How do top players use computers to improve their play?

Question

This question is inspired by yesterday's matches between Hikaru Nakamura + Rybka vs Stockfish. Nakamura was allowed to use an older version of Rybka to help him, and Stockfish was allowed to play with its full strength but not use an opening book or an endgame tablebase.

Question 1: How might players such as Nakamura use Rybka to improve their play against a computer opponent?

I can come up with a few naive guesses: check their own moves for blunders; check the computer's suggestions to see if there are some convoluted move sequences which lead to a huge advantage; and letting the computer play the endgame for them. This does not need to be the full story though. Maybe there are useful features that chess engines have that I am not aware of. I'm also wondering to what degree a human player would rely on his own intuition: would he always play his own moves assuming the computer never suggested a blunder or an amazing winning move which is difficult to see?

---

My second very similar question is inspired by a few comments by Nakamura in a reddit thread, namely

Fischer would almost certainly lose to [people such as myself, Carlsen, and Kasparov], but this is due to the fact that the game has so fundamentally changed. If Fischer had a few years to use computers, I think he would probably be on the same level.

and

I think mainly what can be learned from computers is a deeper understanding that almost all positions are ok with accurate play. In the past, many people assumed certain positions were automatically bad, but computers have shown that the rules and thought processes aren't always accurate!

Question 2: How have computers changed the way top players play today (when unassisted by computers)?

What is meant by the first quote? The second quote indicates that one might not want to give up too early, but does not suggest any change in playstyle.

Answer 1

How might players such as Nakamura use Rybka to improve their play against a computer opponent?

... check their own moves for blunders; check the computer's suggestions to see if there are some convoluted move sequences which lead to a huge advantage ...

I believe you said it yourself here. A computer would be used in order to check if the move that the GM thinks would be a good move is indeed the best move to exploit the existing position. I doubt that it would play the end game fully, I am positive that the GM thinks through every move of the game instead of trusting blindly the computer, but most likely definitely the opening would be played by the machine, at least till about move 5-7, until the position changes known book course.Additionally, the computer would most likely have some kind of a database of the previous games played between other players, so if the position played was already played by someone else a while back, it could be used to not make the same mistake or, vice-versa, to build a long term plan.

How have computers changed the way top players play today (when unassisted by computers)?

OK, I do understand that this might get a bit long, but bear with me. Ill bold the important points. Computers have made the modern top players much more tactically aware and ready than the chess players before them. This is what the quote about Fischer is talking about: Fischer did not have the opportunity to analyse his games and study lines to such a far extent as is done by the modern top players using computer technology. About the second quote: with the knowledge that there is a possible accurate move in virtually any position, top players would dedicate more time to thinking and finding that one single accurate move, than just assuming that the position was lost and thats it. This makes the game much more interesting and thrilling, because you know that the accurate move exists, its just the matter of finding it.

About how computers help top players prepare: so basically, theres 4 reasons why computers are used today for chess:

1. Playing against software program
2. Keeping track of games (Databases)
3. Using software to analyse your games
4. Using software to do chess problems

Out of those 4 points, the one that interest us are the bolded two.

Keeping track of games (Databases) is used usually to analyse the openings, and then comparing them to how they were played by other top-players in order to evaluate one's own position and/or learn something from others. You could've never thought that that move would bring any use, but if Kasparov's been whipping it out right and left for the past 5 years with great success, you might as well try giving it a go.

Using software to analyse your games is the primary use of computer engine by top-players. After every official game they play, the will analyse their games and see where they did wrong and where they could do better. The programs are usually proficient in tactics and opening theory, so for someone like Magnus Carlsen, whose opening is not as strong compared to other GM's like Kasparov and etc, a computer engine would be the best an fastest way to improve. And of course, if they find that what they did in game is not the strongest move in the position, next time they would be aware of it and follow the move that was suggested by the engine. Usually this would happen once again in the opening, because middle games repeat quite rarely if played against different opponents.

Answer 2

When someone like Nakamura plays Advanced chess, he might try out different lines letting the engine play the other side, and then see whether he likes the resulting positions or not. There are some sacrifices that an engine won't initially see the point of, but he can try these out, and generate a line quickly to see if he can get the engine to change its mind once he shows it the payoff. Also in the Stockfish games, Nakamura was trying to use anti-engine strategies, but he doesn't want to get caught in some cheapo tactic while he is maneuvering for position, so the engine can alert him to things like that.

As engines became stronger, the openings played in over-the-board games have become much more tactical and precise. Players prepare novelties in their opponent's pet lines, generate huge trees of analysis, and then try to memorize all the different lines. Fischer did a lot of analysis too, but it took players a long time to analyze, and they might miss a key refutation that a modern engine spots right away. Some openings today have been analyzed almost to the endgame with the players leaving book move 35 or later. In the old days, this kind of deep analysis was impossible.

Because some lines have been analyzed out, players have to find new ideas, and try to get the opponent out of book. This has become much more important especially when playing against an opponent known for his thorough preparation. Thanks to databases and the Internet, players are much more aware of what lines have already been played. Fischer was good at catching oldtimers in known traps, but it would be hard to catch recent superGMs this way.

The opening is the main area where things have changed, but there have been some changes in the middlegame and endgame as well. Certain sacrifices have become standardized, and even though all the nuances haven't been worked out, GMs are more aware of some of these recurring themes thanks to databases.

When he talks about all positions being okay, this has to do with the fact that engines are willing to take on positional weaknesses if there is some tactical payoff. Some players were willing to take on holes and weak pawns even in the old days, but now more players are willing to play like this since the engine seems to think it's alright.

Endgame tablebases have also enlarged the number of positions known to be theoretically won or drawn.

Top players spend a fair amount of time preparing openings and working on their weak areas, and they use computers in a wide range of ways, eg. to draw up a repertoire, to test their memory of their repertoire, to go through tactical puzzles, or positional exercises, to review and analyze games, to compile statistics on their own performance in various openings say, to research their opponent's openings and weak spots.

Answer 3

Although you asked about players such as Nakamura, you might want to consider how correspondence chess players use computers. They play the most "perfect" games, in terms of reliably picking the best move in each position, and they do it with the assistance of computers.

For an interview with the most recent CC World Champion in which he talks about his approach to using a computer, including databases and engines, see Better Than an Engine: Leonardo Ljubicic, ChessBase.com, 2016.02.21.