The Education of a Chess Player

Source: see below

Note: updated 21/6/2019.

Above, we see a page from Israel Rabinovich-Barav's notebook of annotated games, which he kept as a young man. The notebook was loaned to me by his son, Ami Barav. The games are mostly famous ones from the 19th and early 20th century, of which the first, Schiffers vs. Chigorin, St. Petersburg, 1897, is a typical specimen. The game above is Blackburne vs. Marco, Monte Carlo, 1901.  

There are also some games where one of both of the opponents are unnamed, and one curiosity: a game allegedly by J. J. Rousseau (vs. an unnamed opponent, given in other sources as prince Conti). Barav understandably didn't know that this game, widely circulated in various sources from the mid-19th century onward, was proven by the chess historian H. J. M. Murray to have been a forgery as early as 1908, as shown by Edward Winter in a definitive article about Rousseau's chess career. 

Why did Barav copy out famous games when he could just read them in a book? The reason is obvious. Barav uses the tried-and-true method that (almost) all strong players used to get better: learning from game collections of other players, and trying to understand what both players were doing, each side's plan, where the crucial mistake was made, why it was a mistake, etc.

For this purpose, Barav annotates on his own many of the games, sometimes in great depth. His annotations are sometimes made at great length; e.g., a Tarrasch - N.N. brilliancy is annotated in five whole pages (10 columns), almost own his own.

In the Blackburne-Marco game sample page above, he notes, for example, that 5...Bxe3 is bad since it 'opens the f-file for the [white] rook', and similarly that 10,,,Nxb3 is weak since 'Black wants to gain the upper hand on the queen's side, but by this weakens the king's side'.