This bet is worth +€25 in expectation: played long enough, it pays. Most people refuse it anyway, and the gain has to reach about twice the loss before the bet feels acceptable. Tversky and Kahneman eventually put a number on it: a loss counts 2.25 times more than a gain of the same size (Tversky & Kahneman, 1992).
Tversky, A., & Kahneman, D. (1992). Advances in Prospect Theory: Cumulative Representation of Uncertainty. Journal of Risk and Uncertainty, 5(4), 297-323.