Simple: You mix the colors one way, paint for a while, go off to some other part, mix the first color again, and paint again. If you can't tell the differences between two colors, it's easy to come up with two similar/'identical' colors that are really quite different to someone with a different color perception.

While this article doesn't remotely prove that this is what's happened, it seems like a possibility to me.