You can't argue that the mere existence of risk makes a bet not worth taking, because it's already handled by the lopsided ratio of wagers.

Actually, you can. Sort of. At least, your reasoning is not quite right, because the *value* of money to an individual may not necessarily be a straight linear function. E.g. if it's your last $5 to your name and you haven't eaten in a week, then losing that $5 'costs' a lot more than you would 'gain' by winning another $5.

Betting based on straight odds assumes a simple linear value/utility function for money.

Also, there's a good argument that log odds is the better scale to judge risk, rather than straight odds, based on the idea/theory of martingales (https://en.wikipedia.org/wiki/Martingale_%28probability_theory%29). Under this concept, the ideal betting strategy is to bet on all the possibilities, in proportion to the expected *logarithm* of the value of the bet. I may be mutilating the terminology, but the idea is straightforward when one actually applies the theory. See https://en.wikipedia.org/wiki/Gambling_and_information_theory for some intro and further links.