George Atwood's Idea Rehabilitated
George Atwood (1745-1807)

George Atwood's Idea Rehabilitated

CM ArnieChipmunk

Analysing old chess games with a modern engine can lead to interesting insights. Recently, I did some historical research in the Philidor Countergambit for which I looked at some of the earliest games played with this variation. This is a line that was already analyzed in rudimentary form in the oldest manuscripts on modern chess. The basic position arises after 

The line was promoted by François-André Philidor (1726-1795) in his influential book Analyse du jeu des échecs, first published more than 250 years ago (in 1749) and translated into English as Analysis of the Game of Chess. It was also played by several strong players in the 19th century, including Paul Morphy and Adolf Anderssen. In more modern times, the line has often been considered "refuted" by all kinds of authors, though their judgement was (and is) often based on mechanically repeating other authorities or, in recent times, the laziest kind of computer-checking.

In Philidor's days, the main line (which he also gives in his book) ran as follows:

4.dxe5 fxe4 5.Ng5 d5

Black establishes a strong pawn centre in the spirit of Philidors philosophy that "pawns are the soul of chess". In his book, Philidor now only gave the modest 6.f4?! but it was soon discovered by the Italian analyst Ercole del Rio that 6.e6! was a much more dangerous try. (Later, still other moves were found for White. Bobby Fischer once played 6.Nc3!? in a simul game, which also isn't bad.)

After 6.e6, White obviously threatens the devastating Ng5-f7 so Black has to develop some pieces rapidly. The move 6...Nh6 suggested itself.

This position was an important tabiya in chess opening theory in the late 18th and early 19th century. It was the basis of analysis by legends like De La Bourdonnais, Staunton and Morphy (who uncorked the surprising novelty 6...Bc5!? in the middle of the 19th century). The position also was tested in several games by Philidor's contemporaries (and probably also by Philidor himself, although these games have not been preserved.)

In a number of games played between 1795 and 1798 by George Atwood against Joseph Wilson,the natural move 7.Nc3 was seen, and after 7....c6 Atwood played the ingenious piece sacrifice 8.Ngxe4!? (followed by Qh5+). With this he managed to beat Wilson 5-0 (although some authors have pointed out poor Wilson was probably very ill prepared and put up rather weak resistance.) This gave the whole variation a bad reputation at the turn of the century, which lasted for more than 50 years.

Much later, this variation was also tried in a consultation game between Staunton and Morphy (one of their only two recorded encounters behind the board), in which the American introduced an interesting new idea for Black in an old line, and went on to win the game in brilliant fashion - much to Staunton's chagrin, who later claimed in Chess Praxis that a mere "hasty slip" had cost him and his consultation partner the full point.   

However, Atwood also came up with other ideas for White in the diagram position. In a game with Count John Brühl (1736-1809), who also played many a game with Philidor, he invented a concept that hasn't been taken seriously by any theoretician since. Instead of 7.Nc3, Atwood here played the quiet but logical 7.g3!? with the simple idea to just defend the e6-pawn by means of Bh3.

This move isn't even mentioned by most authors who've looked at this variation in some depth. For instance, it's entirely ignored by both Tony Kosten (1992) and Christian Bauer (2006) in their respective monographs on the Philidor. Going back to earlier days, the move is also not mentioned in the Handbuch by Von Bilguer and the Modern Chess Instructor by Steinitz. Morphy doesn't mention the possibility either in an article he wrote together with De Rivière in the magazine La Nouvelle Régence in 1863. Besides 7.Nc3, virtually all analysts focus almost exclusively on moves like 7.c4, 7.Qh5+, 7.Nxh7 and 7.f3.

To their credit, the authors Jim West (The Dynamic Philidor Counter-Gambit, 1996) and Macon Shibut (The Philildor Counterattack, in the book Pour Philidor, 1994) do mention the possibility of g3, but they don't recognize its value. To be fair to them, the move does look rather insipid, but entering the position in Stockfish 11 just the other day, it struck me how favourable it evaluates White's position after 7.g3.

Actually, Stockfish alternates between preferring this move to Nc3 depending on how long you let it think. And even after 7.Nc3, the machine wants to play g3 on the next move. It seems Atwood was on to something when he played this more than 200 years ago!  

Let's look at the opening phase of the game so we can get an idea of how the old masters interpreteted the position. Bear in mind this game was played more than two centuries ago - some concepts which  to us seem pretty basic, had not been uncovered yet. 

Clearly, the level of play from both sides wasn't exactly up to modern standards - let alone when scrutinzed with the latest generation of chess engines. Let's not criticize these pioneers for their middle game treatment but instead admire their contributions to the evolution of chess ideas. These still speak to us over the centuries. 

The truth is that until very recently, nobody seems to have understood just how valuable the idea of g3 and Bh3 is in this position. Not because of a potential follow-up with f4-f5, as in the game, but for material reasons - and because it can be combined with a timely c2-c4! undermining black's centre while holding on to the strong pawn on e6. Modern engines can in this way enrich old opening ideas and create new theory!  

Because of his loss in this game - and maybe also because 7.Nc3 and 8.Ngxe4 was a much more attractive line! - Atwood's idea of g3 and Bh3 was ignored for a long time, but it deserves a much fairer treatment in the history of chess opening theory. I think it's time to rehabilitate the line and rename 7.g3 the "Atwood variation".