BTW, in the video above Kasparov is waaaaaaaaaaaay off in his estimate of the number of possible legal chess games. The actual number is colossal, ~10^80. That is waaaaaaaaay more than there are atoms in the entire [visible] universe. The oriental game of Go completely dwarfs chess in the number of possible legal games, estimated at a number so enourmous, it leaves the realm of human experience, ~10^800. Of course, the key to any game of skill is that only a tiny tiny fraction of those possible games are interesting. This gets me thinking on a different note, on information theory, and one wonders the amount of information that say a grandmaster has to compress (technique, patterns, theory etc) in order to achieve their level of mastery. I think it is estimated at about 25,000 for weak grandmasters, and as much as 100,000 for world champions. No wonder chess is hard. On the other hand, Go I believe requires far less than this, even though the possible legal games is vastly larger. I have always marveled that there are 1.5 billion Chinese, and not a single one has ever entered the top five chess players in the world.
People that aren't trained in science look at for example, 10^10 and 10^11 and think those two numbers are pretty close to each other, when in fact one is gigantically bigger than the other. Exponential notation hides this on purpose.