Can't remember if they named amount of metals used - something like grams? 1 gram of Cs is able to produce 84 mL of H2 (at STP). Not much.
Anybody here knows how amount of hydrogen translates into kaboom size?
But that's not the only problem. I wonder if the reaction speed really depends on the reactivity here? Perhaps transport of water to the metal surface becomes limiting factor, as metal is covered with the cushion made of hydrogen; it have to move away to make place for more water, this requires time.
I think that explosion power in such a system should be a (partially) random thing and I doubt in the easily observable dependence between power and reactivity. Amount of energy should depend on the amount of hydrogen released but left close to the ignition point. That i turns depend on how long reaction proceeds - and I don't think one can predict how fast hydrogen will get ignited. With all that fizzling and splashing speed at which metal gets hot is not constant, so sometimes ignition point is reached sooner, sometimes later - and that in turns means different explosion size.