By Frederic Friedel

This paper appeared as a contribution to the book "Advances in Computer Games 9", edited by Professors H. J. van den Herik, University Maastricht, and B. Monien, University of Paderborn. It was published by the Universiteit Maastricht in 2001. The paper on Cheating was written and submitted by the author in 2000.

Abstract

Nowadays, players at all levels of chess can profit from computer assistance during a game. This is a new development and a serious problem for the game. This contribution lists the main forms of cheating and provides some occurrences from practice. The most prevailing one (Allwermann at the Böblinger Open) is placed in a historical context by describing previously noticed cases of cheating. Finally, the problem of cheating is addressed at the highest level of play. What are the possibilities and how can we prevent cheating at this level? Since there is no clear solution, the issue of cheating remains on the list of issues to be addressed very seriously in the near future.

1. Introduction

The problem of cheating in chess is certainly not new. It has become acute in recent times, but we must not forget that cheating has been going on for a long time. Probably the first dishonest players emerged just after the invention of the game (around AD 600), and cheating continued throughout the history of chess, more or less unchanged, up to the present time.

The main forms of cheating are:

1. Manipulating the position (removing or changing pieces, playing more than one move, etc.). This is rarely seen in tournament chess, it is more of a coffee-house practice.

2. Purposely losing or drawing in return for some reward. This is the most common form encountered in international tournaments.

3. Consulting a stronger player during the game. This is unfortunately quite common in team championships, where a player from the top board may assist his weaker colleague, or in fact help a player from a different team if he is playing against a strong rival team.

4. Other miscellaneous methods of cheating.

Already in the middle ages we have examples of dubious practices by chess masters. The best known – certainly the most often quoted – are by Lucena (“Try to play after your opponent has eaten or drunk freely”) and Ruy López (“Place the board so that the sun is in your opponent’s eyes”). More recent examples include bringing a cat to the playing site when you know your opponent is allergic to these animals (allegedly done by Alekhine), smoking cigars or, more devastatingly, threatening to smoke them (as in the famous story of Vidmar against Nimzovitsch), fidgeting, humming, even placing your pieces off-centre if you know your opponent (e.g. Meking) is obsessed by neatness, and countless similar ploys.

1.1 A Classical Swindle

In chess it is important to distinguish between cheating and a “swindle”, which is a perfectly legal way of tricking your opponent without breaking any of the rules of chess. Here’s an example of a classical swindle.

Short,Nigel D (2440) - Radulov,Ivan (2465) [A36]
Hamburg op Hamburg (6), 1981

In 1985 the 15-year-old Nigel Short was playing in a GM tournament in Hamburg. In round six he had outplayed the Bulgarian grandmaster Ivan Radulov with white, reaching the position shown above. The only reasonable course of action for Radulov would have been to resign. Instead, assuming that his young opponent was fairly inexperienced, he tried a clever little trick:

38...Qf3 39.Rxd7 Qd1+ 40.Kg2. The moves were executed swiftly, as if Black had seen some kind of a perpetual check. And now came the trap: 40...Qe2+. Naturally Nigel was expecting his opponent to deliver the next check on d2 or c2, and he was all set to play 41.Kh3. Had he done so mechanically he would have lost the game.

Luckily Nigel did not fall for the trap. When he saw what his opponent had done he punished him in an amusing way. He went into a deep think. There was a crowd of spectators watching the game, and everyone began to laugh. Radulov could only grin sheepishly and stick out his hand in resignation. Nigel played 41.Qxe2 before he accepted.

1.2 An Example of Genuine Cheating

A case of genuine cheating occurred during a tournament in Zagreb, which Fischer was dominating ahead of Smyslov, Petrosian and Korchnoi. In his game against the relative outsider Kovacevic the unstoppable American set the following trap:

Fischer,Robert James - Kovacevic,Vlatko [C15]
Rovinj/Zagreb Zagreb (8), 21.04.1970
1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.a3 Bxc3+ 5.bxc3 dxe4 6.Qg4 Nf6 7.Qxg7 Rg8 8.Qh6 Nbd7 9.Ne2 b6 10.Bg5 Qe7 11.Qh4 Bb7 12.Ng3 h6 13.Bd2 0-0-0 14.Be2 Nf8 15.0-0 Ng6 16.Qxh6 Rh8 17.Qg5 Rdg8 18.f3

It looks as though Black can win after 18...Nh4 19.Qe5 Nd7 20.Qf4 Nxg2 21.Kxg2 Rh4 and White must give up the queen or be mated: 22.Qe3 Rxh2+ 23.Kxh2 Qh4+ 24.Kg1 Rxg3+ 25.Kf2 Rg8#.

However, Petrosian and Korchnoi, who were watching the game, spotted Fischer’s deadly intention: 18...Nh4 19.fxe4! Rxg5 20.Bxg5 and it is White who is going to win. However, Petrosian’s wife, famous for her radical partisanship, had followed the analysis of the two Soviet GMs. To Korchnoi’s horror she walked across to the board and whispered the lines to Kovacevic. He played 18...e3 and the disconcerted Fischer actually lost the game. It was Fischer’s only loss in the 17-round tournament, and he finished first, two points ahead of Korchnoi and Smyslov (Petrosian was half a point behind them).

2. Illegal assistance during the game

Coaching players during the game is probably the most wide-spread form of cheating (rivalled only perhaps by bribery and the throwing of games). Although this practice began long before the advent chess playing machines, it is computers that have added a new and dramatic dimension to this method of cheating in chess.

There are two ways in which cheating can occur with computers involved:

1. Human masters influence the moves of the machine so that it plays better than it would normally be able to do.
2. A human player will use the assistance of a machine during the game to play better than he would normally be able to do.

In this article we are mainly concerned with the second possibility, although there would be a great deal to say about the first as well, a historic example of which was the famous automaton of Wolfgang von Kempelen.

In the late 18th century this Hungarian nobleman constructed a machine in the form a mechanised Turkish figure. It played chess and beat almost all opponents. Of course it was operated by a human being cleverly hidden within the apparatus.

2.1 First Use of Computers to Cheat

The first time (to the best of my knowledge) that a computer was used to clandestinely help a human player during a game, occurred in Hamburg, Germany, in August 1980. The perpetrators of the deception were the author of this article, a few colleagues from a German TV station, and Ken Thompson of the Bell Laboratories. The victim was German grandmaster Dr Helmut Pfleger.

Unix inventor and pioneer in computer chess: Ken Thompson

At the time we were making a science documentary about computer chess and wanted to perform a chess Turing test of sorts. Dr Pfleger was giving a simultaneous exhibition at the Hamburg chess festival, and we decided to secretly play a computer against him. Ken had just finished constructing his new Belle machine. We hid a radio receiver under the hair of a young colleague of mine, Dieter Steinwender, who had a place in the simul. I was able to talk to him from a vantage point high above the tournament hall. Ken was standing by in New Jersey to deliver the moves by phone.

The first special-purpose hardware chess machine Belle

Using a pair of binoculars I followed the moves on Dieter’s board. As soon as the GM made a move I immediately relayed it by phone to Ken, who entered it into the computer. When Helmut approached the board again I would warn Ken, and he would give me Belle’s current best move. This I dictated by radio transmission to Dieter’s earphone, and he executed it as naturally as possible on the board.

Pictures from the original TV feature: Steinwender getting the receiver, Pfleger at work,
Thompson and Friedel cheating, Pfleger amazed to hear what we had done...

After some hours Helmut Pfleger was winning all his games, including the one against Belle. However, at move 49 he missed a clear win.

Pfleger,Helmut - Belle [D14]
NDR Turing Test, 09.02.1980
1.d4 d5 2.c4 c6 3.cxd5 cxd5 4.Nc3 Nc6 5.Nf3 Nf6 6.Bf4 Bf5 7.e3 e6 8.Bb5 Nd7 9.0-0 Be7 10.Qe2 Rc8 11.Rac1 Bg4 12.h3 Bh5 13.g4 Bg6 14.Ne5 Ndxe5 15.Bxe5 0-0 16.Bg3 f5 17.Bd3 Bd6 18.f4 Bb4 19.g5 Qe8 20.a3 Bh5 21.Qd2 Ba5 22.b4 Bb6 23.Kh2 Rf7 24.Nb5 Rd7 25.a4 a5 26.bxa5 Bxa5 27.Qb2 Bb4 28.Be1 Bxe1 29.Rfxe1 Qd8 30.Qa3 Bf3 31.Kg3 Bh5 32.Rc2 Qa5 33.Rec1 Ra8 34.Nd6 Qxa4 35.Qxa4 Rxa4 36.Nxb7 Rxb7 37.Rxc6 Ra3 38.R1c3 Rxc3 39.Rxc3 h6 40.h4 Kf8 41.Ba6 Rb6 42.Bf1 hxg5 43.fxg5 g6 44.Kf4 Ke7 45.Ke5 Bf3 46.Rc7+ Kd8 47.Rg7 Bh5 48.Ra7 Bf3 49.Ba6 Bg2

Here Ken was saying he thought Belle should resign. The computer saw that 50.h5 wins easily: 50...Be4 51.h6 Rb2 52.h7 Rh2 53.Kd6 Ke8 54.Bb5+ Kf8 55.Kxe6. However, instead of playing this line, after which Ken would have resigned for the machine, Pfleger went for a complicated line on the other side of the board: 50.Ra8+? This gave Belle good counterplay and gradually the machine turned the tables. 50...Kc7 51.Bc8 Rb3 52.Kf4 e5+ 53.Kxe5 Rxe3+ 54.Kf6 f4 55.Be6 Kb6 56.h5 gxh5 57.g6 f3 58.Bxd5 Bh1 59.Ra1 f2 60.Rb1+ Kc7 61.Bxh1 Re1 62.Bg2 Rxb1 63.g7 Rg1 64.g8Q.

In New Jersey Belle saw that the h-pawn is going to queen and it is all over. 64...f1Q+ 65.Bxf1 Rxg8 66.Bh3 Kd6 67.d5 h4 68.Be6 Rg3 0-1.

Immediately after the game we pointed a camera at Helmut Pfleger and asked him whether he had noticed anything unusual in the games. Nothing. One of the games, we told him, had been played by a machine. Helmut was very surprised. “Which game was played by a computer?” He was amazed to hear it is the one he lost. “I really noticed nothing. Wow, these things are really playing quite well these days.”

2.2 The Chess Turing Test

A few days after the event we printed out five of the games that were played in the simultaneous exhibition and sent them to experts all over the world, asking them to identify the computer. Most of the chess experts guessed wrong, most of the computer experts got it right. An exception from the first group was the 17-year-old Garry Kasparov, who was participating in the World Junior Championship in Dortmund. When I showed him the games at the closing dinner he read all five simultaneously from the printout and unerringly identified the computer. All the other games, he explained, contained short-term tactical errors which a computer would never commit.

The Friedel-Thompson-Belle game was of course not strictly an example of cheating. It was an experiment in which the deception was immediately revealed. The same cannot be said about the following examples in which the perpetrators tried to hide their activities as best they could.

2.3 The von Neuman affair

At the World Open 1993 in Philadelphia a completely unknown player appeared, unsubtly calling himself John von Neumann (he was black and had dreadlocks). He played excellently, drawing against GM Helgi Olafsson in the second round. But in round four he suddenly stopped playing at move nine and lost on time. Here’s the game:

J. von Neumann - D. Shapiro [C44] Philadelphia Open, 1993: 1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.e5. Very strange move. Maybe von Neumann had missed a black move, e.g. ...Nf6. 4...Nge7 5.Be2 Nf5 6.0-0 Be7 7.Nbd2 0-0 8.Nb3 d6 9.exd6 Qxd6 and now White stopped playing and lost on time. 0-1.

This is how the game from round nine went:

J. von Neumann - NN [B40] Philadelphia Open, 1993: 1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 Bb4 6.e5 Nd5 7.Bd2 Nxc3 8.Bxc3 Bxc3+. Here von Neumann thought for forty minutes, although there is only one reasonable move (pawn takes bishop). Then he disappeared for a while, came back, played 9.bxc3 and won the game. Obviously there was some communication problem that had to be solved.

Von Neumann won a prize in the category of players without an Elo rating. Naturally people had become suspicious of this unknown and highly unorthodox player. Before the organisers handed over the $800 check they asked him to solve a simple chess puzzle. He refused, turned and left, and has never been seen again at chess tournaments.

3. Allwermann at the Böblinger Open

At the Hoogovens super-tournament in Wijk aan Zee, in January 1999, there was one main topic of conversation. Every morning at breakfast Garry Kasparov would come over to Vishy Anand's table to discuss the matter with him. The object of their interest was the new German chess star, Clemens Allwermann. This hitherto unknown player had won the Böblinger Open, ahead of GMs and IMs. He had played wonderfully courageous attacking chess, scoring 7.5 points from nine games, without a single loss. His Elo performance was 2630. This sensational result made the headlines all over Germany – chess on front pages of straight newspapers.

So who is this Clemens Allwermann? If I said this was a 17-year-old super-sharp kid, rising meteorically, as they sometimes do, you might still find the sheer magnitude of his success difficult to believe. It turns out that Allwermann was 55 years old and had had a stable rating of around 1900 for the last twenty years. He ranked around number 10,000 in Germany, but his remarkable performance could theoretically elevate him to number two (behind Artur Jussupow) in this country.

News of Allwermanns achievement spread quickly – but perhaps not in the way the player himself may have anticipated. “Was a pocket Deep Blue used for brain doping” asked one newspaper, while Der Spiegel put it more sarcastically: “Move over Goethe, Beethoven and Einstein, here comes a new pinnacle of Teutonic genius”.

Nobody was willing to believe that the amateur had done it all on his own. Especially not when other players discovered that you could reproduce virtually all of Allwermann’s moves with the chess program Fritz. All of this was revealed by Hartmut Metz in newspapers and Schachmagazin 64. He gave many examples that show it was not just the tactically brilliant shots played by Allwermann but also bad, anti-positional moves that Fritz will duplicate.

The game that caused the greatest suspicion was the one Allwermann played in the last round. It was an encounter that would decide the winner of the tournament, and the amateur player was facing the leading Grandmaster. Completely undaunted Allwermann proceeded to outplay his opponent in fine attacking style, reaching the following easily winning position:

Allwermann,Clemens - Kalinitschew,Sergey (2505) [B32]
Boeblingen op 15th Boeblingen (9), 30.12.1999
1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 e5 5.Nb5 d6 6.c4 Be7 7.Be2 a6 8.N5c3 Nf6 9.0-0 Be6 10.Be3 0-0 11.Na3 Nd7 12.Qd2 Nc5 13.Nc2 f5 14.exf5 Bxf5 15.Bf3 Kh8 16.Bd5 Qe8 17.Rad1 Qg6 18.Na3 e4 19.f3 exf3 20.Bxf3 Ne5 21.Nd5 Bh4 22.Nf4 Nxf3+ 23.Rxf3 Qe8 24.Nd5 Ne6 25.Rdf1 Qg6 26.b3 Rf7 27.Nc2 Ng5 28.Bxg5 Bxg5 29.Qf2 Bxc2 30.Rxf7 Bf6

I showed the position to a number of players in Wijk aan Zee, and all gave me simple wins – for instance 31.Rxb7, 31.Rd7 or even 31.Rxf6. Remember, the first time control is looming and tournament victory is in grasp.

So what does our hero play? 31.Qa7?!! “Fritzy!” squealed Anand and went into uncontrollable fits of laughter when he saw this and the following moves (I filmed his mirth and included it in my multimedia report in ChessBase Magazine 69). He and the other players immediately recognised the “hand” of the computer.

If you switch on multiple-variation mode you will see that Fritz thinks 31.Qa7 is a tenth of a pawn better than the other alternatives. A computer program simply doesn't understand the difference between cast-iron moves that cannot fail, and a tight-rope walk on the edge of the precipice. I venture no human would undertake the latter course in the given tournament situation. Play through the continuation and judge for yourself.

31...Rg8 32.Qxb7 Be4. Black is threatening mate on g2, and there is only one move that avoids immediate disaster: 33.Nf4. These are razor-sharp lines, the kind Fritz thrives on. 33...Qf5 34.Qd7 Qe5 35.Kh1 g5. Most human players would now play 36.Nh5 – the knight moves to safety and at the same time attacks an enemy piece. But Fritz and Allwermann cannot be bothered with such timid strategies. 36.Nh3 g4 37.Nf2 Bf5.

In this position White just needs to move the queen out of the line of attack. He can then sit back and wait for the opponent to stretch out his hand in resignation. But Fritz and Allwermann seek a further escalation of tension: 38.Nxg4. Nerves of steel – or maybe silicon? Cross your heart, is this a move you would play – two moves before the time control, with the overall tournament victory depending on the outcome of this game? 38...Be4 39.R7xf6 Bxg2+ 40.Kxg2 Qe4+ 41.Kh3.

Here Kalinitschev resigned. While they were shaking hands Allwermann couldn't resist mentioning that the final position was mate in eight. “I don't think so,” said Kalinitschev. “Check it out, you'll find I'm right,” replied Allwermann with a wry smile.

Well, some of the more experienced players did exactly that. They switched on their computers and fired up Fritz. The program reacted exactly as Allwermann had done (except perhaps for the smile).

On a PentiumII-400 with 64 MB hash tables Fritz 5.32 finds a mate in nine – you have to deduct the initial black move when you announce it as a mate for White.

I showed the position to a number of players in Wijk – nobody was able to spot the mate. Vishy Anand said that if I had asked him with no prior information he would have guessed it is mate in about fifteen moves. Otto Borik, the editor of Schachmagazin, showed the position to players, including many GMs, at a team championship match. They couldn't find a mate either. Borik tried to work out all the lines. He says they would fill an entire issue of his magazine. Here are some critical variations: 41...Qe8 42.Qf5 Qc8 43.Qf4 a5 (43...Qxg4+ 44.Qxg4 Rxg4 45.Kxg4 Kg7 46.Kh5 a5 47.Rf8 a4 48.R1f7#; 43...Qc5 44.Rf8 Qh5+ 45.Kg2 Qg6 46.Rxg8+ Qxg8 47.Qd4+ Qg7 48.Rf8#; 43...Qe8 44.Rf7 Rg7 45.Qd4 Qe5 46.Nxe5 Rg3+ 47.Kxg3 Kg8 48.Ng6 hxg6 49.Qg7#) 44.Rf7 Rg6 45.Rf8+ Qxf8 46.Qxf8+ Rg8 47.Qxd6 a4 (47...Rxg4 48.Rf8+ and 49.Qf6 mate) 48.Qe5+ Rg7 49.Rf8#. Only Fritz is sure it's mate – after running through 33.117 million positions.

After the Böblinger Open a number of experts also got to work, analysing Allwermann’s games. They all discovered the same thing: on tournament time controls Fritz plays most of the other moves Allwermann executed on the board. There was a lot of speculation on how he might have got the moves from the computer. The tournament director in Böblingen recalls that although the temperature in the playing hall was usually 32 degrees Centigrade Allwermann always appeared with a tie and a dark blazer. Some participants speculated that a miniature camera might have been hidden in his tie. The long hair and spectacles could easily have hidden a receiver. Hartmut Metz reports that Allwermann used to run a store for electronic equipment, so the expertise was probably available.

Unfortunately Allwermann talked himself into a corner. Soon after the tournament he published a letter on the Internet saying that his result had been the crowning achievement of a forty year chess career, attributing his play to careful preparation, especially in the Sveschnikov. “I now have a good knowledge of this variation, which I was able to use to defeat GM Kalinitschev in the last round.” This made it very difficult for him to admit any wrongdoing after the accusations had been raised in the press. He could have conceded he had used a computer, perhaps claiming that it had been his intention to demonstrate an important and dangerous development in chess today, and returned the DM 1660 prize money. Instead he threatened to sue anyone who accused him of cheating – although he didn't follow up when Der Spiegel published their story anyway.

After the articles had appeared in German newspapers and magazines Allwermann wrote a long letter protesting his innocence. Among other things he complained that nobody had published games that would exonerate him. “The second game I played should be of interest to an objective reader. Anyone can see that a computer program would have easily won this game.”

Okay, let’s take a look at this game – with Fritz.

Giacopelli,Vincenzo (2145) - Allwermann,Clemens [A46]
Boeblingen op 15th Boeblingen (2), 26.12.1998
1.d4 Nf6 2.Nf3 e6 3.Bg5 c5 4.e3 h6 5.Bxf6 Qxf6 6.Nbd2 cxd4 7.exd4 Nc6 8.c3 d5 9.Bd3 Bd6 10.0-0 0-0 11.Re1 Bd7 12.Nf1 Rac8 13.Ne3 a6 14.Ng4 Qf4 15.h3 Bc7 16.g3 Qd6 17.Qd2 Ne7 18.Nh4 Bb5 19.Bc2 h5 20.Ne5 Be8 21.Qd3 g6 22.g4 f6 23.Nef3 hxg4 24.hxg4 Bf7 25.Qe3 g5 26.Ng2 Kg7 27.Kf1 Rh8 28.Ke2 Qb6 29.Kd2 Ng6 30.Bxg6 Bxg6 31.b3 Rh3 32.Qe2 Rch8 33.Rg1 Be4 34.Nfe1

As usual in this tournament Allwermann has outplayed his opponent and is easily winning. Fritz5.32, which reproduces all the moves so far, now shows an evaluation of 4.06 in favour of Black and expects the continuation 34...Bh2 35.f3 Bxg1 36.fxe4 Rxc3 37.Nc2 Bxd4 38.Rc1 Qa5 39.Nxd4–+.

So what does Allwermann play? 34...Bxg2? The wrong bishop! After this move Fritz's evaluation drops by almost four pawns. The game continued 35.Nxg2 Rh2 and a draw was agreed.

There was some speculation as to whether Allwermann was receiving moves in a kind of morse code: dit-dit-dit-dit means “bishop”, dit-dit-dit-dit-dit-dit-dit-dit is h, and dit-dit is 2. Oops, did somebody miscount a dit in the above position? Remember, it was round two and things probably had perhaps not settled down yet.

Some months after the event the District Attorney’s office began investigating Allwermann for embezzlement of the prize sum of DM 1,660 – (about $850). GM Rainer Knaak was consulted, and the Fritz expert (Knaak works for ChessBase) confirmed that all the games were almost completely reproducible, move for move, with Fritz5.32 and the Fritz Powerbook '99. Even a small transposition error in the PowerBooks was faithfully reproduced in one of the games. In the meantime Hartmut Metz had located an electronics supplier who had sold Allwermann the equipment he probably used to transmit the computer moves. According to the store owner Allwermann had insisted on a modification that would allow him to enter four-digit codes in the hand-held radio transmitter. He had also purchased the very smallest receiver possible, one that could be completely concealed in his ear and hidden behind his long hair.

However, after many months the DA’s office dismissed the case due to “lack of sufficient proof”. A speaker expressed the view that “moves by good chess players often coincide with those of a computer”, and apart from that there was no direct evidence – nobody had seen or documented the use of electronic devices during the tournament. The Bavarian Chess Federation, on the other hand, has taken drastic action and barred Allwermann from participation in further tournaments.

Before this was enforced Allwermann had played in one more tournament, closely watched by large numbers of spectators and journalists. He scored exactly what is to be expected of a player who is below the 2000 Elo mark.

4. Computer assistance at the highest level

Potentially computers can play a decisive role at the very highest levels of chess. This was made very clear to me during the Super GM tournament in Las Palmas in 1997. In round four of this tournament Garry Kasparov played a very nice attacking game against the world’s number two Vishy Anand. I was following the moves with Fritz in the press room, together with some of the grandmasters present there. Here’s how the game went:

Kasparov,Garry (2785) - Anand,Viswanathan (2735) [B92]
Las Palmas Las Palmas (4), 12.12.1996
1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 a6 6.Be2 e5 7.Nb3 Be7 8.0-0 0-0 9.Kh1 b5 10.a4 Bb7 11.Nd5 bxa4 12.Rxa4 Bc6 13.Ra3 Nxe4 14.Na5 Nf6 15.Nxc6 Nxc6 16.Bc4 Nd4 17.Rh3 g6 18.Qd2 Nf5 19.Nxf6+ Bxf6

At this point Kasparov went into a deep think. Jan Timman started to speculate whether White couldn’t play the very forceful 20.g4. Kasparov’s second Juri Dokhoian immediately confirmed: “That’s what he’s looking at!” Yuri understands Kasparov’s thinking better than anyone else in the world.

We started analysing the position with Fritz, and soon we had the following lines: 20.g4! Qc8 (20...d5 21.gxf5 dxc4 22.Qh6 Qd5+ 23.f3 Rfd8) 21.Bd5 Nh4 22.Rg1! g5 (22...Rb8 23.Qh6±; 22...Bg7 23.Rxh4) 23.Rxh4 gxh4 24.g5 Bg7 25.g6 Kh8 (25...Qf5 26.gxf7+ Kh8 27.Bxa8 Rxa8 28.Qd5 wins) 26.gxf7 Qf5 27.Bxa8 (27.Rxg7 Kxg7 28.Qh6+ Kh8 29.Bg5 Rxf7 30.Bxa8=) 27...Rxa8 28.Qd5 Rf8 29.Bh6! This final point, found by Fritz, is especially important and clinches the line. Our full analysis was published in CBM 57.

Juri Dokhoian checking the 20.g4 line with Fritz in real time during the game, Kasparov and Anand analyse after it six hours of play.

Meanwhile White had played 20.Bd5. The game lasted six hours, Anand defended very tenaciously and at around 10 p.m., much to the disappointment of Kasparov, a draw was agreed. When he left the stage Garry spotted me and walked straight over. “I couldn't win it, could I, Fred?” he asked, with a troubled look on his face. It was a bit shocking: the world champion and best player of all times consulting a chess amateur, asking for an evaluation of the game he has just spent six hours on!

Naturally Garry wasn't asking me, he was asking Fritz. He knew I would have been following the game with the computer. “Yes, you had a win, Garry. With 20.g4!” My answer vexed him deeply. “But I saw that! It didn't work. How does it work? Show me.” He and Anand listened in horror while Juri dictated the critical lines. All of this was captured on video and published in ChessBase Magazine 56 (Feb 1997).

Kasparov and Anand learning the truth about 20.g4

The next day Garry did an interview with the German magazine Der Spiegel. He spoke about “Advanced Chess”, a new concept he has developed, which involves playing games in real time with computer assistance. He used the game against Anand from the previous day to illustrate his point. This is what he had to say: “That game provides us with new arguments for Advanced Chess. If I had had a computer yesterday, I would give you the full line with 20.g4 within five minutes. Maybe less. I would enter g4 and check all the lines. I know where to go. It would give me the confidence to play moves like this. Can you imagine the quality of the games, the brilliancy one could achieve?”

In the time since those remarks there have been two Advanced Chess matches in León, Spain. In the first Kasparov was unable to defeat Bulgarian GM Veselin Topalov, who made efficient use of Fritz to defend against the world champion. The match ended in a 3:3 draw, although Kasparov had just demolished Topalov 5:1 in a match without computers. In the following year Vishy Anand played against Anatoly Karpov. Both players were assisted during the game by ChessBase 7.0 and the chess engine Hiarcs 7.32. Karpov was quite inexperienced at operating a computer, while Anand happens to be one of the most competent ChessBase users on the planet. The result was that we were witness to an (unplanned) experiment of man and computer vs man. Karpov didn’t have a chance and was trounced 5:1 by his opponent. I am convinced that a player like Anand, using a computer to check crucial lines during the game, is playing at a practical level of over 3000 Elo points. However, this is the subject for a different article.

5. The problem

The fact that players at all levels of chess can profit from computer assistance during a game is a new development and a serious problem for chess. Ironically the world chess federation FIDE has chosen this moment in time to attack another far less critical matter and wants to introduce doping controls in chess. At the Linares tournament in March 2000 top players were asked for their opinion on the FIDE initiative. Vladimir Kramnik and Garry Kasparov both said it was fairly silly (Kasparov: “What are they going to check for, coffein in your blood?”). Especially Kramnik was very outspoken. “Instead of a doping control, which is meaningless in chess, they should have airport-style security systems at all major tournaments to make sure nobody is getting help from computers during the game. That is a much greater danger.”

Most top grandmasters understand all too well how computers can affect the outcome of a game. In contrast to an amateur playing 600 points above his true strength, for whom the computer must dictate practically every move of the game, a strong grandmasters requires only occasional assistance to improve his performance dramatically. There are usually a few critical positions in which a player must decide whether a promising plan can work, or whether it is tactically flawed. One can only speculate how many brilliancies have not been executed because the player just wasn’t able to check all the lines in the time available (or because no human is able to do so in very complex positions).

Naturally the temptation would be great to have computer assistance available in these situations. The problem is made more acute by the fact that only very little information needs to be passed on very few occasions. If we return to the example of the Las Palmas game given above, Kasparov’s second knew what he was looking at and actually uncovered the solution before the move was played. All that Kasparov needed was one bit of information: “Yes”. He didn’t need to get the message “20.g4! wins” but simply “There is a win” or even, in this specific case “The move we know you are looking at works!” That would have been enough to decide the game.

The problem is more serious the higher you go up the Elo scale. Not only do the players need progressively less information, they also stand to gain progressively more from cheating with the computer. Whereas in a big national open like Böblingen players stand to gain less than a thousand dollars, the sum increases to tens of thousands in top international tournaments, or even up to a million dollars in the case of a world championship (the difference between winning and losing – and not including the monetary benefits derived from winning the title).

Cheating with the computer has, to the best of my knowledge, thus far not been attempted at top levels in chess. But the players are all well aware of the fact that there are bound to be attempts to do so in the future. It should be noted that cheating at higher levels can probably only occur with the help of computers. Normally no human is able to reliably assist a top grandmaster who is immersed in a game and understands it better than any onlookers.

6. No clear solution

When the Allwermann affair surfaced, a number of other top players turned to us for advice on how to counter cheating in chess. The assumption was that since ChessBase makes the software that would be potentially used for clandestine computer assistance, it is we who would know how to counter such attempts. This was unfortunately not the case – there was no special expertise in our company with regard to electronic intelligence and counter-intelligence techniques.

In the months following the Wijk aan Zee tournament, however, I was able to gather a lot of information on the subject. The main source was, of course, the Internet, where there is a profusion of sites explaining the technology at every level of detail (not to mention retailers offering to sell you every form of legal and illegal devices). I also discussed with Ken Thompson the kind of information that needs to be passed on and how this would be best generated by an assistant working with a fast tactical chess program like Fritz. Finally I got some invaluable information from Mark Lefler, chess programmer and US government employee who is an expert on security at US embassies abroad. Apart from describing advanced electronic methods Mark opened a whole new can of worms by describing to me all the non-electronic ways in which information can be passed on – especially in the miniscule amounts needed for cheating in chess.

This brings us to a general problem I have in writing this article. If I were to now present a detailed account of all our findings it would be akin to writing a general instruction manual for cheating in chess. This is not my intention. On the other hand one must note that a lay person like myself can become a relative expert on the subject in a short period. It is frightening to think what a real expert with criminal intent and a substantial budget could come up with in the same time.

In this article and at the current time I have decided to be guarded in what I say and omit many of the details that would be of interest mainly to those actually interested in cheating. I will only draw attention in general terms to some of the problems faced by the counter-intelligence side of the equation. These are quite substantial.

1. It would seem to be a general rule that the cost of installing an electronic cheating mechanism is far lower than the cost of the equipment and infrastructure required to detect it. The factor appears to be about ten to one. You can start cheating in chess on a budget of around $100, and it would cost a tournament director around $1000 to reliably prevent this method of cheating from being used. If the perpetrator spends $1000 in electronic equipment then the other side must invest around $10,000 in order to thwart the attempt.

2. Airport-style detectors as envisioned by Kramnik will not pick up some of the electronic equipment that could be used to receive information during a chess game. There are currently devices on the market that use a very small amount of metal – in fact they have been specifically designed to elude detection by the usual methods. Apart from that there are ways in which a player could pass through a security detector without the receiver and obtain this later on during the game. A whole set of additional measures, many extremely restrictive, need to be implemented to prevent this from happening.

3. Electronic receivers can be very cleverly concealed in many different parts of the body, and detecting them would require unacceptably intrusive searches. Remember that the signal does not have to be acoustic or verbal – you can easily transmit chess moves or yes/no information using a buzz, click, vibration or pulse. It is very difficult to detect a receiver that is may be smaller than some of the dental appendages the user is wearing – or may in fact be part of such appendages. Certainly this cannot be achieved using a standard airport metal detector.

4. The detection of the signal used to cheat is a similarly daunting task. It is important to remember that at the higher levels of chess we are dealing with just a few very short bursts of information, occurring within a five to seven-hour period. The transfer can take place at one of many different frequencies, and the signal may be more or less defused. Detecting the receiving device itself by its tell-tale electronic emissions is possible, but exceedingly difficult, especially if the perpetrator is using modern spy equipment that is designed to avoid exactly such detection.

5. It has been suggested that future matches must be held in Faraday cages (i.e. that the playing site would be shielded from all radio signals passing in or out by wire mesh covering the walls, floor and ceiling). There are a number of problems here as well. It may be feasible to put two players into an electronically shielded room, but how do you do this for an entire tournament with many players on different boards? Even with a two-player match special arrangements would have to be made for the participants leaving the table for refreshments, exercise or other human needs. If a player may step out of the electronically shielded area whenever he pleases then there is hardly any point in making it secure in the first place.

6. Another unfortunate aspect is that the information required for cheating at a higher level of chess can be transmitted in a variety of non-electronic ways [1]. Visual methods are already quite well-known, and for many years in world championship matches there have been arrangements to restrict the visual contact between the players and their seconds. But of course it is quite easy to use an unknown person to transmit signals. If the players can see any people in the audience then there are countless methods in which someone can pass him game-decisive information. If it is done in a fairly sophisticated way a single glance is enough to read the message. In fact the player doesn’t even have to look directly at the messenger, so that detection becomes even more difficult. The only reliable way to prevent this is to make sure that the players have absolutely no visual or acoustic contact with the audience at all.

7. If countermeasures are put into place to prevent all of the above there is still no absolute guarantee that a player is not receiving outside assistance. The simple reason is that there are other, even more subtle methods that can be employed. Since there is no known precedence for their use in chess I will not start a discussion of them here.

The conclusion of our studies is that from a technical point of view it is extremely difficult to guarantee that there will be no cheating in chess by the clandestine use of a computer at decisive points in the game. The closest one could get to absolute certainty would be by implementing very radical controls: one would have to introduce rigorous body checks, isolate the players completely from the outside world during the games, perhaps even keep the playing site a secret until immediately before each game. Possibly the most effective method would be to broadcast the moves of a game only after it is over – if nobody outside the playing site has access to the moves while they are being played it becomes impossible to provide assistance by any of the methods discussed above.