Innovation strategy
The landscape is the (hidden) quality of every possible idea. You pay one "test" per idea you evaluate.
The science (and what is simplified)
The innovation process is modelled as search on a fitness landscape q(x, y) ∈ [0, 1]: a sum of random Gaussian "good-idea" hills (heights differ, so there are false peaks) on a 2-D idea space. Each test costs one evaluation of q.
Brainstorming (divergent thinking) draws ideas uniformly at random. If q has empirical distribution F, the expected best of N draws has the closed form E[max] = ∑i vi(FiN − Fi−1N) over the sorted grid values vi; that is what the check compares with Monte Carlo. It grows slowly (diminishing returns).
Prototyping (convergent thinking) is stochastic hill climbing: perturb the best idea, keep it if better, shrink the step on failure. It converges fast but only to the nearest hill, so a single start usually settles on a local, not global, optimum. Diverge, then converge (the "double diamond") spends 25% of the budget sampling, then climbs from the top 3 samples. On rugged landscapes it typically beats both pure strategies; on smooth ones climbing alone wins. Simplified: real innovation has no known objective, evaluation noise, or fixed dimensionality; this is a metaphor with exact arithmetic, not a model of a company.