In reply to Jimmi:
It's totally dumb. The W3P2 rigging is used in some rescue work where very high loads are anticipated. The entire point is mitigate the strength loss in the sling material that comes from the knot efficiency. Because of the friction of the two pulled loops around the tree, very little load is actually applied to the knot, so the weakening effect of the knot is diminished (and, with enough friction, totally eliminated).
By putting an overhand knot in the two pulled strands, this genius introduces a knot that will be fully loaded, thereby negating the only purpose of the W3P2 protocol, which he clearly doesn't understand.
For any ordinary climbing application, the strength loss in something like 1" webbing due to the knot is of no consequence, and so things like W3P2 are pointless. Moreover, if a W3P2 installation is left in place and reused by many parties, the potential to strangle the bark and kill the tree is greater than for a simple sling loop.