Question about SQLT (specifically, sqltcompare) - how is actual sreadtim calculated?

Good day,

We have been diagnosing a production performance issue whereby the optimizer generates a poor plan even though better plans are available (eg, via hints, sql profile, etc). In a cloned non-prod database, I have discovered that even though the system and object stats are identical, sqltcompare reports that the "Actual SREADTIM" is a bit higher than the value stored in SYS.AUX_STATS$ on the production server (the "Actual SREADTIM" is pretty close to the value in SYS.AUX_STATS$ in the Dev server). As a tangent, the difference makes sense because we recently upgrade the back-end SAN in non-prod, and I would expect