A discussion post I recently started got me thinking about chess computers. We all know that off-the-shelf software on relatively standard equipment can beat the worlds strongest players with ease. But is correspondance chess the same? If, say, a computer running Rybka was pitted against a top-ten GM, with three days to calculate a move (and therefore, presumably, have a chance at the brute force calculation that computers excel at) using the resources available to a CC player (opening books/databases etc), who would win? A human with 1.5 hour time controls could just not compete. I'd like to see who came out on top.

I'm fairly confident in my answer that the result would be a draw. Clearly, the computer isn't going to lose, and with so much time a top-ten GM will be able to calculate all variations of importance. Would anyone like to disagree? 

The reason most freestyle CC games are turning into draws is because it ends up being Rybka 3 vs Rybka 3 when both sides use computers.

If your talking about a GM not using a computer against someone who gets to use one, I'm fairly confident that the computer would win, if the player using the computer is competent enough he could just let it run for days.

The difference between a otb GM and a CC GM is the later will have "centaur skills" such as backsliding the engine evals what Arno Nickel calls digging for gold. It still takes a human's judgment to steer it. I think one of the most important things about computer assisted CC chess is the test on the theory that it gives. These types of encounters are at a higher level for all facets of the game than your run of the mill GM vs GM otb game. Your not going to see the gross blunders or not even the positional errors because its all underpinned by deep computer calculation.

I would say a human GM would have a significant advantage, however, I'd also say the chance of converting that advantage into a win is small.

A simple proof would be the following:

Player X allways seeing 1 more ply than Player Y

Y sees 1 ply ahead, which means X sees 2 -- this is a 100% increase and moreover, there are many tactical sequences with in chess with a depth of 2 ply. Thus X has significant advantage and will win many games. 

Now suppose that Y sees 10 ply, and X 11ply. -- Cleary X has an advantage, but this time it is much smaller mathmatically (100% improvement compared to 10%) but also that there are not as many 11 ply deep tactical sequences as there are sequences of 2ply depth . thus, X will still win more games than Y, but the winning percentage will not be as high when compared with the above case.

For those of you that think that this conclusion is obvoius, consider Y seeing 20ply ahead, and X seeing 40ply. We have the same 100% increase as in the first case, but I posit that the winning average will be lower, simply because tactical sequences of 2ply are very common, whereas, tactical sequence of 40ply are much rarer.

This is what I think of corr chess vs computers ... the human player is able to see further than the computer, but, he is let down by the very nature of chess -- very long tactical sequences (where the correct first move can only be discovered after a search of 40 ply) are rare.  

Consider naughts and crosses -- a player might be able to see 6 moves ahead, but, due to the nature of the game, another player that can only see 3 moves ahead can still do enough to secure a draw.  Since, to see 3 moves ahead, is sufficient enough to avoid making a bad move.

I would say it would be a draw to be honest,

the computer gets 3 days to calculate.. and so does the human, when they both play like geniouses, no1 comes out on top.