But have you considered

Show

actually if you consider the hypothetical world in which the stakes are, indeed, increasing monotonically and perhaps exponentially over time, and superimposed some pseudo-random wavelets to represent sentiment perturbations from the baseline increasing stakes function, then you might get a curve that looks very much like this. arguably, from a kind of crowd-sourced bayesian perspective, this ngram is evidence in favor of constantly increasing stakes.

Stakes haven't been higher so far.