I highly recommend reading up on the genetics of ball pythons. A good starting page is at NERD's website: http://www.newenglandreptile.com/gen...recessive.html.
But to clear up a few points mad in Lokua's post:
Actually, it's a little more complicated than that. When you breed a heterozygous to a heterozygous each egg has a 25% chance of being homozygous albino (dom, or dominant, is an entirely different conversation). So often times, the odds don't always add up. There are times when you may end up with 75% of your offspring being homozygous, and there are times when you may end up with 0% of your offspring being homozygous. It's incredibly variable... the statistical values only give us what "we could expect if the world were perfect"then 1/4 will be dom trait albino.
That's not quite correct. If you breed het to het (of the same trait) each offspring has a 25% possibility of being visiual (homozygous) albino, 50% possibility of being heterozygous albino, and a 25% chance of being entirely normal.they can however pass it down as a dom trait (i belive) to their offspring so 3rd gen some will be albinos.
So the only way any of the third generation are coming out albino, is if the breeing pair of second generation carry the albino allele.
I'm not aware of anyone producing a true De novo mutation in ball pythons. And the likelyhood of producing a de novo albino borders on the extremes of unlikelyhood.the only other possible solution i could see coming from 2 normal balls with no recessive albino traits is if a natural mutation actualy happened which is not a frequent occourance