Well, this genius got it wrong, lol.
At first glance, the right hand side being heavier seemed like the obvious answer, didn't it? Still, I carefully weighed up everything and it looked to me that the left side should be heavier by the weight of the ping pong ball since it's attached to the container and since the ball bearing is suspended, hence shouldn't be putting any force on the scales. Clearly that's wrong, but his explanation didn't quite make sense to me even after hearing it a few times.
Thinking about it now however, the ball bearing is displacing the weight of water equal to that of the ball bearing's volume and hence there's an upward boyant force on the ball bearing that must in turn push down on the scale by the same amount. So, since that's much heavier than the ping pong ball, that side goes down. Note that the ping pong ball would still be lighter than the water it displaced and hence would still keep the string taught and the boyant force that this generates is irrelevant as it's tied to the container, which cancels it out.
One could engineer an experiment where the ping pong ball was sufficiently large that it's now heavier than the water displaced on the other side and hence the left side would drop.
I got the other two right though.