But there is a difference between those examples. HT is such an effective counter to Minion that Minion becomes counter-productive. That is not the case with Cellar. Sure, in the few games where Warehouse, Cellar and Moat are available you may buy a moat but not a cellar. But that's not nearly the same as conceding Moat to now all of a sudden be the better card. I stand by my (usefulness of function)x(effectiveness at function) metric.
All I'm saying, and I will now be done saying it, is that you ought to decide whether to rank by price or by function and be consistent. I find the rank by function more meaningful, but enjoy the rank by price as well.
I'm sorry you thought I misunderstood your post. What I was trying to point out was that the two situations are actually very similar. As you noted: Minion goes from effective to counter-productive in the presence of Horse Traders. Does this effect the "absolute effectiveness" of Minion? No, it effects it's value only when Horse Traders is in the supply. Additionally, as you noted Minion becomes a bad buy, so in other words Minion's strength is lessened because you are less likely to purchase it. The important thing is that Minion's effect hasn't changed so if it has some "strength" property which is independent of likeliness to purchase, then it actually hasn't moved. If Horse Traders doesn't improve the effectiveness of Laboratory when you purchase it, then it's strength hasn't changed either. So, Minion and Laboratory must compare the same way...but you argued (In my opinion correctly) that Horse Traders actually makes Minion a weaker card. It does this because there is a situation where you would otherwise buy Minion, that you would change your mind because of the presence of another card. (It doesn't matter that the effectiveness of Minion changed on this board...that would only matter if you purchased Minion anyways). This is exactly the situation of Cellar and Warehouse, you are less likely to purchase Cellar if Warehouse is available...why? Because you would rather purchase Warehouse. So, basically it all comes down to how you rank the cards. The argument here is that there are 2 factors that contribute to the ranking:
1. Likelihood that you will want to gain the card. (aggregated over all possible sets of cards)
2. Effectiveness of the card when you would gain it. (aggregated over all possible sets of cards)
The two factors are related, both are effected by other cards which are available. No single card will have great impact because they are aggregated. Arguably there are many factors which contribute to Cellar being weaker, and Warehouse may be more useful for pointing that out. But it is a fact that some proportion of sets of cards will see both Warehouse and Cellar, and in those sets the likelihood that you gain Cellar goes down. So Cellar's value has gone down, relative to Moat (which isn't weakened by the presence of Warehouse).
You mentioned another interesting thing, which I'm going to run with a bit:
suppose there is a set with Cellar, and Moat. You have $2, and Cellar and Moat are both determined to be equally useful in this set, and you want both (and actually useful...though no such set may exist currently, such a set could conceptually exist). Additionally, suppose that one of the cards is unimportant to this situation existing. You'll get the one which is more useful now, and get the other one later. Now suppose you have the same set, replacing the unimportant card with Warehouse. You have $2, I would say you get Moat, almost no question. Sometime later you'll probably have $3 to pick up Warehouse, so you totally ignore Cellar. Warehouse has actually in fact made Moat better than Cellar in this situation. The relative value of the 2 cards has changed due to the presence of Warehouse.
It's marginal, yes...maybe not worth mentioning. But if Bridge makes Native Village better, then Warehouse makes Cellar worse. There are lots of other factors that come to play, but this is a simple easy-to-understand factor. Card combos are frequently discussed, even though those 2 cards may never show up together. Card counters are discussed with the same caveat. IF these discussions have any merit, then one card superseding another has the same merit, and serves a lot of the same purpose in these ranking discussions.
Man I'm wordy, sorry about that.