Rog's — for three reasons.
Sample size. 166 rounds vs 53. 3x the data means individual outlier rounds have less leverage. Rog's metrics are more stable estimates of actual skill.
Course diversity. 76 unique courses vs 32. Rog's averages span a wider range of slope and rating, so his numbers are less context-dependent. Chuck's 9.9 handicap may partly reflect course selection, not pure skill.
Data hygiene. Chuck had 47 misdated imports excluded — nearly half his raw dataset. Even after cleanup, that signals a less disciplined tracking process. Rog's data had one anomaly (2026 GIR jump), but that's one cell, not a systemic pattern.
Counterpoint worth flagging: the question isn't whose data is cleaner, it's whose data reflects their ability. Chuck's two 78s at Troon North and Gamble Sands are legitimate signals of upper-tier capability — the issue is whether he plays to that capability consistently. His numbers may underestimate his ceiling while overestimating his floor.
Bottom line: Trust Rog's numbers as measurement. Trust Chuck's best rounds as capability.