Can chess grandmasters beat computers? Human vs Machine
Chess grandmasters can generally beat lower-middle tiered chess computers just fine. High-tiered computers can only be beaten if there are time restrictions and some other handicap. At their strongest (fullest depth), supercomputers will not be beaten by any super grandmaster.
Grandmasters are the pinnacle of human strength when it comes to chess as they are someone who has taken the steps to be that good. With the rise of chess computers however their humanity is being exposed right in front of our eyes.
And that is what I am going to talk about in this article, whether chess grandmasters are capable of beating modern chess computers. This is very interesting if you are curious about the strength of computers compared to us humans.
The uttermost product of modern technology is right in front of our eyes and we are here to observe its effects on modern chess. Man vs. Machine is a common theme in the contemporary world and chess is one of the fields of this clash.
Without further ado, let’s get started.
Grandmasters can beat mid-tier/lower-tier computers
This is the thing with this question, all chess computers are not the same. All of them will have different modifications, designs, and strengths that make them different from each other.
This meant that depending on the engine, they would have varying strengths that we couldn’t just treat them all the same. If the engine is not something that is generally considered strong, grandmasters will usually have an easy time dealing with them.
Grandmasters can generally beat lower-tiered to mid-tier chess computers, the problem starts with high-end computers that are able to implement brute force calculation.
These so-called supercomputers literally out calculate their opponents by considering hundreds of thousands of variations in just a minute or so. This is why it is called “brute force” since everything is so calculated thoroughly.
But if the engine in question is not really in the names of stockfish, komodo, houdini, leela chess zero, etc. (famous chess engines) then a traditional grandmaster will have a good shot. It is the supercomputers that are giving all the trouble.
There are super grandmasters that beat supercomputers
You might be thinking that yeah, grandmasters can beat unpopular chess engines but there’s no way supercomputers would ever lose a right? wrong. There are popular examples of super grandmasters beating supercomputers in chess.
We do have popular examples of human players beating chess computers such as with Nakamura vs Rybka. There are also a lot of videos online of stockfish for example getting bagged.
In the Nakamura vs Rybka game, Nakamura implemented an anti-engine strategy that was popular at the time.
In a nutshell he beat Rybka by playing a closed position and waiting for a chance to sacrifice, giving him an opportunity to convert. Rybka is actually a supercomputer that can be considered modern (engines created in the 21st century).
Rybka terrorizes a lot of chess tournaments and has bitten a lot of famous players. So this is not some cheap engine that Nakamura has beaten, it is a real functional supercomputer.
If you are wondering how this is possible, it might be worth taking a look at how chess engines “think” about a position. This will shed light on the inner workings of a chess engine.
Now, chess engines have been much improved since the days of Rybka. But it only takes you some youtube searches to see grandmasters beating stockfish, unrepeatable modern chess computer
Grandmasters defeating Stockfish on Youtube
I just did a simple video search on youtube literally typing “grandmaster beats a supercomputer” and I already stumbled upon some interesting findings. I have seen a grandmaster beating stockfish 12.
This is literally a titled player beating stockfish 12! a version that is considered to be the god of chess by some people. And here is this guy beating it convincingly.
This is not a top 10 player either, which puts into perspective how strong these individuals can be.
Don’t get me wrong this does not happen every day, even grandmasters will struggle for the most part. The reason that this guy won is because of the time, stockfish 12 cannot process as deeply with the time given (will discuss later).
But it is indeed proof that grandmasters, which is the strongest form of a human competitive player, can be supercomputers here and there.
Grandmasters can at least win a game against supercomputers
Though chess engines will generally beat human players in a tournament setting, a super grandmaster can win a game against an engine every now and then.
If we are only concerned about the capability of a human beating a computer in a single game then sure, grandmasters will be capable of that. The problem arises when the format is in a tournament style.
Humans have a hard time beating computers in a consecutive match format throughout the 21st century, chess computers have just become more advanced that they can win more than a regular grandmaster (statistically).
But there will be games (depending on the engine) where even a regular grandmaster could naturally convert, I think we are underestimating the strength of a grandmaster too much.
There are grandmasters that could play a game with minimal inconsistencies and super grandmasters that could play perfectly (engine-like).
It would make sense that a super grandmaster can beat even the top engines in just one game (even if they lose a lot of other games).
Grandmasters beat chess computers in lower time formats
This is something that I think not a lot of people are aware of, that chess computers are not really good at dealing with time. Primarily because they can’t calculate at a greater depth if only forced to calculate in a minute or so.
The time control is also a factor for a grandmaster to beat a supercomputer. Even strong chess engines generally do not do well in bullet (time format) for example.
I would even dare say that a human grandmaster will be able to beat a supercomputer in a tournament setting if the format is a bullet game. Chess engines are just not built to respond to fast calculations.
Chess computers can generally think much deeper with a brute force calculation, but they do need time to think of the said moves. Humans on the other hand can respond with patterns, humans can play without calculating everything thoroughly.
So if the time control is set to be in a blitz or bullet format, then a grandmaster will have a good chance of defeating a strong chess computer.
Grandmasters can win against chess computers with odds
It is easy to think that grandmasters can only win if they have the advantage of time, another handicap that we can give to chess computers are with odds.
Giving odds can definitely help grandmasters overcome the brute force calculation. Grandmasters can beat chess computers if they are up a pawn at the start of the game for example.
You might say that this is a cop-out since he didn’t really beat the computer if you are given the advantage at the beginning of the match. But it does prove that there isn’t an insurmountable gap between humans and computers yet.
That with a single pawn advantage for example, a regular grandmaster can beat the so-called “unbeatable chess engines”.
Grandmasters can beat chess engines if they have the better opening
Some of you might not know this but there is a meta when it comes to opening choices, some lines early in the game do not work as well as the others. This means that it can be used as a handicap to close the gap between humans and engines.
The opening choices also matter in the state of the match. If a chess engine is forced to play a suboptimal opening then a grandmaster might be able to beat it.
French defense, Queen’s Indian defense, Scandinavian defense, or even a King’s gambit are some of the openings that are not considered good for their particular color in modern times.
These openings are not so bad that they single-handedly lose the game, just that they are less optimal than other opening choices out there. If a grandmaster is given the opening advantage then they will have a good chance of beating a strong computer.
There are matches where strong chess computers are forced to play a particular line in the opening, though it hasn’t really been implemented yet. In theory though, this will give chances for grandmasters.
There are moves that the engine could not analyze well
The engine is not absolute, there are certain moves that are proven best in the position but cannot be seen by the engine because of some technicality. Usually it has something to do with the depth.
There are certain moves that are invisible to an engine depending on the depth, a chess grandmaster can beat strong chess computers if it’s not in its fullest depth (not in its full potential).
This is the origin of the term “anti-engine move”, which are moves that win the position but cannot be identified even by strong chess engines. This is usually because the depth of the engine is not set in its fullest mode.
If a grandmaster faced a strong chess engine that is not in its fullest depth, then there will be chances for victory.
Generally, chess computers are better than humans
If everything else is fair and both sides are given a lot of time (while a strong chess engine will have its fullest depth), human players might not even get a single game.
Now I did say that human players can win here and there but that is only if the opponent is not a supercomputer in its max potential. If there is not a timer handicap or anything, then supercomputers might just totally crush grandmasters.
Because if all things are fair then the gap between humans and computers is just too much (not insurmountable though at the moment). Out of 1000 games, a grandmaster’s not even winning one game is not too far from reality.
Even with all the time spent trying to be a chess grandmaster, computers are just too strong.
But that is the thing with this question, there really wasn’t a study that has been conducted to really answer this issue.
Games between top grandmasters and top chess engines are limited since professionals are not winning to brawl with computers most of the time.
There is a lack of lab data in this concern.
The gap between humans and computers are hard to evaluate
One of the best ways to really answer this question is to just have the best player in the world right now compete in some 100 to 1,000 games against the strongest chess engine until he wins one game.
But that is the thing, the strongest grandmasters in the world will have pride in their craft and are usually not willing to taint it by losing convincingly to a chess computer.
Plus, a 100 to 1000 game setting is a very tiring match that would probably last a ridiculous amount of logistics and time. The actual preparations for the event will be challenging if somebody is planning to do it.
It is very unlikely in the future but without it, we don’t really know how the strongest human fare against the strongest chess engine.
But so far, it is hard to evaluate the gap between the strongest human and the strongest computer, since there is no way to really determine if a human player can get one game out of 1,000.
My own personal opinion
In the case of supercomputers in their furthest depth, I doubt that even the top chess player can even beat it once. If we are talking about the latest version of stockfish with maximum depth and a lot of time, the possibility of losing is slim.
It will probably be a mountain of draws at best, l mean winning and losing are not the only possible results in chess, most likely the games would be a bunch of draws.
Unless human players improve dramatically over the next few years while the improvement in chess computer technology declines, then I think the gap between humans and computers will just keep getting bigger.
I personally believe that with full potential, chess engines are superior to humans and the strongest grandmaster won’t probably win even a game out of 1000 games (this is just an opinion since there is no experimentation yet).
Final thoughts
A lot of people are becoming nervous about the advancement of computers in chess, but I think they are missing the point. Every field will be visited by technology in order to improve the industry, we just have to adapt.
It’s not like human games will be dead even if engines are still around (engines are around today, and people are still playing chess!). I think their existence is a really interesting point in the works.
If you look at elite games there is really a need for precision, perfection is becoming the goal for professionals. It might not be visually appealing, but precision can become beautiful at times even if the game ends in a draw.
Plus we can use computers to improve ourselves! They are the best coach out there since they play so perfectly (and they are free). I am happy to live in this era, sleep well and play chess.
