checking routine, 2 is daily tactics, 3 is the middlegame, 4 is conversion technique, 5 is endgames and the clock.
That exercise costs an hour and is the closest thing there is to a private lesson without paying for one. What makes it work is that the answer comes from your games rather than from general advice, and general advice is exactly what chess has too much of.
And an uncomfortable part, because sometimes a plateau has no technical solution:
You do not improve all the time. Nobody rises continuously. Everybody has stretches of months without moving, professionals included, and expecting otherwise guarantees frustration.
Your ceiling depends on the hours you are willing to put in. If chess is your evening hour, your ceiling is where it is, and no method changes that. That is a legitimate decision and it is worth making on purpose rather than making it and then being frustrated by it.
And sometimes what is missing is rest rather than study. Exhaustion disguises itself as a technical plateau with remarkable ease. If you have been playing worse than usual for three weeks and changed nothing, the answer is probably two weeks of playing less, not a new method.
None of this is an excuse not to work. It is that the right question is often not what to study but how much you want to put in and whether you are rested, and those two get answered far faster than the first.
The pattern among players who break out of a plateau is always similar, and there is nothing spectacular about it:
It is slow, it is unglamorous and it works. The alternative — five hundred more blitz games hoping something shifts — you have already tried, and that is why you are reading this.
How long is it normal to be stuck? Three or four months is completely normal, even working well. More than six months with no change at all — neither in rating nor in the type of error — is a sign the method is badly allocated.
Could I have reached my ceiling? Below 2000, almost certainly not. Almost everybody's real ceiling is far above where they get stuck, and what runs out is not capacity but method, which stopped attacking whatever is now holding you back.
Should I change opening to unblock myself? It is almost never the cause and it is the solution everybody tries first. Before changing, count which move your losses are decided on: if it is after move twenty, the opening has nothing to do with it.
What if I study but do not play? Then you will not improve, and it is one of the commonest cases. Knowledge is installed by playing; without long games, study stays information that never reaches your hand.
If you want to put this into practice with structured material:
Four classes and eight hours of training to master modern evaluation, read the engine, handle complex positions and improve your endgame technique.
ARBy GM Andrés Rodríguez
A six-stage course to improve position evaluation, understand pawn structures and develop practical resources in openings, middlegames and endgames.
ARDFBy GM Andrés Rodríguez and GM Diego Flores
A collection of six courses by GM Andrés Rodríguez to expand your opening repertoire, discover brilliant combinations, and improve your understanding of pawn structures and piece trades.
ARBy GM Andrés Rodríguez
The five-step method for getting everything out of a game, why the engine goes last, and what to write down so the analysis becomes a study plan.
Read the articleWhat each time control trains, which one is good for what, and the ratio to play them in if your goal is to climb.
Read the articleThe four ways to throw away a winning position and what to do about each. An advantage does not defend itself.
Read the articleARDFBy GM Andrés Rodríguez and GM Diego Flores
