In a normal tournament, you meet so many different people (and you often don't know in advance exactly who you will meet of all the participants), so it's hard to specifically prepare. For a match of this kind, however, you know your opponent will prepare specifically against you. So you kinda have to do the same.

That entails analyzing all of their games and finding defenses and weaknesses. But also trying to find new novelties in the openings etc. And since the opponent will do the same, you yourself have to prepare defenses for lots of potential new openings 16 moves deep or so. It's an insane amount of studying.

Of course, that pays off for future games outside the match as well. But when you know your challenger will spend half a year on this once-in-a-lifetime opportunity to beat you, you probably have to do the same. So have to say no to everything else.

the top 100 players in the world is a small list, and at this level you don't need to worry about anyone below that (until they get higher) as you are enough better than them that you can beat them. Thus for tournaments you don't have worry as you have already studied everyone you need to worry about before the tournament. Odds are you know ahead of time who is in the tournament and so you have a couple weeks to study all possible opponents.

Remember for the best players in the world chess is their full time job. They spend 12 hours a day on the game. They are earning enough from tournaments to support life.

Say a player just rote memorised moves from a chess engine and just played it. Would that be considered cheating?

No, that's part of the preparation. But remember that you have around ~16 moves to choose from each time it's your turn. And you have to prepare a response for all the possible choices your opponent can make.

This of course gets unwieldy very fast. 16^5 is already over a million different games.

Actually it's so much worse than that. Chess has a branching factor of ~30, not 16, IIRC. A lot of those moves will be nonsensical though.

That is correct. And of course you don't just need to prepare for your own moves, but also consider your opponent's moves. So "just" 6 moves = 30^12 = about 500 quadrillion possibilities.

Though, like you said, the vast majority of those moves will be somewhat nonsensical though.

As long as it's all in memory, that's fair game. Everybody at the high level is using engines for training/preparation.

This is balanced by the fact that the sheer number of possible lines is huge. Too big for any human to "remember" them all.

It's also not enough to memorize "good" moves from an engine perspective. "Good" engine moves and "good" human moves are different things.

An engine might see a move as "good", because it calculated 50 moves down the line and found an advantage. This means that you'd have to play the next 49 moves in a very precise manner to reach this advantage. One way to think about it, is that there are multiple positions that are analyzed as "draw" by en engine, but 9 grandmasters out of 10 will see one side winning. That's because one side would need to play perfectly, while the other side has much higher tolerance for mistakes (engine assume perfect play on both sides).

TLDR: engine is fair game, during preparation. It's a big deal, it transformed high level chess, but it didn't break the game, and is unlikely to do so.

No, using engines to find lines is a common way to practice (at least, that's my understanding. Obviously, I'm not personally a world class chess player.)

No, that is exactly what they do in opening preparation.

No. But the opponent will do that too. The trick is to find a move that is just slightly worse according to the engine, but that gives a complicated position that's hard to figure out behind the board. That you will then learn all the in and outs of, and memorize. And then hope you get it on the board...

They need hundreds such ideas for a match.

That is the strategy of opening preparation. For the most common openings, GMs often have a 20 move sequence memorized. The time limit of chess matters a lot, so the strategy is to break your opponent out of their preparation with a move that is unintuitive or considered untenable, after you have determined the ways in which it is actually better than widely considered.

The ideal at top level is to get to move 20 or so with a slightly better position, essentially all of your time (because you've been playing memorized moves), and your opponent already having used most of their time.

It isn’t, partly because you needed to predict the scenario would come up and partly because your opponent can do the exact same thing.

That would depend on the opponent doing the same, and picking the same engine and settings for their side too.

> Is there something they're memorizing

Yes. Their upcoming opponent's past games (including analyzing them for strengths and weaknesses), and engine lines mostly.

I remember Caruana mentioning how he wasn't prepared for how much Carlsen "understood" him (not exact words from the video, but something to that effect).

It sounded sort of like a side mind game that Carlsen plays on his opponents. It made it clear that Carlsen really studies his opponents and not just their past games.

Going through all of opening theory to invent enough new ideas that might give them a tiny edge, somewhere. And then eventually memorizing what they found.