Originally Posted by Spark
The quality of the attacks, the nature of the pitches, the timeframe across which the runs were scored (and hence how many "form windows" the sample size actually captures).
I mean otherwise if you're silly you end up with statements such as Samaraweera > Ponting, and, well...
Yeah. I think people confusing rating players on stats/records with just looking at their overall career average and making a call on that. You can still analysing someone on purely a performance-based level and come up with a different conclusion to the Tests column in his cricinfo stats bio.