Stock market cap is what you would have to pay to acquire the business but it is inherently unreliable as a guide to value because it is reflective of uncommercial factors such as sentiment. The grossest recent example of this was the dotcom bubble, in which robust companies' shares collapsed while dotcoms' prices rocketed despite the fact that they didn't make any money and never would. The reasons for this, and how it unwound, are complex, but it make it clear that stock market cap reflects price rather than value.
You don't consider only one possible outcome either; you look at everything that may affect value and form a view based on what you think will happen. Other valuation methodologies do not lend themselves to this as well. Thus NPV is the most reliable in that you can at least consider what you think will happen rather than either shrug or guess.
In your example you wouldn't value the business at a mix of X or Y because that creates a third valuation which unlike the other two is unsupported by any conceivable outcome. You'd go for X or Y and assign a probability to their happening; and you'd then consider what the implications are for the business - it may be that the deal is either still good or still bad.