Difference between chess puzzle and chess problem?

Question

While a chess puzzle is any puzzle involving aspects of chess, a chess problem is an orthodox puzzle in which one must play and win or draw a game, starting with a certain composition of pieces on the chessboard, and playing within the standard rules of chess.

There's this episode of Queen's Gambit where Beth says she doesn't do 'chess problems' because they don't come up in real games. Meanwhile, it appears Lichess gets all its 'puzzles' from games. Soon after Beth is enticed to a mate in X 'problem' (though described as a 'puzzle' here).

Is there any relevant difference between puzzle and 'problem' here?

I cannot imagine any chess player not doing puzzles if they want to increase their rating, except for my dad who once told my (paternal) cousin (who attended university in a chess scholarship) and me half-jokingly 'Stop doing those puzzles. You will never get into that position.' (My dad was and is nowhere near the level my cousin was at the time. He seemed to favour more learning from pro games than puzzles.)

Idk perhaps the show really means 'problems' as puzzles and then it's an exaggeration of players who prefer more to study openings, positions, strategy, etc. than tactics.

Answer 1

The character, Beth, says she doesn't do chess problems because those situations don't come up in real games.

1. She may be speaking in a more metaphorical sense, to say that chess problems don't come up in real games the same way that they happen as chess problems, so that training with them isn't useful (to her).

   1. Chess problems are always (supposed to be) solvable and often have "clever" solutions. In chess matches, there isn't guaranteed to be a solution, and sometimes moves that appear clever backfire.
   2. Chess problems usually come with a very limited, well defined task, e.g. "white to mate in four moves". Same as before, chess matches generally don't come with that.

2. This is a reflection of the character, which may not be objectively true. The character has deliberate flaws and overconfidence or ignorance (of better methods of training) may be coming into play.

This is the same type of objection that some people have to "math problems" (of the type assigned as homework). If we consider a math problem:

Alex had two apples. Cameron gave Alex three more apples. How many apples does Alex have now?

It's surely true that people somewhere do have to occasionally solve a problem where they have two things, receive three more things, and have to know how many things they have. But they probably also don't get the problem this cleanly, it may not even be obvious what they need to calculate, and there may be a lot of other information provided that clutters up the problem. Someone who can look at a worksheet that's labeled "Addition Practice" and decide to add 2+3...may still have difficulties when faced with the more complicated, unlabeled real-world situation.

On the other hand, if this is part of the student's path to learning addition, then I think it's a reasonably useful tool.

But the general objection that chess problems aren't the same as chess matches, coupled with a flawed character, makes this statement coming from this character plausible as part of a narrative drama.

Answer 2

As a chess composer, seeing most problems being called puzzles is rather frustrating. A while ago, I personally revamped the problem and puzzles tags.

The problem tag says:

Chess problems consist of a board position and a task. Most ask for a line of play that mates black in a set amount of moves, or a combination that results in a winning position. Many other types exist, including, but not limited to, construction tasks for various positions and proof games that have numerous goals, selfmates, helpmates, and retrograde analysis. Tournaments and championships exist for composing of original problems and for speed-solving.

The puzzle tag says:

Often, a puzzle type of problem in which a tactical or positional solution is required, like White/Black to move and win”, in what is often a middlegame of endgame-esqe position. But a puzzle can also be a chess riddle, chess math. or any other non-normal chess problem. Questions about solving puzzle, puzzle apps and websites, how to understand the idea behind a puzzle, etc. are also on-topic.

To my knowledge, these are generally accepted definitions by the chess composing community. I expect @Laska to have more input here.

Answer 3

tl;dr: Chess problems must meet defined quality standards. Chess puzzles need not. (Disclosure: I am a chess problemist.)

---

Chess problems are all subject to rigorous conventions that are not that widely known, even among experienced chess players. Off the top of my head, the most important are:

• There must be a unique solution. For example, in a mate in 4 (white to move, by convention), in the intended variation, white will only have one correct move against black's replies at every step.
• Every piece is necessary. (Not in the Codex, but a basic principle nevertheless.) On the board, every single piece and pawn has a role to play to make the solution unique. Pieces whose only role is to "confuse the solver" or "make it look realistic" are not accepted.

Of course there are further aesthetic conventions; I won't bore you with the details here. (Note to other problemists: I also simplified by saying "unique solution" since it's clearer than explaining twinning and "h#2 3.1.1.1" here.)

The result is that composing good chess problems is hard. A good mate-in-two, proof game, or endgame study is really hard to get working just right, so problemists tend to get a little miffed when their creations are referred to as "puzzles".

---

Chess puzzles, as a catch-all, are:

• Everything else that is not a chess problem.

Chess tactics puzzles, like the lichess timed tactics, fall into this category. Nowadays these tend to be automatically generated or pulled from real games - relatively lower-effort (on the human side) than composing a problem. Problemists may sometimes also self-deprecatingly refer to their problems as "puzzles".

"Joke puzzles" may be those that involve giving an upside-down board or an en passant capture without justification (don't confuse those with "joke problems" that do the same things but follow the internationally accepted problem conventions for such - much harder to do, and infinitely more satisfying!)

A somewhat grey area are problems/puzzles of a more mathematical bent. "Place all 32 standard pieces on the board such that none are attacking each other," would be an example of a construction task that could be called a problem or a puzzle. "Find the unique game ending with 3...Qd4#" is sometimes called a synthetic game. These have been referred to as both problems and puzzles.

---

EDIT to address the show referenced: When Beth Harmon says "problems", she means problems as I have explained above. Not chess tactics, not the lichess puzzles, but composed problems. "Problems" is a precise term, used correctly in the show.

There is an impression among chessplayers that composed problems can be useless for improvement, as the positions may be unrealistic. To some extent this is true; complex directmates aim to show particular themes, and no problem aims to be "realistic" with respect to played games. Helpmates, reflexmates and other stipulations are also at odds with the goal of "standard" chess games.

However, doing such problems can improve your board vision (it would have to, as regular pattern recognition might not apply because the positions aren't "realistic") and your creativity (seeing the coordination of different pieces in ways that you might not have come across before).