One Fortress Leads to Another

When I started the category on Posts with label Fortress, the goal was to create an article on the subject. I expected to find more examples than I have. The next position is from an early FIDE World Championship match.

1954 World Championship Match (game 3)
Botvinnik, Mikhail

Smyslov, Vasily
(After 41.Bd2-e3)
[FEN "3b4/p7/Pp2k1p1/1P1p1pp1/3P2p1/4B3/6KP/3B4 b - - 0 41"]

Botvinnik wrote,

The sealed move. The players agreed a draw without resuming play. Black does not have any problems. After the exchange of Bishops (for example, on the square e5), he only needs to exchange his five Pawns in the center and on the Kingside for the two Pawns on d4 and h2, in order to establish another fortress. The King will occupy an impregnable position on b8.

For the complete game, see...

Vasily Smyslov vs Mikhail Botvinnik
http://www.chessgames.com/perl/chessgame?gid=1032342

...on Chessgames.com. Wikipedia also has a page on the subject -- Fortress (chess) -- with many examples from theoretical endgames. The page says, 'Fortresses pose a problem for computer chess: computers are unable to reason about fortress-type positions except to the extent that their endgame tablebase allows.' This was true once, but Monte Carlo methods now allow chess software to detect fortress positions.