Hypothetically, you can bet the Bucks-Spurs over 198.5 total points (With a no-vig line of -136) and over 99.5 points for the first half (no-vig line of -130), both at -110. How much of your bankroll would you allocate to each of these bets? I'd assume if one knew the correlation coefficient between the first half over and the game total this would be simpler. I believe that there would an advantage to be had in betting both of these, as it would slightly decrease the variance of your expected return.