The straight calculated (IE: no DR applied to the equation, and all values weighted equally (IE: all used as multiplicative modifiers)) resists for both scenarios are very close to observed numbers.

For A it's a 42.6% calculated resist versus a 43.5% observed resist. 1 - (.65*.82*.8)

For B it's a 45.5% calculated resist versus a 48% observed resist. 1 - (.65*.82*.9*.95)

Which seems to lend credence to there being both a DR of some sort and resist modifiers being weighted equally (IE: none flatly applied outside a multiplicative scenario). Obviously there could also be something crazy going on in the equation that just mimics what I think I'm seeing, and so I'd want to see considerably more data points before I jumped to anything resembling a conclusion however.