Since we have about zero reviews now, I think that any kind of review will be acceptable, so long as it shows a significant attempt to grapple with the contents of the paper. I think a good review is something that shows how a calculation method can be extended, points out related literature that is not in the reference list, extends the paper with a nontrivial observation, to show importance, or gives a negative review with detailed explanation of what is wrong.
The quality will only improve once some sort of competition starts, but the first reviews will be setting a tone. But it shouldn't take ~4months as is sometimes standard on journals. Just my 2c. Agree with Vladimir too.