Much, much smaller than 1/(2^48). You don't get whole chromosomes from each parent. Suppose one of your father's chromosome pairs is (ABCDEF,abcdef) and your m

Much, much smaller than 1/(2^48). You don't get whole chromosomes from each parent. Suppose one of your father's chromosome pairs is (ABCDEF,abcdef) and your mother's corresponding pair is (PQRSTU,pqrstu). Then you might be, e.g., (ABcdef,pQrSTu). Except that actually there are hugely many points at which crossover (i.e., a switch between ABCDEF and abcdef) can occur, and it can happen a variable (though usually not very large) number of times on each chromosome. Oh, and after that you'll have maybe 100 mutations (which could be anywhere) relative to your parents. Crude back of envelope calculation: you have on the order of 10^8 base pairs on each chromosome; there might typically be two crossovers per chromosome pair from each parent, giving you on the order of 10^16 possibilities for the chromosome you get from your father, the same for your mother, and hence ~ 10^32 possibilities for each of your chromosome pairs, so ~ 10^700 possibilities in all. Then you have ~100 extra mutations in your ~10^9 base pairs, giving another 10^900 possibilities; so there are ~ 10^1600 ways to make a typical human child's genome from a given pair of parents. That's a super-crude calculation, of course, but it's much nearer the right order of magnitude 2^48.