Well, there's an additional tax on the negative bet: if you lose, you lose $100k. If you win, you get no money.
The first question for this bet is "How much would I pay to make Sam donate $100k to a charity of my choosing." Would you pay $10k to make that happen? $1k? $50k? $100k?
If Sam donating $100k to charity is only worth $20k to you, and you think that Sam is 80% likely to lose his bet, then the bet has an expected -$4k utility to you.
EDIT: Sorry, my model above does assume that paying $100k to a charity of Sam's choice has 0 utility to you. If, for example, you strongly expect to donate $100k+ to charity in 2020 anyway and aren't very choosy about which charity, then obviously the bet is a pretty sure winner to you.
The first question for this bet is "How much would I pay to make Sam donate $100k to a charity of my choosing." Would you pay $10k to make that happen? $1k? $50k? $100k?
If Sam donating $100k to charity is only worth $20k to you, and you think that Sam is 80% likely to lose his bet, then the bet has an expected -$4k utility to you.
EDIT: Sorry, my model above does assume that paying $100k to a charity of Sam's choice has 0 utility to you. If, for example, you strongly expect to donate $100k+ to charity in 2020 anyway and aren't very choosy about which charity, then obviously the bet is a pretty sure winner to you.