> this relies on the idea that the purity numbers are based on "time of test" not "date of seizure"

No, the idea doesn't rely on "time of test" vs "date of seizure". There is no real provenance for any of this. There is no auditable trail for when any given batch of narcotics was manufactured, when it appeared in the US, how long it took to disseminate to domestic dealers, when it may have been further cut by domestic dealers, when it was sold, and when it was actually used. Even the seizure dates are dubious, given haphazard and inconsistent law enforcement handling and record keeping. There are also sampling biases, because some legal jurisdictions and law enforcement organizations are more or less cooperative than others.

All I claimed was that a delay was plausible. I am not obligated to become a narcotics market researcher in defense of my modest claim, and given the nature of all this, no amount of such effort is likely to be sufficient for you in any case.