Indeed. Given that this research showed these particular studied effects were specific to particular muscles that were being exercised, it seems unlikely to transfer through entirely different set of cells and into offspring.

In fact, my uneducated first guess would be that the different methylation patterns are related to the actual buildup in strength, which itself would require some way of coding for the expression of certain genes differently in one set of muscle cells than in another.

But with indications that it happens

http://www.sciencedirect.com/science/article/pii/S0092867414...