Beware of the Old Books
When the "Most Instructive" games aren't instructive the way the author thought.
Having grown up in the pre-computer age, I’ve long been a fan of chess books and old games. Some of it is an appreciation for genuinely great chess, but there’s a touch of nostalgia and sentimentality as well. Still, I think I’ve managed to avoid an uncritical appreciation for older games and books, and while the old material can be a source of forgotten treasure there’s a lot of fool’s gold there as well - watch out.
When I was a youngster starting out, I generally bought best-games collections, ideally by the players themselves, and didn’t buy “instructional” books after the first year or two. Years later, when I started working with students, especially younger/newer players, one or two other chess coaches mentioned Irving Chernev’s books, like *The Most Instructive Games of Chess Ever Played*. The title, I can assure you, is the 1965 version of clickbait. (Are the 62 games there really the most instructive games ever played up to that point? And instructive for whom? Is it the games that are instructive, or Chernev’s notes? Etc.)
Leaving aside the hyperbole, how’s the book? How are the games?
Perhaps I’ll do an occasional series on it; in this post, I’ll discuss the second game in the book, Tal-Lisitsin (rendered “Lissitzin” in the book) from the 1956 USSR Championship in Leningrad. It’s an interesting game in that Tal wins playing against type. Rather than winning with the spectacular slash-and-burn approach typical of his youth, he plays what seems to be a positional masterpiece. First he creates and controls some weak squares in the enemy camp, then he leaves poor Lisitsin with a sorry collection of weak pawns, and then a clever king walk leads to the pawns’ collection and a winning rook ending he easily converts.
I was tempted to use that game with students and/or do a ChessLecture video on it. Unfortunately, once I took a closer look - yes, with the engine - the game didn’t hold up. The game is in fact far messier than Chernev’s comments would suggest, and Lisitsin was much better, even winning, at times, and wasn’t worse until very near the end. Basically, Lisitsin outplayed Tal in the opening and early middlegame, and found himself with a (briefly) winning endgame. However, he started to drift while Tal played very purposefully, and after wasting numerous tempi that ending went from won to better to equal to lost.
It is unfair to blame Chernev for not having access to Stockfish, obviously, but it is fair to say that he is guilty in this case of annotating by result and/or narrative. Tal, a legend of the game, won, and it fits so nicely the picture of winning a model game that Chernev tells a pretty story - an “instructive” story - that leaves out a lot of the game’s reality.
I’ve annotated the game twice. In the first, I present the game with very few notes, offering a few exclamation points where Chernev provides them, and I throw in a few comments that are not quotes from the book but which summarize the general narrative. In the second version, I provide fuller annotations that give a more realistic sense of what actually happened. (The games are here.)
The point isn’t to pick on Chernev as to warn readers of two things - things that are true of contemporary books as well as the oldies, but the problems are endemic to all but the very best of the old-time material. First, serious analytical errors are going to be ubiquitous (errare humanum est, as I’ve noted on the blog countless times over the years). Second, the temptation to commentate by the result and spin an “appropriate” narrative was common as well. So beware of that, and for that matter beware of it in your own games as well.
I do own the book and it never gave me the sense the games, while entertaining, provided key insights to chess. Chernev's books have always a bit problematic. My understanding was that he was a USCF Expert at best. However, his very readable books had a lot of entertaining games to expose his readers to not only well known, even forgotten players. They also captured his love of the game and were a good attempt to pass that on to casual players.
At least he did not provide analysis that was deliberately wrong or games that were more fiction than fact. Before chess engines, the amount of incorrect analysis found in books, written by Grand Masters, was not insignificant.
I wonder if fitting the narrative to the result is a form of confirmation bias. Which may be inevitable when trying to understand the thought process of others, butt is good to be cognizant of.