Board Thread:Wiki discussions and announcements/@comment-1990160-20140521040036/@comment-1330314-20140522145058

The best way to explain it would be to imagine a graph with an upward sloping line: if you were to shift the graph up or down or left or right, it would not affect the slope of the line. Since you're shifting the gold value of every item with the same stat left or right, it's functionally the same as moving the graph for that gold value to the left or right. The numerical values may change, but the information you're getting does not, even if it looks different.

Which brings me to one of your main points in this discussion: there is absolutely no point in focusing on the "true" gold value of a stat, because it's not the specific numerical value that gives information regarding gold efficiency as a whole. Suppose a Rioter kindly joined the discussion and put up a table listing the official gold values of every stat in the game. Suppose that one stat, say, attack damage, is actually worth twice the value we assigned it. We would then bump up the gold value of attack damage across the board, which would then bump up the gold value of every AD item out there. Because every AD item would become more gold efficient using the same ratio for attack damage, any comparison that could have been made before would still give the same results after. If you're comparing two AD items on the value of their AD alone, bumping up the gold efficiency of AD would not impact on the comparison. Effectively, we could just multiply the value we assigned to every stat by 100 and we'd still get the same results. The advantage to the values we have is that they are based, one-for-one, on the price of the components, so the standard for gold efficiency is 100% instead of 10000% or some other value.