Far away in the infinite-dimensional quantum space, there is an Everett branch that looks much like our own, except that Spice Merchant and Pixie have never been printed. In this world, Queta and Alice are having a conversation about Dominion cards.
Q: Hey! Wait a minute! Come over here! Thanks. I have this awesome idea for a Dominion card I want to discuss with you.
A: Hi. Sure. Shoot.
Q: Okay, so get this: it's a lab that can trash Treasures. What'cha think?
A: (thinking) ... hm. Well, trashing without hand-size reduction is extremely powerful. So it's strictly better than a Lab, which means it would obviously have to cost at least 6$-
Q: Ah, but there's a twist! You
have to trash Treasures, and it stops working if you can't.
A: You trash before or after drawing?
Q: Before. Does that make it weaker?
A: Marginally. Okay, in that case, it would probably be fine at 5$. Your card starts off very strong but gets weaker over time. However, early turns are more important than late turns, and you start out having seven Coppers, so it'll take a while. Probably a fairly strong 5$ ... (thinking) ... I guess the thing that worries me is, since it does get weaker, you'll always buy it as soon as possible if you do buy it, and because it consumes Coppers, you'll never buy two. So overall, it seems to only add a single decision to the game. That may not be-
Q: Actually, do you think it would be okay at 4$?
A: A Lab that trashes being okay at 4$? No.
Q: Suppose for some reason that it really wants to cost 4$. Can you steelman a case for it?
A: (thinking) ... well, I guess I'd say, it's important not to commit the fallacy of treating powerlevel as a terminal value in Dominion. The questions, "what is the cost such that the card is most similar in power to other cards at this cost", and "what is the cost that makes the game maximally fun if this card is added with that cost" are
correlated, but they're not inherently tied together. There are cases where they come apart, and if they do, it's important to remember that the first question doesn't matter at all. The clearest example of this is Chapel. The answer to the first question is probably 5$, but the answer to the second is probably either 2$ or 3$... I'm not sure. Anyway, the reason this is fine for Chapel and not Laboratory is because Chapel does a thing with diminishing returns, which means it doesn't prescribe a strategy itself but just transforms the game into something else. In that way, Chapel is arguably a good card for the game, and the fact that it's bonkers relative to its cost just means that powerlevel isn't a useful abstraction in this case.
A: Your card also has diminishing returns, so it can't ruin the game, regardless of how much it costs. At 4$, it would essentially be like having a Heirloom that trashes a Copper every time it's played, except that it's a card that only does that, rather than a card that does something and has that effect on top of it. If that's what you want... but why not just do that Heirloom instead?
Q: Suppose it were added to the game at 4$. Would there be any effect other than removing one choice in the opening and making your Coppers gradually disappear?
A: Well, sometimes Coppers are good, like if there's a Garden rush or whatever, and then you don't buy it. Those cases are rare. Other than that... (thinking) ... a 5/2 opening comes to mind, but I think you'll probably just have to buy it anyway.
Q: Hold on, you said earlier that it was just 'fairly strong' at 5$, so doesn't that mean many other 5$s are stronger?
A: If it costs 4$, then whether you have 5/2 or not, you would buy it in the opening. Your card is stronger the fewer non-Coppers there are in your deck, and trashing is generally stronger early, so your card has become mower powerful since I said that. I don't think there are many 5$s that can compete -- remember that almost every card that can trash without reducing your handsize is among the strongest cards in the game. You wouldn't buy it for 5$ if there is Recruiter on the board, but otherwise...
Q: And after the opening?
A: (thinking) ... well, it's kind of an absurd idea, but if it really cost 4$, you would probably buy two sometimes. It sounds silly, but trashing without hand-size reduction is just so very powerful, so even if you can only play them for a few turns, if there's Butcher on the board or something...
Q: Would it be worth buying Coppers once you're out?
A: (thinking) ... not usually. But sometimes, yes. And, I guess, every once in a blue moon, you could do something really neat with Banquet or Beggar or whatever.
Q: I think you could try harder to steelman the cost. What if there's a pure big money board? Trashing Coppers isn't important for Big Money, right?
A: Pure big money boards are like boards where it's correct to open Chariot Race. Frequently sighted, rarely encountered.
Q: Okay, but BM+X. That still happens sometimes, even after Menagerie.
A: Sometimes... well, if X is Butcher, I still want your Copper-trashing lab for 4$.
Q: What if X isn't trash-for-benefit?
A: (thinking) ... it's close. I'm not sure. May still depend on X.
Q: Cool, that sounds like sufficient reason to make it cost 4$.
A: Seriously no.
Q: Thanks for your feedback, I'll make it cost 4$.
Q: Actually, there's another part to this card. Instead of being a Lab, it can also be a Woodcutter for some reason. Isn't that cool? Do you think now it makes more sense at 4$ because people will sometimes choose a Woodcutter instead of the Lab, and that'll make it weaker?
Q: Hey, come back! Don't go! I still have another card to ask you about.
Q: So get this: it's a card that's sometimes like a Village but usually either like a Lab or like a double-Lab. And the coolest part: It also costs 4$. Isn't that super cool?
Q: Noo, don't go!
Please, come b
aaaaack