Given the use of the web that allows individuals to play chess regardless of how far apart they actually live, we may make the following unremarkable claims. The barrier of space is no longer insurmountable, and each year’s discoveries, by facilitating communication, bring the distant nearer together. Moreover, the speed of the web renders an approximation possible to the conditions of actual play over the board.
Both of these statements are so clearly true now in the early 21st century that they are not particularly interesting. But what if I told you that, but for a single word, both statements were also true in the latter part of the 19th century? Would your interest level rise?
In 1890 an Englishman named Edwyn Anthony published a pamphlet in London entitled Chess Telegraphic Codes in which he wrote those two sentences, but using ‘telegraph’ where I wrote ‘web’. While long obsolete, his work is nonetheless fascinating for its creativity to reduce the cost of transmitting chess moves for correspondence play via telegraph. Anthony devised a system by which two consecutive moves could be transmitted as just a single word.
Of course, in a single telegraphic game a player could only transmit one move at a time, not two. But Anthony hoped that his work would facilitate the long-distance play of tournaments and other multi-game events such that the two moves could be for different but simultaneous games. As he put it, “Since by their use the rate of inland telegrams becomes one halfpenny for every two moves sent, the question of expense need scarcely debar players and clubs…from engaging one another.”
He actually devised two telegraphic notation systems, one based on 5-digit numbers and the second based on a special chess cipher dictionary that used words from the eight languages approved by the International Telegraph Convention. Both systems relied on yet another system called ‘Compass Notation’, which Anthony credits to an unknown inventor.
Compass notation relies on the observation that almost any chess move can be indicated by designating the piece followed by one of eight compass directions and the number of squares the piece is to move. The compass directions are the abbreviations of ‘N’ and ‘S’ to indicate travel along the files, ‘E’ and ‘W’ for the ranks, and travel along the diagonals indicated by NE, NW, SE and SW. This suffices for all chess moves except for knight moves.
Thus, by ‘Queen’s Bishop SW 5’ we indicate a bishop traveling 5 squares along a diagonal from the Black side toward the White side of the board. While ‘King E 2’ or ‘King W 2’ could be used in this system to indicate kingside and queenside castling, respectively, for some unknown reason Anthony chose simply to designate those moves as ‘Cas (K)’ and ‘Cas (Q)’.
For knight moves, he used the digits 1-8 from a clockwise perspective as shown in the following diagram. Thus, a Nf3 move from g1 becomes “Knight 8”.
Finally, Anthony chose to employ a simplifying assumption that a game would rarely occur where more than three pawns of the same color would reside on the same file. With this assumption, any chess move can be represented by one of 268 symbol strings in compass notation. He put each move into a 1-to-1 correspondence with the numbers 1-268. The numbers 269-315 were either put into relation with game results or actions, such as ‘resigns’, or else one of as many as 42 different players, e.g. Player ‘E’ or player ‘g’.
Finally, he also associated each compass move with a range of numbers in the interval 1-99541. For example, the compass notation move “Queen’s Rook E 2” was not only associated with the single number 103 but also with the range 32233 to 32549. Anthony placed all of these compass moves and related numbers onto a large chart, a snapshot of which is shown here. This chart comprises what he called the Figure Code, which is the first of the two systems I mentioned above.
We are now ready to see how one number can be used to represent two chess moves. We select a number in the Table A range such that it satisfies the equation Num = Low + Single, where ‘Low’ refers to the lower bound of the range in Table A and ‘Single’ refers to the single number in Table B. The two moves thus represented are (a) the move that is associated with the Table A range, and (b) the move associated with the Table B number.
For example, suppose that the first move is represented in compass notation as “King N 1” and the second with “Queen’s Rook E 2”. From the chart above, we see that the first move is associated with the range 1 to 317 so we want to select one of the numbers in that range in order to refer to this first move. Since the second move is associated with the single number 103, the number we are looking for is found by solving the trivial equation Num = 1 + 103 which is 104.
When ‘104’ is transmitted by telegraph the sender is charged for one word and the receiver, who also has a copy of the table, finds the first move by locating that number in the Table A range and the second move by subtracting the lower bound of that associated range from 104 and then looking up that result in Table B. As an exercise, you can verify that the number 3887 represents the two moves “Queen N 3” followed by “King’s Rook W 7” since 3887 = 3793 + 94.
Anthony wanted to limit his system to numbers of 5 digits or less because those numbers were charged as a single word by the telegraph office.
Here is an 1834 game played between McDonnell and La Bourdonnais followed by the moves of that game in both compass notation as well as Anthony’s numeric Figure Code notation.
If you had the full chart you could determine that the first number 87842 refers to McDonnell as player ‘E’ and La Bourdonnais as player ‘n’; and the last number 59362 represents the final knight move followed by resignation.
Anthony referred to his verbal alternative to the numeric system as the Word Code, which required the construction of a 99,255-word cipher dictionary to be constructed as 315 large pages where each page held 315 words of ten letters or less, again because that was the length limit set by the telegraph office for the cost of one word.
The words were to be taken from any of the languages Dutch, English, French, German, Italian, Latin, Portuguese or Spanish. Anthony only constructed one such page due to the sheer effort of such an undertaking, and appealed to the world of chess to assist him in the completion of this dictionary. A sample of his page is shown below, the use of which is simpler than his Figure Code above.
Each of the 315 pages represents one of the compass moves or one of the additional pieces of information as previously discussed for the Figure Code. The example page below is for the move “King N 1”, which can be seen at the top of the page. Any word from this page of the dictionary refers to this as the first of the two-move sequence. The second move is indicated by which specific word is selected from the page. Thus, the word ‘Abecquer’ not only designates “King N 1”, but also “Queen’s Rook E 2” as can be seen below. These are the same two moves I illustrated above with the Figure Code number 104.
I do not know if anyone ever helped Anthony complete his multilingual chess cipher dictionary, and I did not find any evidence that his numeric Figure Code ever caught on. Sadly for him, his work was completely unnecessary.
By 1890 the Gringmuth notational system was already in use in which each square on the board was named by a consonant-vowel syllable and a single move was represented by concatenating two such syllables, the first for the source square and the second the destination square. For example, White’s king knight move Nf3 could be represented as the single word ‘KAHI’, where ‘KA’ is the syllable naming the g1 square and ‘HI’ is the name of f3. Concatenating two moves would allow the transmission by telegraph of one 8-letter word, thus falling within the ten-letter limit set by the telegraph office.
But neither Anthony’s system nor the Gringmuth could transmit more than two moves at the cost of one word. Anthony even wrote, “We consider it a practical impossibility to devise a system which would enable more than two moves to be sent at the cost of one word.”
The web, of course, has made the cost of playing an entire game virtually free.