And on this topic, I'm a big fan of the way Cribb put it about 4 or 5 iterations of this discussion ago.
He basically argued that, as you go up to a hypothetical next level against the Martians, the difference between two players in their secondary skills converge. For example, Shane Warne averaged 17 with the bat, and Murali about 11. Presumably, stepping up a level, the difference in the batting would be completely negligible; picking someone slightly better in their secondary skill doesn't help a whole lot.
Cribb used the example of Glenn McGrath and Stuart Broad. If you were playing park cricket, you'd take Broad every time -- McGrath may be a better bowler, but at that level the difference would be negligible; the level of the batting means they'd both dominate irrespective. Meanwhile Broad's batting would be of use compared to McGrath's, because McGrath was such a shunt with the willow.
Meanwhile if you were playing against the Martians, you'd take McGrath without thinking. The difference in their bowling ability suddenly becomes huge when you're pitting them against a higher level of batting. Meanwhile against a higher standard of bowling, Broad would become closer to absolute uselessness and McGrath would shunt it equally hard because he genuinely couldn't bat at either level. Diminishing returns and whatnot.
Perhaps not quite as applicable to the Pollock/McGrath debate because their bowling averages are so similar, but framing Pollock as providing 25 more runs per innings in an ATG scenario is probably rather misleading; the difference wouldn't be quite that pronounced.