This question has nothing to do with "metaphysics", it is an effort to establish reality thru reason alone -- in contradiction to Kant's Critique of Pure Reason.
What the argument does specifically is try to misuse a property of theoretical objects, the property of infinity, and combine that with an inappropriate application of statistics, thru a process of limited imagination and confirmation bias.
Theoretical objects are infinite. One can postulate an infinite number of variations of any hypothesis. Therefore, the sum of one such set of hypotheses, cannot be "greater" than the sum of a different set that proposes a different nature to our world. One cannot count these hypotheses, as they are infinite.
One CAN try to identify ranges of plausible reality space, and then try to show that one range is wider than another, IF one has a probabalistic theory of what the world ought to look like. This is a different process than counting, though it does arrive at a similar probability calculation.
Physicists can do this in some cases, most notably with the application of quantum mechanics to cosmology. We think the values of the constants in the Standard Model of QM were established thru random draw, and we have an idea what the possible range of them were, primarily by comparing them to teach other, to conclude that the plausible universe space that can sustain any kind of life at all for variants of QM, is very tiny. This is because most possible universes disappear almost immediately, or quickly spread out to disperse matter to effective density of zero, or quickly collapse into nothing but inert black holes. This calculation supports that a life compatible universe appears to be WILDLY improbable, and is called the Fine Tuning problem. Trying to provide a possible answer to the Fine Tuning problem is why so many cosmologists postulate a multiverse. QM theory is well enough understood to do these sorts of speculative evaluation of probabilities.
All we have for afterlife are logic categories of ontology, we don't have a logic hypothesis that translates these categories into probability space, nor measurable phenomena that have been at least partially quantized by science theory. The limitations of what your debate partner can imagine are not anything like a probability space, other than a measurement of his inclinations to use his imagination to reinforce confirmation bias.
We can TRY to quantize this, at least partway, just by looking at logic categories. Afterlife requires spiritual dualism to be true, AND that spiritual entities can exist independently. We have four types of speculative ontologies: idealism, spiritism, materialism, neutral monisms, and dualisms and triplisms of the three main types. This is eight logic categories, only 4 of which even possibly allow spirit as a independent ontic state, so can even support an afterlife. As yes/no on independent state is another logic division that removes half of the dualisms and triplisms, assuming they are in the "no" category, so only about 1/4 of this logic space could support an afterlife. Trying to establish probabilities between these logic states is unjustified, one cannot go from logic space to probability space.
So -- contrary to your disputant, the possible logic states that support an afterlife are a minority of possible ontological logic states. But this does not actually say anything about the probability of one of those states actually occurring, as we don't have a probabalistic theory that relates logic states to possible realities, as we sort of do with QM.