I've found more and more that it's so hard to judge a lot of cards depending on how often they shine and how good they are under those conditions. For example consider the following card:
Super Village:
Action - $3
At the start of the game, roll two dice. If you roll two 6's, this card reads "+3 cards, +2 actions" for the rest of the game. Otherwise, it reads "+1 action"
Now this is obviously quite contrived, but one could argue it as being both the best, or the worst $3 quite legitimately. 1/36th of the time, it's totally bonkers, but the rest, it's worse than a $2. How much do we value that ~3% chance of dominance* when it's so bad otherwise? Some people would probably see it's dominance in those games, and give it a high rank because when it's good, it's so ludicrously good. Others would probably see that it's terrible 97% of the time, and give it a low rank. The truth is probably somewhere in the middle, but it's extremely hard to judge how to rate such a thing.
The issue is that, while this is a lot more extreme than what's really going on here, Ironworks is similar. It's dominating on Silk Road and Gardens boards, very good with Bridge, Highway and Vineyards, occasionally still good but otherwise, is very bad. There are other examples as well - Coppersmith is occasionally amazing, in a draw-your-deck engine without trashing, it's a terminal +$7, but usually, it's a card to avoid.
The natural human tendancy, I feel, is to swing in one direction or the other.
*I am assuming that when the double 6's are rolled it's dominant, but even then I guess there are boards when it wouldn't be.