etc.
It doesn't matter that you only get it once. Again, consider IGG. Without the on-gain, it would cost far less.
Well, that's a good point. But, I think a $5 IGG and a $2+$3 stonemason don't exactly compare. That is, I don't think IGG would be the same as $2 for IGG-curse effect, you may overpay by $3 to give each player a curse. Either the $2 base would be useless since everyone would just pay $5, or it would actually make it too powerful, because then it becomes much easier to rush the IGG pile in the cases when curses run out before IGGs, letting you three pile earlier. Or it would break it the strategy and make it not a good card, because the defender could buy a couple for $2 and prevent you from emptying the curse pile.
I'm think in general a $X+(N-X)overpay card is not the same as an $N card. Aside from all the cards that care about costs (from which there are many), a $5 card should in general be (slightly) more valuable than spending $X+(5-X) on an $X card, because there should be some opportunity cost for that extra utility. A Stonemason and two $3 cards is usually not as good as a power $5 attacker, though of course there are some cases where those options turn out to be better. It is very situational, of course, but we have a general principle of convexity of use vs. cost. That Stonemason has this extra ability makes it, in my view, a valuable $2 card. Kind of like Fool's Gold is a valuable $2 card, but hey you can't just spend $2 and grab one. You need to be buying all of them, ideally with +buys. It's almost as if an effective overpay is built into the mechanics of Fool's Gold.
And anyway, if Stonemason cost a fixed $5 it wouldn't be a great card. Good in some cases, but not as useful. It's the variability that gives it its value. And its cost of $2 is tied into its balance; it would be much different if it cost $1 or $3 or $4.
Stonemason's on-play effect is about as valuable as IGG's on-play effect. Both are niche and usually weak. If IGG is worth comparing at $5 (and it is), then it's also worth looking at the value of a $5 SM with the on-buy effect of gaining two $3s. And hey, you say that SM+two $3s is not as good as power $5 attack... that is not obvious. Not at all. That is an interesting thing to discuss, to compare. That is a reason why it would be interesting to rank SM at $5 with other $5 cards.
But let's be even more clear about this. Let's look at Masterpiece again. This is a card that is obviously useless at its base cost. It is a $3 Copper. Why would you ever get this over Silver? Why would you ever get this over a real Copper, which costs less and has interaction with cards like Apothecary, Counting House and Coppersmith? There are the usual name-caring edge cases (Menagerie, Fairgrounds, etc.) but those apply to everything. No, there is no way to argue otherwise -- you would pretty much never buy a base cost Masterpiece. Not even if it cost $1. Given the popularity of Copper, you would rarely buy it even if it were free.
Masterpiece's function and power is ENTIRELY in its overpay effect. That's it. That's all. That's what should be ranked. Who is going to buy Masterpiece with only $3? Nobody. Who is going to use a gainer on Masterpiece? Nobody. And, let's face it, it's also terrible at $4 -- with rare exception, you would prefer to buy Silver and not have to take this Copper junk. In the end, you are almost always going to be buying Masterpiece for a minimum of $5. That's why I believe that Masterpiece should be put with the $5 cards, or maybe even $6+.
With Doctor and Herald, I don't know. I expect Herald's on-play to be the main attraction, and Doctor is sort of mixed. But the overypay is the star for Stonemason and
especially Masterpiece.
So yeah, I would put Masterpiece with the $5 cards, maybe $6+.
I would put Stonemason with the $6+ as well.
Doctor I have no idea.
Herald probably $4 as I expect that the overpay would be used rarely.
And while we're at it, I now think I would move Peddler to the $2 category (that's also where PH is, right?) because that is usually the price you pay for it; when ranking it, one would account for the fact that you have to jump through hoops to get it at that reduced price, as well as the fact that TfB performs much better with it than with the other cheap cards in that tier.
I recognize that ranking by base cost is a fine option. It is the simplest method. It means people don't have to take into account that TfB will be less effective with it than other cards in the category. In exchange, they have to account for the overpay. I, on the other hand, think it would be easier, more appropriate and altogether more useful to compare overpay cards at the total cost they are likely to be paid for. If I rank Masterpiece against $5 cards, I would account for the fact that TfB is less effective with it because of its nominal $3 price. I find that far easier to factor in that it would be to rank Masterpiece against $3 cards and try to account for all the other prices I might pay for it. But like, that's just my opinion.