One weakness with Gordon Clark is his lopsided fixation on epistemology. Perhaps his neglect of metaphysics is due in part to his radical skepticism regarding sense knowledge.
In that regard, Van Til is more balanced. Van Til shares Clark's interest in epistemology, but that's complemented by an interest in metaphysics. Indeed, Van Til attempts to use metaphysics to ground epistemology: the metaphysics of epistemology.
Take his view of God as the Absolute (i.e. absolute personality). That's a metaphysical principle with some epistemological implications.
Likewise, his belief that the one-over-many problem is harmonized in the Trinity. Admittedly, that's not a conceptual solution. Indeed, he thinks the Trinity is paradoxical. But by the same token, he thinks the Trinity proves the compatibility and equal ultimacy of the one and the many.
Ironically, Greg Bahsen share's Gordon Clark's overemphasis on epistemology to the neglect of metaphysics. However, Vern Poythress (chance, coincidence, randomness, risk, probability, prediction) as well as James Anderson and Greg Welty (theistic conceptual realism) have been developing transcendental theism in ways that explore and exploit the theistic (indeed, Calvinistic) metaphysical underpinnings of epistemology. In that respect they are doing what Van Til failed to do, which is to turn his programatic suggestions into detailed models.