The vendor quote is a bit much, but I agree that a convex tamper piston does seem
to reduce the likelihood of channeling, assuming an otherwise correct tamp. Maybe. FYI, the poll Flat or convex?
shows an overall preference for convex (28% vs. 53%).
On a related note, I've recently been experimenting with "nutating motion" while tamping, something I privately derided as silly. And yet, for some espresso machines, it does seem
to reduce the likelihood of channeling. That said, the difference between flat/convex/nutation, if it indeed exists, is quite small.