when you have $5, what's a good thing to buy?
You know, maybe this is my problem with this whole ranking vs. rating system. I think this statement, from Donald X of all people, underscores a significant problem with gauging the usefulness of the system.
Ranking the cards by cost encourages this sort of thinking (what should I buy with $5) and this sort of thinking is fundamentally flawed.
I've watched people in my playgroup plunk down $5 on the table and then, one by one, stare at the $5-cost cards trying to figure out what to buy next. That's after having bought six or seven cards already. And time and again, I berate them: "Don't you have a plan?"
Sometimes (well, a LOT of the time, actually) the best buy for your deck with $5 is a $4-cost card, or even a $3-cost card. (Or that Herbalist you forgot to get earlier for the extra buy. Don't wait and pay $10 for it. Better yet, pay attention to the fact that it's the only +buy and open with it if you're engine-building.)
Cards simply shouldn't be compared to other cards at the same cost. Cards should be compared to other cards, PERIOD. A strong $5 on the board means that opening with at least one Silver might be more important than a strong $4 and a strong $3. When you look at a kingdom, the first thing you need to do is make note of the power cards, then look for special combos, then figure out what the kingdom wants to do (engine/rush/slog/other.) Don't wait until you have money on the table, then look at the most expensive cards you can buy and figure out which ones are better. If that's how you're making your decisions, you've already lost.
Now, you might say, yes, Dominion is a complex game and you can't just rate cards in a vacuum, so why do we even bother having ratings and rankings at all?
Because the cards really do have different power levels that are not immediately obvious to all players, even ones with some experience. Certain cards are very situational. Certain cards seem like they should be good, but no matter how many different combinations of strategies you throw at them, they just don't seem to get you there (I'm looking at you, Counting House.)
And for the record, I do think that split piles, castles and knights should all be rated separately, in that they have the same problem as every other Dominion card: They are only good when you take other cards into account, and they have a cost that can be measured in more than just coin. (Even a simple card like Village can measure its worth only in terms of the Terminals available in a Kingdom, and has a cost that can be measured in coin and opportunity.) Some castles are useful even if you're not trying to piledrive the castles. The bottom split pile cards should be downgraded somewhat if the top card isn't that great, in the same way that potion cards should be downgraded if potion is a prohibitive opportunity cost.
If the presentation of the data were user-selectable, allowing users to sort cards by cost, type, etc., and filter by a number of other parameters, then it would be possible to do things like look at a whole stack of knights or castles and see their ratings relative to each other. Maybe you're planning on playing a knight or two, but the top knight is one of the weaker ones; it might be better to allow your opponent to buy it so you can buy the second one. This is useful information. It might also be worth knowing that Bridge is such a colossally game-warping card in the right kingdom that wasting terminal space on any knight at all would be foolish.
But maybe I'm putting the cart before the horse.
Perhaps I should simply ask the question: Why rate the cards? What exactly are we trying to accomplish here, and who is the information for, and for what purpose? I may be misunderstanding the entire exercise.