Annals of Mathematics referees accepted Thomas Hales's Kepler conjecture proof with a disclaimer stating 99 percent confidence in its correctness