Savielly Tartakower, My Best Games of Chess 1905-1954 (Russell Enterprises, 2015). 437 pp., $34.95.
Let's start with the finish and then go backward: I recommend the book to anyone interested in Savielly Tartakower, or chess history in general, or entertaining games (especially with readable and instructive annotations), or unusual opening variations.
Tartakower was one of the strongest players in the world for almost the whole of the first half of the 20th century, and was one of the game's wittiest annotators and its greatest aphorist. (See here and here for some of his classics.) He was also a great innovator, serving as at least one of the trailblazers behind the Tartakower Variation of the Queen's Gambit Declined, the O'Kelly Variation of the Sicilian, the Orangutan/Sokolsky, the Catalan, the Fantasy Variation against the Caro-Kann and plenty of others. He played solid, classical lines, Romantic lines (e.g. the King's Gambit, with the eccentric 3.Be2), gambits, hypermodern variations and if there's anything left, he played that too. His risk-taking style and love of innovation may have been both a great strength and a great weakness in his chess career: he could do just about anything, but he wasn't solid enough to handle the absolute top players like Emanuel Lasker, Jose Raul Capablanca or Alexander Alekhine. But he was neck-and-neck with practically everyone else during his long tenure near the top, and as he was one of the greatest writers of chess history his legacy is one worth celebrating.
The book comprises 201 complete and well-annotated games, 49 (also well-annotated) game fragments, Tartakower's tournament and match record, a number of crosstables and pictures, some brief introductory material and a number of indexes. (The usual material you'd expect, especially from Russell Enterprises, which generally aims to include extra photos in their books on historical players and events.)
A few remarks about the annotations. Tartakower provides variations, as you might expect, and while they're rarely as deep as the material you might see in today's chess books they compare favorably in their depth to pretty much anything written at that time. Tartakower could be sloppy (see Taylor Kingston's errata on the publisher's website, along with his discussion at its end), but despite the errors his comments are typically instructive and insightful, both from a purely chess point of view and on occasion from a psychological perspective as well.
In today's games, top players are almost always playing against other top players, and the games are very difficult to understand - despite the illusion of understanding supplied by having Komodo, Stockfish and Houdini at our side. The level of resistance is high, the opening play is subtle and refined and generally limited to a relatively small number of lines. In Tartakower's day, the top players faced mixed opposition far more frequently, leading to livelier and more accessible play, while opening experimentation was far more prevalent at the top than it is today. And even by the standards of the day - or really, any day - Tartakower stood out by his penchant for experimentation. The book gets off to a flying start with a Wing Gambit Sicilian in game 1 and the truly eccentric 1.e4 e5 2.Nf3 Nc6 3.Nc3 f5 (played by Tartakower and not some simul opponent) in game 2. Not all the openings are quite this odd, but there's plenty of variety and something for all tastes.
You can find an excerpt from the book here.