In its current form item comparison makes no sense considering that there is masterworking available.
Both items shall be compared to its basic statistics not to the statistics after the masterworking has been applied.
This should be obvious but somehow it is apparently not, so below an explanation.
After masterworking the items affix power increases. It means that at some point even if you drop a new item, lets say close to 100% of their affix power and you compare it to the masterworked one it will be weaker, although after reaching the same masterwork level it will become much stronger than the one equipped.
There is an easy solution that I see which would not change anything and allow players to choose to what would they like to compare. This solution would be to leave the system as it is and add a “comparison view” in the inventory, where after clicking a button all the items in the inventory are temporarily shown in their base form not considering the masterworking buffs. This would allow to either compare to the basic stats or to masterworked stats.
There is also an issue with masterworking itself. Each few steps one of the affixes gets an extra 25% boost, but it makes no sense for the affixes which add skill points. Lets take Fractured Winterglass as an example. One of its affixes is “+x to Conjuration Mastery” [where x can be in range 1 - 2]. This means that if you have +1 and this affix gets buffed it is still +1, so you have basically lost 25% which could successfully boost one of the other affixes and actually work.
BTW. The fact, that there is no way to save builds with its skill trees in wardrobe to easily change in between the builds depending on what scenario a player wants to play is laughable.
It is radiculous that until now such trivial issues are not working properly. Are there no engineers in the development team?