Chess engines

Since the advent of the silicon era, computers have been programmed to play chess. Paul Shannon once wrote a paper on how to program a computer in this manner and he said that the number of possible positions in an average 40 move game was of the order of 10 to the 120th. This is now called Shannon's number. Even then, engines calculate the best moves as if it were child's play. Earlier on, their programming was buggy, so they could only calculate moves in forcing positions or where there were captures. They couldn't strategise at all and could not make long term plans. Only in 1 minute or 3 minute games were they amazing. The first mainframe computer which could calculate more than a billion positions every second was Deep Blue, which was created by IBM as a part of a project called Deep Thought. IBM challenged Garry Kasparov to a match which the latter won in 1996. A newer and more improved version of Deep Blue challenged Kasparov again in 1997. This time he lost. In open positions, the machine was a monster and would rip off pieces. Recently, new machines like Stockfish, Houdini, and Komodo have come out. Tournaments where these electronic beasts face off against each other have also been arranged. I think that engines, even though tactically unparalleled, are strategically vulnerable and may not anticipate exchange sacks, out of the blue moves, etc. Engines also enter into infinite loops, where a drawn position is seen.