I think I spotted my mistake. My r_1 was slightly too small, even after taking the Earth diameter into account. When calculating with r_1 = 6711.94 km (which takes into account the radius of the Earth under where the spacecraft was at TLI) rather than 6569km, I get a required velocity of 10.804km/sec for the Hohman transfer rather than 10.923km/sec. 10.804km/sec is indeed smaller than the 10.834km/sec actually used. With r_1=6711.94km and v_1=10.834km/sec, we'd get to an apogee of ((v_1 * r_1)^2 / (2mu - v_1^2 * r_1)) = 566,380km, far greater than the semi-major axis of the Moon's orbit. It's amazing that by just reducing r_1 a tiny bit to my incorrect original r_1=6569km, then even by increasing v_1 to my v_1=10.923, the apogee radius still is only 383,050km. If I leave v_1 at 10.834km/sec (the correct figure) but keep r_1 at my incorrect r_1=6569km, I get an apogee of 193610km. So just picking an r_1 that's 2.1% too small, I get an apogee that's 66% smaller than it should be. Wow!