The run expectancy matrices that I have seen have slightly different numbers than Law's, but that depends on what sample of data you're looking at. I'm good with Law's numbers as well.
Out of curiosity, what does Law say about batting order? Is there anything different than what the previous studies have found?