Model probabilities Taylor Fritz 66.8% · Christopher O'Connell 33.2% Why Taylor Fritz is favoured Taylor Fritz's Elo (1752) is +233 over Christopher O'Connell (1519); low confidence. But — counter-arguments Elo confidence is low (5%) — ratings are based on a thin sample, so the prediction itself is wider than the headline number suggests. Calibration Raw model: 86.4% home · Platt-calibrated down to 66.8% based on historical reliability across this sport's graded matches. Confidence note Elo data on these teams is thin (5% sample-size confidence); headline number is sharper than the underlying signal — treat it as directional, not absolute.