The sharpie and the grandmaster are playing a game. Who and why is forcing Russia to play by the rules that change during the game?
On this page you can play chess with one of the most complex computers in the world - Shredder.
Free online chess with a computer: why is it so popular?
In any art it is impossible to rise without strong opponent. Chess is no exception. Previously, opponents were looked for in chess clubs, through recommendations, even through advertisements on poles. And the arrival of a grandmaster for a simultaneous game was a rare event that all chess players in the region or even the city were looking forward to. Now everything has been simplified thanks to the ability to play chess online with a computer.
A computer program can be no less powerful and tricky an opponent than an experienced chess player. That is why competitions between the world's strongest grandmasters and computers are held regularly. Online chess has long become a popular intellectual entertainment.
How a chess computer works
A computer chess program is built on a tree principle: after a player’s move, it first evaluates all positions that will arise after it, then analyzes possible positions as a result of the subsequent move, etc. The analysis lasts until the program determines the final outcome of the fight - checkmate or stalemate. Next, from the entire tree, the program cuts off “weak branches” - positions whose successful assessment is minimal and selects the strongest position, on the basis of which it makes its move. The ability to see ten steps ahead is a quality that every strong chess player possesses, so we can say that in this regard online chess no different from humans.
the site is your online grandmaster!
On our portal you can play chess online without registration with a Schroeder computer - one of the most popular chess programs used to teach the skill of the game. You can choose any of three difficulty levels, ranging from easy to difficult. Thus, our online chess will be useful to both beginners who are just learning fascinating world chess and experienced players who want to “pump up” their skills to improve the speed of the game.
To play chess with a computer, you don’t even have to register on the site. The game mechanism is simplified to the minimum: just press the “New Game” button to select a side and start the game! Our community was created specifically for beginner chess players to develop their skills by playing chess online without registration for free. We have a VKontakte group where participants look for partners for the party and exchange experiences. In addition to playing with a computer online, on our website you can familiarize yourself with
Chess passions at the Assembly: grandmasters against cheaters (photo report)
It seems that from time immemorial, chess lovers have been gathering on Cathedral Square, laying out boards under the Odessa chestnut trees and engaging in intellectual battles from morning to night. These people are not stopped by either political ups and downs or crises - they go to the Assembly as if they were going to work. They improve their skills, earn, let’s be honest, a pretty penny, and simply spend their retirement leisure time well. TIMER visited the main amateur chess arena of the city to find out what brings people there throughout
It seems that from time immemorial, chess lovers have been gathering on Cathedral Square, laying out boards under the Odessa chestnut trees and engaging in intellectual battles from morning to night. These people are not stopped by either political ups and downs or crises - they go to the Assembly as if they were going to work. They improve their skills, earn, let’s be honest, a pretty penny, and simply spend their retirement leisure time well. TIMER visited the main amateur chess arena of the city to find out what brings people there for many years.
When there weren’t even tables here, chess players played right on wooden benches. We discussed the latest city news and gossip. Football fans also gathered here, learned the latest sports news, shared their intense impressions of past matches, and played amateur betting. However, most visitors to the informal “Club on Soborka” prefer sedentary sports - chess and checkers. The tradition is still alive today.
“I’ve been coming here since 1977,” a visitor to Soborka with thirty years of experience, the director, told us. chess school Vladimir Andronov. – Then I lived on Bolshevik, and there was no one to play chess with.
Now amateurs come to the Soborka intellectual game from all over the city. Both professionals and amateurs gather. It cannot be done without material interest. The game of “bet” is a feature of this chess arena. This peculiarity sometimes attracts even chess champions and masters to the center of Odessa. Chess players say, of course, you can also play for fun, but it’s better to combine interest and money. True, not everyone agrees with this.
“The game of chess is so attractive that all financial issues disappear here,” admits Vladimir Pavlovich. - True, I myself also played for a bet, I remember everyone chipped in 3 rubles, and I offered 5, so everyone who wanted to was immediately eliminated.
Another old-timer at the open-air club, Alexey Ivanovich, is so popular and respected that partners are willing to pay good money for the right to play with him.
“I’m a pensioner, I have nothing else to do, so I’ve been coming here for more than 20 years,” says Alexey Ivanovich. – I used to be an amateur, but thanks to Soborka I became a real professional. I’ll tell you a secret, sometimes they give me 100 hryvnia for one game.
Once upon a time, real chess cheats were spotted at the Soborka. The fact is that it is very difficult to hit the same jackpot at a black and white board as at a green table. The stakes here are low and there is no particular excitement - only tactics and strategy. And with a chess player who is obviously superior to the rest, no one will play for money. So failed grandmasters have to pretend to be “dummies” for a long time - to give in and imitate an average game, until someone loudly declares that they are ready to bet a large sum. It is then that the self-confident player is cut to pieces.
The assembly remembers how one day two such “chess pieces” “found” each other. There were 200 rubles at stake - a crazy amount for the early 80s. Partners who had not shone before chess genius, suddenly they started playing like at the World Championships. They played all day until late at night. They even began to record their moves. Then, having drawn the twentieth game, they stood up and left together, even leaving their pieces behind. No one else saw them in Odessa.
And the Soborka is also frequented by young, gifted chess players, who often give experienced elders a head start.
“There was such a boy, Kirill Reshetkov, he sat down and knocked out all the old-timers,” recalls Vladimir Andronov. – Now he is already an international master of sports, and he won his first victories here.
Soborka is predominantly a men's club. Women are very rare here. Basically, the fair sex condescends to the arena to pick up their overstaying chess-playing husbands.
Yes, there is such a form of gambling addiction: “dependence on Soborki.” Vladmiir Andronov remembered one such chess fanatic:
– There was one player who didn’t care whether he lost or won, the main thing for him was to play, play and play. Once he sat for seven days and nights in a row. A whole week - can you imagine? Moreover, neither snow nor rain stopped him. This is the kind of enthusiast we had.
Others, gambling addicts, are relieving their withdrawal symptoms at Soborka. After being separated from the one-armed bandits, gamblers also flocked here. Now they share the territory with chess players, drafts players and those who like to “beat the goat.”
I managed to scare away one gambler with difficulty. The man, seeing a journalist with a press photographer, simply begged on his knees to evaluate his cheating skills. In the end he showed card trick, which I forgot back in first grade, thus “won” five hryvnia from me and went to celebrate the victory in the direction of Moldavanka.
The previous presidential elections helped refine Soborki's playing space. There are tables, benches, and most importantly, a canopy that saves players from rain, wind, snow and falling chestnuts. However, there are still not enough places. Only 12 chess players can play comfortably at the same time. The rest are waiting their turn - the game is on"to take off." True, you don’t have to shift from foot to foot for long: a prolonged game is very rare.
The article does not specifically indicate all the disadvantages of chess programs, but only some. This is done for my personal reasons. In addition, appropriate examples and conclusions have not been selected. Also, the article is somewhat complicated so that it is not used by “attackers”. However, an intelligent person will be able to see useful advice in it.
Everyone remembers two significant Human-Computer matches, they received a lot of publicity and had a solid prize fund. First world champion Vladimir Kramnik played an eight-game match with the chess program Fritz world champion computer programs. Then it was the turn of the leader of the world ranking list Kasparov will compete in a six-game match with the Junior chess program. Also world champion, but from a different year. These versions of the programs did not play official matches with each other, so we were able to find out which of them was stronger only at the next World Computer Championship. However, the styles of these programs are known to differ. If Fritz tends to play positionally, then Junior can be called an attacker. Someone even compared the style of the Fritz program with the style of Kramnik himself, and the Junior style with the style of Kasparov. It was said that grandmasters play with their electronic embodiment.
Let's now take a closer look at how our chess champions fought against artificial intelligence. We will start with the famous match Kasparov Deep Blue chess supercomputer, specially created by the IBM corporation.
So this happened in 1997. At that time, personal computers were not powerful enough to compete with grandmasters, and IBM specifically created a giant multiprocessor chess computer. The first version of the machine was not particularly perfect, and lost to Kasparov in all respects. But the second version managed to beat current champion world chess. This caused a lot of noise and attracted the attention of all media. What happened? Let's look at it from a scientific point of view.
There were six games in total. In the first game, Kasparov played with white. So that the chess program could not use the opening library stored in the computer’s memory, Kasparov chose “crooked” openings in almost all games, which are not in the database. Openings with strange wait-and-see first moves don't play like that against people, because it's just giving away the initiative. However, Kasparov was able to gain an advantage in this game, using the positional errors of the computer program, and win.
In the second game, Kasparov played black. Unexpectedly, the computer played very strongly in this game. He managed to make several very strong planned positional moves in a row, which put Black's position at risk. Next, we will also look at the principles of operation of chess programs, but now we note that planned games are not good for programs; they operate not with plans, but with individual moves. To save the position, Kasparov sacrifices two pawns, in return receiving chances to attack the king, perpetual check due to the activation of the pieces. And then a miracle happened, the computer thinks for fifteen minutes, this has never happened, it always thought for exactly three minutes on each move. And rejects the victim! All computer chess programs big materialists. They don't see positional threats, but they know how to count pawns. And one of the good methods of fighting against programs is to sacrifice something for positional compensation, which the program is not able to evaluate. Upset and surprised, Kasparov makes a few more moves, the computer strengthens its position, creates threats, but unexpectedly still provides Kasparov with a chance to check the king and try to save the game with a perpetual check. But Kasparov gives up. However, he is immediately surprised to learn from the commentators that he resigned in a drawn position, he had a perpetual check! Kasparov explained his decision by saying that he did not believe that the computer deliberately went for the perpetual check option, if several moves earlier it had evaded it without being tempted by two (!) pawns.
For the next three games, Kasparov put a lot of pressure on the chess program, but it miraculously escaped to a draw all the time. Some tactical nuances saved the computer.
In the last, sixth game, Kasparov played with black; he went for the opening variation, in which white sacrifices a knight for a strong attack. Chess players never go for this option. The computer sacrificed a knight and won easily, not right away, though. After the game, Kasparov explained that computers never donate material for an initiative that they are not able to properly evaluate. However, many commentators said that this variation could already be in the opening library that the chess program used. Kasparov, for completely unknown reasons, ruled out this possibility. It was suggested that Kasparov simply mixed up the order of moves, that is, he made the second move of the variation first instead of the first. This happens to chess players, especially since Kasparov was already tired and exhausted from previous games.
So, the match ended with a score of 3.52.5 in favor of the chess computer. IBM fully promoted this success of its brainchild. I even heard in some American film in the intro that such and such a year has come, the Computer beat Man in chess, almost the end of humanity. After this match, IBM shares rose significantly in price.
Based on the strange behavior of the computer in the second game of the match, Kasparov's team suspected the IBM team of foul play. The fact is that they invited the grandmaster to suggest strategic points to the chess program. That is, the chess player and assistant told the program not to accept the sacrifice of two pawns due to the threat of an attack, which the program could not fully calculate and correctly evaluate, and when a decisive advantage was achieved, the adviser lost his vigilance, giving the opportunity to declare a perpetual check.
An indirect confirmation of such a hypothesis can be the commercial interest of the company as a result of the match. The chess computer lost the first match cleanly, and a lot of money was invested in the development, so there shouldn’t be a misfire a second time. And also the fact that after the match the computer was dismantled.
However, with the same success one can accuse Kasparov himself of playing a fixed match, having invented all the games himself and learned them by heart, adding a dramatic element. Why did he give up so early without trying to give several checks, why did he take a very risky option in the last decisive game? There are more questions than answers.
There were strong rumors that during games the huge multiprocessor computer sometimes froze and had to be rebooted. The technical staff revived him like a beaten boxer between rounds. I can hypothesize that in the second game the computer first relatively accidentally made several good moves in a row, even a weak chess player is capable of this, and then froze. And such a machine takes a long time to reboot. And when I rebooted, I had not yet entered the game, I played according to the principle of reliability. Maybe he has such an option, to play immediately after the reboot, when, perhaps, not all the blocks have turned on yet, from protecting the king. And then he began to play normally, but he overlooked the threat of the perpetual check, since it was multi-move. In general, the program froze, which surprised and amazed Kasparov so much that he could no longer come to his senses.
Gradually, the power of processors for personal computers and small multiprocessor systems grew, which also allowed improving chess programs to reach the grandmaster level of play without the involvement of supercomputers.
Let's start with the Kramnik-Fritz match that took place a little earlier. The match consisted of eight games, and it so happened that it was divided into two halves. In the first four games, Kramnik clearly adhered to a competent anti-computer strategy. He exchanged queens and moved into a complex multi-piece endgame. There wasn’t much tactics in it, and the large number of pieces did not allow the computer to deeply consider the options in order to see the long-term consequences of its strategic miscalculations. That is, the computer could not choose the moves correctly, since their results would appear beyond the limits of its counting capabilities. Let's say he counts for 6 moves, and the consequences occur on move 10, and the program cannot distinguish a bad move from a good one. The absence of queens seems to slow down the passage of chess time; with queens there are a few moves, and that’s it, checkmate. And in the endgame, dozens of obvious but planned moves are often needed for a positional advantage to develop into a material one. As a result, Kramnik won two games; in one game, the computer, thanks to Kramnik’s inaccuracies in realizing the advantage, managed to make a draw. And in the very first game, the computer had the best endgame, but it was completely useless, since it didn’t know how to play it. But in the second half of the match, Kramnik became bored with winning like this, and to begin with, he blundered his knight in one move and gave up. In the next game, he decided to attack, simply sacrificed a bishop for several pawns, and lured the enemy king out of his hiding place. But he failed to complete the mating attack. And the computer is very strong in counting-type positions, and Kramnik may have made an inaccuracy somewhere. However, the initiative was sufficient to obtain an endgame with draw possibilities. It was possible to build a so-called fortress; the queen cannot overcome the defense of the rook and the pawn supporting it. The computer does not see such things at all and allows it. But Kramnik, as is usual with participants in such matches, hastened to give up in an almost draw position. After that, his rage disappeared, and in the last two games the opponents swarmed each on their own half of the board, not risking crossing the border.
This match was contractual in nature, in the sense that the parties simply agreed among themselves that they would play such a match. If there had been a fair selection for the match with the world champion, then, of course, this version of Fritz would not have made it anywhere. She has the obvious weakness game in a multi-piece endgame, but won only due to unforced mistakes of the opponent. There, on the outskirts of the chess Olympus, there are such “deprived” grandmasters that they wouldn’t “make peace” with her. They would win once, and then calmly wait for a mistake, which is inevitable when trying to play for a win.
In fact, an earlier version of this program played together with grandmasters in a super tournament in Dortmund. At first, the chess players didn’t know how to fight it, they just tried to save themselves. Then one managed to win, and then the program was butchered, as its weak point was revealed. She played closed positions poorly.
Let us now consider Kasparov’s match with the Junior chess program. Although this program is not related to the brainchild of IBM, the match was perceived as revenge for the defeat by the supercomputer.
Kasparov won the first game, this time he did not play the wrong openings, he stuck to classical schemes. Even though the chess program was able to use the opening library, it was unable to solve the opening problems. After the opening, she was already doing badly, and Kasparov calmly beat her.
In the second game the computer played White, Kasparov again did not hide and chose the sharp Sicilian Defense. However, he managed to outplay the program positionally. At one point, the greedy short-sighted program was tempted to sacrifice an exchange for a strong attack, and ended up with a losing position. However, Kasparov made a hasty move, allowing the program to declare, in essence, a perpetual check.
In the third game, Kasparov played white and attacked. But the position was of an open counting nature, and the computer confidently defended itself. However, he had nothing but the opportunity to make a draw. However, Kasparov violated the rules of the game with the computer, carried away by the attack, he left his own king unsecured. And, instead of securing a draw by repeating moves three times, Kasparov blundered the threat of checkmate and was forced to surrender. Human attention tends to focus on one thing, in this case an attack, which allows for plans to be made. The computer plays without a plan, but all moves are equal. Therefore, he tends to combine counter-threats with defensive ideas. I played chess with the computer many times and noticed that when defending, it likes to unexpectedly jump up with a queen on one flag, creating a threat to the king, and then suddenly transfer the queen to help its attacked king. This defensive maneuver is easy to overlook.
In the fourth game, Kasparov played for a draw with Black and received rook endgame without a pawn. The stupid computer happily went after him, not realizing that, despite the extra pawn, he was completely worthless. And so it turned out to be a draw.
In the penultimate fifth game, Kasparov prepared to attack, but, quite unexpectedly, the computer sacrificed (!) a bishop to him on the h2 square for the attack. Kasparov thought for a long time and did not go to a position with an extra bishop, but richly saturated with tactical threats. He later explained that such a position is very easy to lose to a perfectly calculating computer. Moreover, having little time. It turned out to be a draw by repeating moves.
This game is being touted as a great success for artificial intelligence. The computer made a sacrifice of material for initiative, this was previously only the prerogative of the human mind. However, one should not rush to conclusions too quickly. This sacrifice was made in the opening, perhaps it was already included in the program in preparation for the match. After all, strong grandmasters worked with the program, they studied debut repertoire Kasparova.
In the last game, Kasparov again outplayed the program and got a difficult endgame with a big advantage. Some commentators argued that it was technically won. However, Kasparov chose to record a draw. He explained that it was more important for him not to lose the match than to win.
The match this time ended in a draw, but the interesting question is whether Junior played stronger than Deep Blue? On the one hand, Deep Blue had great computing power at its disposal, but, nevertheless, over 7 years, the development of computers did not stand still; the power of personal computers increased significantly. However, it can be noted that there has been a significant development in chess programs that has compensated for the power gap.
However, let's take a look at these twenty games. In them, the computer won 5 times, lost 4 times and drew 11 games. Not a single intelligible victory. It’s all just blunders and giveaways. Not to mention surrendering in drawn positions. And the defeats, on the contrary, are absolutely devastating, revealing large positional errors. It is completely incomprehensible how a computer is going to win without human help.
In general, you know what the game style of computer programs reminds me of? There used to be, and still are, players who play blitz in coffee shops and parks for money. They always have a rich set of traps and memorized opening variations in their arsenal; they love to confuse their opponents with provocative moves and catch them in small tactics, but they have no real understanding of chess.
How did it happen that artificial intelligence plays like a chess sharper? How do chess programs that beat world champions work, and on what principles are they built? The basis is to search through all possible options and select the best from them. Let's say there is a position and a computer move. He goes through all possible moves in a given position, each move leads to a new position in which he also goes through all possible moves, and so on to some depth.
At the end, he evaluates the resulting positions and chooses the best move. In this case, he will choose the first option, since in it he will be able to obtain a position with the greatest advantage for any response move of the opponent.
How does the program evaluate a position and assign a number to it? It relies on factors that can be calculated analytically. First of all, this is the ratio of material, expressed in conditional pawns. This, of course, is the most important factor, so all other factors are also usually measured in conditional pawns. For example, in this position the activity of the pieces is one and a half pawns. Relatively speaking, you can transform the activity of the pieces into winning a pawn, and a small advantage will remain.
What else can be calculated algorithmically? It is enough to simply calculate the same activity of the pieces, that is, the number of possible moves; it is easy to evaluate the control of the pieces and pawns of the center. You can calculate the advantage in space, the number of controlled fields. Take into account the simplest factors related to the location of pawns. Somehow the presence of doubled pawns, weak isolated pawns, passed pawns, backward pawns. You can introduce a simple factor that reflects the security of the king, for example, the degree of control of the fields around him.
But it should be understood that these factors are calculated purely formally, mechanically. It simply states the presence of doubled pawns, which is usually bad, or an isolated pawn. But the weakness in this case or the strength of this isolated pawn cannot be understood. To do this, you need to sort through the options, but they have already been calculated to the maximum depth.
Whether this figurative activity is imaginary or real is also unknown. It simply formally states the presence of a large number of possible moves.
This operating principle makes programs extremely strong in scoring positions, where it is possible to achieve a material advantage through tactical maneuvers. The programs simply go through all the options and go for the one in which they have more extra pawns. This is how they can be caught, which was successfully demonstrated in the matches described above.
Let us now evaluate the depth of calculation of options that can be achieved by implementing this simple exhaustive search method on modern personal computers.
Let's say we have a gigahertz processor, and 100 seconds are allotted per turn. That is, the program has 100 billion clock cycles at its disposal. Let it take 1000 clock cycles to evaluate the position. This is an approximate, but of course underestimated value. Taking into account various optimizations of the evaluation calculation. Related, for example, to the fact that positions obtained from one position by different moves have a similar assessment with starting position. They should not be completely overestimated. It is clear that the main contribution during work comes from the assessment of positions obtained at the very last level.
Now we need to estimate the number of possible moves in typical chess positions. In the opening, middlegame, and multi-piece endgame, this number fluctuates between 30 and 60 moves. For simplicity, let's take it equal to the square root of 1000.
We divide 100 billion clock cycles by 1000 clock cycles and get the number of positions at the deepest level of calculation. After this, we take the logarithm to the base equal to the number of possible moves. It turns out that the depth of calculation is equal to 56 half moves. That is, almost three full moves. It's not thick at all. Exponential nature of search chess moves did its job and destroyed productivity.
But maybe, if given more time and a more powerful processor, it will be possible to go deeper? Let's give a table, let the time be not 100 seconds, but 1000.
Estimated potential:
It is clear that chess does not lend itself to overkill. All hope for increasing the skill of chess programs lies in improving the algorithms for their operation.
Indeed, even candidate masters count for 5 moves. What did they do with this search algorithm to make it possible to fight with world champions?
Two main things were done with the fluffy tree of variants: in some places they shortened it, and in others they lengthened it. Chess programs have a function that “cuts off” the calculation of a variation when a critical advantage is reached. So as not to count the numerous mutual captures of protected pawns by queens. However, not all so simple. This take may be the beginning of a mating combination, after which material ratio and other positional factors are irrelevant. Therefore, the option cannot be cut off immediately, but only after a certain number of moves.
The second fundamental idea is that, since we can’t count everything, let’s count the most important options, but deeply. The trouble is that identifying these most important options is extremely difficult. Therefore, first of all, the computer counts all captures and checks in all positions. So that it doesn’t turn out that the calculation is stopped at the moment when you can eat the queen or give mate in two moves.
When you play with chess programs, they write in a small window the options they are thinking about and the depth of moves. The typical value of this depth is 12 14 half moves. But in fact, they did not look at all the options to this depth, and some went much further, dozens of moves. Thus, the computer can checkmate in a hundred moves if it is forced, that is, on every move check. And you can’t checkmate in ten moves if there are continuous “quiet” moves without captures or checks.
It is the manner in which the option tree is optimized that most determines the playing style of a computer program. It is not the position evaluation function that has a greater influence, but the balance between the calculations of options in depth and breadth. If the program quickly discards suspicious options, then it may miss hidden tactical opportunities, but spend more time assessing positions.
Let us now consider in practice the operation of chess programs. Next I will give an example of the unexpected tactical (!) blindness of the Fritz program, but first we will look at the moments from the Kasparov-Deep Blue match.
Game 2 of the match Kasparov Deep Blue. Before the move 46. Ra6? In this position the move Ra6 was made. Instead of just going into this absolutely won endgame.
This endgame is hopeless for black. They have great amount weaknesses and passive figures. The computer gave Kasparov the opportunity to declare a perpetual check by 45. ... Qe3 46. Q:d6 Re8! 47. h4! h5! and there is no protection from the eternal check.
How is it that a chess program makes such stupid mistakes? Even a less skilled chess player will understand that it is better to take a technically winning position than to expose his king to the threat of perpetual check.
However, most computer programs, even after much thought, make this move, Ra6. There are two main points here. They cannot understand at all that the above endgame is won. If we approach his assessment formally, then White has only the more active position of the rook on his side. On the other hand, Black has a protected passed pawn, while White has a passed but isolated one. From a formal point of view, bishops have approximately the same number of available squares. The computer program will most likely win this endgame by gradually strengthening the position and position of the pieces, but it does not see a winning plan. Material acquisitions, the transformation of pawns into queens, will occur only after dozens of moves, which lies beyond the limits of calculation. When realizing an advantage, some activation of black pieces and the creation of black passed pawns are inevitable, all of which the program is unable to evaluate.
And in the variant with perpetual check, the program has a large material advantage, a whole bishop, and many more advantages in the position. As for the perpetual check itself, the fact is that the program can register a perpetual check only if in all variations it is either worse or has a threefold repetition of the position. And here you have to wait for dozens of moves to repeat the position three times. First, the white king can go to the queenside, then he will return, begin to cover with the bishop, and then go on a journey again. There are a lot of options, but it’s all pointless, but the program won’t understand this until it calculates all the options.
So don't be surprised by the choice of computers. By the way, all the chess programs with which I had the opportunity to explore this position froze and glitched unusually often.
Kasparov found it strange that in some positions previously encountered in the game, the program reflected the threat of the eternal check. However, you can run a modern chess program even on a personal computer and see that the difference in evaluation of the move made by Deep Blue and the move leading to perpetual check is in the region of one tenth of a pawn. This difference is within the range of fluctuations generated by different program settings. The program can evaluate positional and material factors with slightly different weights.
On the other hand, in previous positions the threat of a perpetual check was more clear, a threefold repetition of the position was obtained much earlier, and modern programs it is discovered relatively quickly.
Knowing the tendency of programs to quickly “cut off” unpromising options with large material losses, we can assume the existence of tactical positions that are completely incorrectly assessed by chess programs.
Indeed, here is an example of such a position (from a game of grandmasters).
Black's move. I do not pretend to have a complete analysis of this difficult position, but, despite the extra pawn, White needs to save himself in it using sketch methods. Black activates the rook, Re4! and it cannot be taken due to the capture of the bishop on c4. On the move Bc3 there is a knight sacrifice on g5 at the right moment with the collapse of White’s position, and on the block Nc6! there is a quiet move Qe8! There are still many tactical ideas in the position. It would seem that chess programs built on the calculation of variations should quickly understand such positions. Not so!
Fritz program former champion among computer programs, it takes about an hour on a one and a half gigahertz processor to just begin to understand that not everything is fine in White’s position. And hours to get closer to the correct estimate. For a long time she does not see “quiet” moves, etc.
What about other programs? All of them are not far from Fritz in terms of playing strength. Even if they understand better, it is at the cost of weakening strategic thinking. By the way, the old ChessMaster 5000, 1996, begins to understand this position faster than Fritz. But it has many other gaps.
Maybe an increase in processor power will have a beneficial effect on the analysis of such positions? Maybe, but the point is that the program, when analyzing the previous position from which this one is obtained, must consider dozens of moves. This will again destroy all the speed gain.
Many chess players, especially amateurs, like to “compile” games for analysis into chess programs. However, one should be wary of computer-based assessments. The position can be won, and the program can be written even after much thought, with an advantage of only half a pawn. And, conversely, the position may be a draw, and the computer will estimate the pawn advantage of one of the parties. In this case, one should not be deluded by the high rating of chess programs; there is a big difference between the practical strength of their play and understanding of the position.
After playing with chess programs, I came to the conclusion that in order to achieve a positive result, it is necessary to try to make the absolutely strongest move in each position, maybe just calmly, and not whatever move you want. Often you try to make not the best move, but the move that you like, with a computer this can lead to bad consequences.
Having such a primitive and slow operating algorithm, how do chess programs manage to defeat grandmasters? How do they trick the human mind? How do chess cheats win?
The chess sharper has an opening cheat sheet with trap variations in his left pocket and looks into it
Indeed, a chess program has at its disposal a gigantic opening library recorded on its hard disk, but a person is deprived of such an opportunity. However, human memory can be considered generally, both internally and on external media such as paper. The eyes in this case are something like CD-ROM heads, and the sheet of paper is like a laser disk. A person is artificially deprived of not just half of his memory, but a special part of it that has separate functions for permanent storage of information. This is good for sporting competitions between people, but unacceptable for competitions between minds.
In the other pocket of the chess sharper is a collection of endgames. Indeed, the developers of chess programs, desperate to strengthen the play of their creations, began to incorporate databases of known endgames into their programs. Without them, the program will float and completely incorrectly evaluate even the simplest pawn endgame with the formation of a distant passed pawn. That's why chess programs have grown Lately, began to take up entire disks. It turns out that a person is not playing with artificial intelligence, but with his own knowledge base. This is good for training chess players, but there is no need to talk about the success of artificial intelligence here.
Cheaters rarely work alone
The matches described above were conducted between humans and multiprocessor machines. The processors told each other moves and exchanged ideas. There's little honesty here. So it was possible to load all computers connected to the Internet with one chess program and force one chess player to play from it. Since one person is playing, he should play with his own desktop computer.
Sharpers trick clients
Swindlers like to first let a person win, get excited, and then beat him up. Likewise, chess programs change their parameters from game to game, and they don’t do it themselves, it’s done by a human team. They almost replace the opponent. How would they like it if they were preparing to play with Kasparov, but got Kramnik as their opponent?
A chess sharper holds a spare board under the table and looks through the options on it.
Chess programs have a huge amount of stable memory at their disposal; you could say they move pieces on millions of boards. And the person, as I wrote above, is left without memory. It can neither be read nor written. If I were Kasparov or Kramnik, I would come to a match with a computer with my chessboard and would look at all sorts of options. The computer can't see anything anyway. Who said that you can't touch the pieces when playing with a program? There are no such rules.
It's like playing blind chess with a sharpie without looking at the board. You will imagine something with eyes closed, and the rogue moves the pieces on a real board.
On the other hand, it would be more interesting for the audience to watch what options the chess player is thinking about than just watching his thoughtful face.
Schuler wins on blunders
Indeed, all games won by computers in the matches described above were won by blundering everything possible. Pieces, checkmate, draw, opening. Without blunders, the sharper does not win.
In conclusion, I would like to say that whether this is good or bad, but for now, artificial intelligence very far from the human mind, which we see in the example of chess programs.
The sharpie and the grandmaster are playing a game. The rules are as follows: in the box there is chess (all the pieces). Players take turns, without looking, take pieces out of the box, two at a time, and place them in front of them. If both pieces are white, the grandmaster gets one point. If the pieces are black, the sharper gets a point, if they are different, no one gets a point. This continues until the box is empty.
Question:
If exactly in the middle of the game the score is 4 - 2 in favor of the grandmaster, it is useless to play further. Why? Who is likely to win, and by what margin?
Answer:
As many of our readers correctly guessed (thank you!), a game with such rules will always end in a draw. The box contains equal numbers of white and black pieces. If two white pieces are taken out of a box, it means that the number of black pieces inside is 2 more than the number of white pieces. If the score is 4-2, then the players have taken out at least 8 white and 4 black pieces. This means that there are 4 more blacks in the box. This guarantees the sharper at least two additional points: at the end of the game the score will be 4-4 or more - but the game will end in a draw in any case.
