Now, we will not be too far off the mark if we think of a chessgame as being s story. The moves and tactics are the story's vocabulary and idioms. As someone who had to learn languages on-the-go, I can tell you that deep-thinking is an ineficient way of approaching foreign vocabulary -- what is needed is a repetition that brings about automatic command of the language. On its own, that repetition will not make you a spellbinding orator -- but without it, you have no chance. (With concentration, one will normally slowly recall a foreign word after some half-a-dozen uses. But after some 50-200 uses, that word will be at the tip of your tongue, whether you put hard work into remembering it, or just lazily let your brain to do its own thing and learn the patterns by whatever means it choses.)

The last point is that of combinatorics: it stands to reason that powerfull 2- to 4-move tactical bursts will repeat over and over, while deep, 12-move marvels you happen to encounter once in your lifetime. Moreover, that marvel is invariably composed of 3-4 of those small combinations blended together. So, if you know your vocabulary well, your skill as an orator will let you complete the game brilliantly.

<You simply cannot learn chess by memorizing a bunch of positions from a book>

In my experience, both chess and language are best learned through practice, not books. If you think learning languages is mostly memorization, it may be due to your personal learning strategy. For me, learning a language requires constant practice (both talking and reading) and reflection. There is a strong element of creativity in trying (and most often failing) to construct novel sentences in the language you are learning. Also the syntax may not be so clear unambiguous. There is also the issue of idiomatic usage - a sentence may be syntactically correct but no native speaker would use it. These are things that have to be acquired by immersion and a subtle feeling for nuances.

Brains of most of us seem to be wired so that they somehow arrive at a synthesis if given a theme-and-variations treatment. In turn, the sythesis is typically a mixture of intuition and conscious understanding -- <Halldor and Ziggurat> said it quite perceptively and elloquently, better than I would.

The question, I guess, is how hard should be our excercises. Whereas I see a great value in tackling very hard problems, I suspect many of us here naturally concentrate too much on the difficulty and not enough on quantity. Well, I do. I naturally yearn to be at the throat of as hard a problem as I can handle. As soon as I overcome it, I ask for a tougher one. But from the standpoint of training and learning, there is a great value in often scaling down to, say, 60-80% max-strenght level. Use the extra mental energy for solving more problems, and/or for free-form meta-analysis of the patterns.

Here is what I mean: At the end of a realy hard problem, I (hopefully) know the solution -- sac, check, takes, check, interpose, if he diverges with a then b -- an enumeration of variations. But after a realy taxing problem, that is often all I have. But at the and of a, say, 75%-max problem, I may have also figured out that the only reason that such particular promotion pattern works is because oponent's pawn was on a the adjacent file, and not one file over... And if we do not make such a small side-discovery? Well we still put a new solved problem into our theme-and-variations databank for the brain to do its own thing with it.

The deflection 26. Nxd8! appears to give White a decisive advantage:

26. Nxd8! Rxd8 (26...Qxd8 27. Be6 fxe4 28. Qxf8+ Qxf8 29. Rxc8 ) 27. Qxf5 Nb6 (27... h6 28. Qf7 Ne7 29. Rxd8+ Qxd8 30. Be6 Kh7 31. b4 Kh8 32. c4 bxc4 33. Bxc4 Nc8 34. b5 Nb6 35. Bd5 Kh7 36. g3 Qb8 37. Be6 Kh8 38. Bf5 Na4 39. g4 Qd8 40. h4 Qg8 41. Qg6 Nb6 42. Qxd6 ) 28. Qf8+ 1-0