Are classic puzzles useless?

No. They train a real skill — seeing opportunities for yourself — and the data shows that skill improves well. The problem is not what they teach, but what they do not. Keep doing puzzles, and add specific defensive work on the side: reading threats, saving pieces, positions to hold.

What exactly is a "missed defense"?

It is a move where the player loses material or the game while a defensive resource existed on the previous move: a retreat that saved a piece, a parry that avoided a fork, a trade that defused an attack. It is the most frequent mistake mechanism in the Observatory corpus, and one of those that recede the least with rating.

Does a defensive player win more games?

That is not what the data says; it measures mistakes, not a style. What it shows is that defensive mistakes cost a lot and do not disappear on their own. A player who attacks well and defends well wins more than a player who attacks well and defends poorly — and the second case is the norm among amateurs.

Where should I start if I want to work on my defense?

With reading threats: after every opponent move, before looking for your own, ask what it attacks. That is the gesture that feeds all the others. Then work on saving attacked pieces, then on positions to hold (king under attack, inferior endgame). Each gesture matches a mistake category you can measure in your own games.

Why is defense harder to teach?

Because a defensive exercise has not an answer but a proof: you have to show that the proposed move holds and that the others do not. The solution is often an unglamorous move, hard to showcase. And defensive mistakes are punished late, which keeps the student from connecting cause and effect. Those three difficulties explain why the market produced twelve attacking exercises for every defensive one.

← Back to blog

Why Defense Never Gets Learned in Chess (and How to Train It Anyway)

October 6, 2026 · ChessPivot · Coaching

Episode 3 of the series "The Immortal Mistakes", drawn from the Amateur Chess Observatory, Fall 2026 edition.

Open any puzzle app. In almost every position, it is your move, and a winning move exists: a fork, a mate, a sacrifice. You look for it, you find it, the score goes up. You have just trained yourself to take.

Now open your own games. Most of your losses do not come from a combination you missed. They come from a piece you left hanging, a threat you did not see, a rescue you did not look for. You lose because you do not know how to hold — and nothing in your training ever taught you.

That is not an impression. It is what 52,005 amateur games measure.

What the games show: two learning speeds

The Amateur Chess Observatory tracks 38 mistake mechanisms, band by band, from 400 to 2399 rating. For each one, it compares the frequency in the lowest band to the frequency in the highest. The ranking that comes out tells a single story.

The mistakes that fade have one thing in common: they are opportunities for yourself that you learn to see. The mistakes that hold on have another: they are things you hand to your opponent. Seeing material, yours and theirs, gets learned by playing. Reading what your opponent is preparing barely gets learned at all.

The headline fact of the edition says the same thing another way: of 13,754 pieces lost after being left hanging, 47% could still have been saved by a quiet move — and that proportion is identical from 400 to 2399. Strong players hang three times fewer pieces. Faced with an attacked piece, they do not save it any more often.

Why: defense is almost never trained

A skill that does not improve over 1600 rating points is not inevitable. It is a skill nobody trains. Two numbers, verifiable by anyone.

Puzzles. The most used tactical-motif taxonomy in the world, the one behind Lichess puzzles, counts 25 motifs. One is defensive ("Defensive move"). The other 24 — fork, pin, skewer, discovered attack, named mates — teach you to win something.
Our own banks. Before this study, ChessPivot’s exercise banks counted 59,629 positions where you must find the winning move, and 5,122 where you must hold. Twelve to one. We did no better than the rest of the market.

That imbalance is not negligence. It has a practical cause, and it is worth understanding so you do not reproduce it in your own training.

The practical cause: a defensive exercise is hard to build

An attacking exercise has one answer: the winning move. You check it with an engine in a second, you grade it without ambiguity, you produce a thousand a day.

A defensive exercise is of a different nature. The right answer is often "the move that does not lose", and to validate it you have to check that every other move loses. Two moves sometimes hold equally. The right move is sometimes an unglamorous retreat nobody wants to showcase. It is easier to produce a thousand attacking puzzles than a hundred defensive positions — and that is what the market did.

What "defending" means in practice

Defense is not a passive posture. It is a set of precise actions, and each one matches a mistake measured in the corpus.

Saving an attacked piece with a quiet move, before looking for a counter-attack. The missed defensive resource is the most frequent mistake in the corpus, at every level.
Reading the threat contained in the opponent’s last move, before looking for your own. Every hanging piece, every fork suffered, comes from an unread threat.
Not weakening your king without a concrete reason. The weakened king is the mistake that best resists improvement.
Trading at the right moment, meaning not simplifying when you are worse, and not keeping pieces on when you are better.
Finding the saving resource in a lost position: the stalemate, the perpetual check, the fortress. These resources exist far more often than they get looked for.

Three habits to train defense without waiting for tools

You do not need an exercise bank to start. Three habits, usable from your next game.

After every opponent move, ask what it attacks now

Before looking for your move, re-read the position from your opponent’s point of view. Which piece just changed its line? Which square just got uncovered? That is the basic gesture of prophylaxis, and it is the one improvement does not install by itself.

Piece attacked: look first for the move that puts it out of danger

Before considering a counter-attack or accepting the loss, exhaust the simplest question: is there a quiet move that saves this piece? In the corpus, that move existed one time in two, and was not played.

Solve puzzles backwards

Take a classic puzzle position — you have the winning tactic. Go back one move, and ask what the other side should have played to avoid it. You have just turned an attacking exercise into a defense exercise, without changing the material.

Common mistakes

Believing you defend by attacking. The counter-attack is a resource, not a rule; in the corpus, it is the first cause of lost pieces that could have been saved.
Only working on puzzles where it is your move and a winning move exists. They train a real skill, but only one.
Treating lost positions as finished. That is exactly where defense is played, and where most players mentally resign before their opponent does.
Confusing caution with defense. Playing h3 "to be safe" is not defending: it is often creating the weakness your opponent will attack.

Conclusion

Chess training, as it exists, teaches taking — almost never saving. The data from 52,005 games shows the result: attacking mistakes fade with rating, defensive mistakes stay. The good news is that the missing skill can be named, measured and trained. ChessPivot’s king safety exercises and defensive banks now exist for that, and an analysis of your games tells you which of these habits are missing from yours.

Frequently asked questions

Are classic puzzles useless?: No. They train a real skill — seeing opportunities for yourself — and the data shows that skill improves well. The problem is not what they teach, but what they do not. Keep doing puzzles, and add specific defensive work on the side: reading threats, saving pieces, positions to hold.
What exactly is a "missed defense"?: It is a move where the player loses material or the game while a defensive resource existed on the previous move: a retreat that saved a piece, a parry that avoided a fork, a trade that defused an attack. It is the most frequent mistake mechanism in the Observatory corpus, and one of those that recede the least with rating.
Does a defensive player win more games?: That is not what the data says; it measures mistakes, not a style. What it shows is that defensive mistakes cost a lot and do not disappear on their own. A player who attacks well and defends well wins more than a player who attacks well and defends poorly — and the second case is the norm among amateurs.
Where should I start if I want to work on my defense?: With reading threats: after every opponent move, before looking for your own, ask what it attacks. That is the gesture that feeds all the others. Then work on saving attacked pieces, then on positions to hold (king under attack, inferior endgame). Each gesture matches a mistake category you can measure in your own games.
Why is defense harder to teach?: Because a defensive exercise has not an answer but a proof: you have to show that the proposed move holds and that the others do not. The solution is often an unglamorous move, hard to showcase. And defensive mistakes are punished late, which keeps the student from connecting cause and effect. Those three difficulties explain why the market produced twelve attacking exercises for every defensive one.