Bet with NBA players 
Giannis of Milwaukee Bucks and  Kyle Lowry of Toronto Raptors are having lunch together at a restaurant. Giannis claims that if their teams play against each other, Raptors will lose with probability 3 out of 5, while Kyle claims that Raptors will win with probability 4 out of 7. 
  The waiter, who overhears the conversation, decides to make some money himself, by challenging one of the two, or even both, to go on a bet with him. He wants to make some money regardless of the outcome of the game. He is thinking of suggesting to Giannis of Kyle (or both) that, if his team wins, the waiter will have to pay him A dollars, while if his team loses, he (or they – if they bet together) will have to pay him B dollars. Obviously the bet must be favorable for Giannis or / and Kyle, in order for them to accept it and, if the waiter proposes a bet to both, it must be with the same amount for each of them. 
What will be the amounts A and B?

I wish I could show you some work but I haven't been able to do anything! I even gave it to two friends of mine, who are mathematicians and they haven't come back with any solution yet!
I assume we must compare the two probabilities for Raptors to win (one says 2 of 5 and the other 4 of 7). But I don't know how to proceed.
 A: Assume that $G$ and $L$ are willing to trade on their exact probabilities (this is unrealistic, since both should demand a positive expectation bet).
Then you can place a $\$100$ bet with $G$ on the Bucks using $G's$ odds (of course you bet that the Bucks lose, $G$ bets that they win).  Since $G$ sees a $\frac 35$ chance of winning, $G$ sees that as worth $\frac 35\times 100=60$.  Since $G$ sees a $\frac 25$ chance of losing, he must offer $X$ such that $\frac 25\times X = 60$ so he offers you $\$150$ should you win.
You also place a $\$100$ bet with $L$ on the Raptors using $L's$ odds (of course you bet that the Raptors lose, $L$ bets that they win).  Since $L$ sees a $\frac 47$ chance of winning he sees that $100$ as worth $\frac 47\times 100$.  To be fair, he then must offer to pay out $\frac 73\times \frac 47\times 100 =\frac 43\times 100=133.\overline 3$.
Payouts:  If the Bucks win, you receive $133.\overline 3$ from $L$ and hand $100$ to $G$, giving you a profit of $33.\overline 3$.
If the Raptors win you receive a payout of $150$ from $G$ and you hand $100$ to $L$, giving you a profit of $50$.
Either way you win (and you can afford to sweeten the odds a bit to induce the players to bet with you).  
Note: nothing above made any reference to the "true" probability for the winner in the game.  The only way that matters is if you want to compute the value of this double bet to you.  As it stands all we know is that this position is worth somehwere betweeen $33.\overline 3 $ and $50$.
