Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[tlloyd]

The problem with Theory's rationale is that he slips from comparing the $2s to comparing the cyclers/sifters. If we're doing comparisons among cards with similar effects (which have popped up recently and are probably more meaningful strategy-wise), then certainly Cellar should drop in relative value given the advent of Warehouse. But if we are ranking cards with a given price, it makes no sense for a new $3 to change the relative merits of the $2s. Of course $3 and $2 are not significantly different prices in most situations (I'm looking at you, Duchess-Rinkworks! ), but that just reinforces the fact that these lists are for fun and shouldn't be taken too seriously.

[Fabian]

FWIW I think the "Cellar is bad because Warehouse exists" argument is completely nonsensical. I've read theory's posts on it multiple times, and I just can't figure it out. Cellar is bad because Cellar is bad, not because Warehouse is better than Cellar. If a $5 treasure called Diamond was printed tomorrow that gave +3 coin but only appeared in 2% of matches, Gold wouldn't become a bad card. It would be pretty damn bad in almost all of those 2% of matches where Diamond appeared of course, but that's kinda obvious.

The idea that a card's strength would be measured by what another card, which will affect Cellar once every x hundred games (or however often two cards appear in the same game), does, kinda blows my mind.

[theory]

But Diamond wouldn't be balanced.  If you accept that every card has a role, and you're looking for the weakest cards (defined as the cards worst at fulfilling their roles), then it seems that if there's a significantly better card in the same role that isn't super strong, then that's pretty good evidence that the original card can't be average or good.

Thief is bad on its own, but the fact that Noble Brigand isn't overpowered is further evidence that Thief is bad, since if Thief is good then a superior version of Thief should logically be a little too strong.  Compare to Lab/Hunting Party: you can't use this argument because Hunting Party really is supremely good, so the fact that Lab is slightly worse than HP does not necessarily mean that Lab is bad.  Whereas Warehouse is not overpowered by any means, it's just average-to-good, so if Cellar is considerably worse than an average-to-good card that does exactly what Cellar does, well, that's pretty solid evidence that it's weak.

[Fabian]

Well I guess we're just arguing in circles now. I think your way of evaluating a card makes no sense. Especially so when you introduce arbitrary "well this comparison doesn't count because that card would be really good and Warehouse is only kinda-good-but-not-better-than-top5-of-$3-costing-cards" type stuff. I don't accept that "every card has a role, and I'm looking for the weakest cards defined as the cards worst at fulfilling their roles". I'm going with "I'm looking for the weakest cards". Again, Cellar is bad because Cellar is bad, not because Warehouse is good. Gold is good because Gold is good, not because Diamond would be better so that makes it worse, or something.

But even so, if I accept that I'm looking for a the cards that are worst at fulfilling their roles, it still just doesn't say anything meaningful. Yeah, Cellar is worse than Warehouse, and under these premises, that is indeed "pretty solid evidence that it's weak". But so what? I'm not looking for pointers on what seems weak, I'm looking to rank cards from weakest to non-weakest. What does it mean that there's evidence that Cellar is weak compared to Warehouse? How does that compare to how weak Duchess is compared to Silver? Why aren't I comparing how weak Cellar is compared to Duchess?

[theory]

Well, because the two do very different things.  So when you compare them, you're implicitly comparing how they perform each of their roles, rather than some other metric.  Like, Lab is really bad at performing Cellar's role, but a better card because it does the "job" it's supposed to do with, let's say, an A-.  Whereas Cellar is maybe a D in its category?  So I think of "The Five Worst $X Cards" as looking for the cards that cost X and that are F or D in its role.

And yeah, assigning these grades in a category is arbitrary, but you look for telltale signs.  One way to tell is to look at similar cards in the same category.  If you think (as I do) that Warehouse is maybe a B, B+ in the same category, and you also think that Warehouse is much better, then you have no choice but to consider Cellar a C-, D card.


Of course, not everyone thinks of these lists in the same way.  This just happens to be how I compare cards that do very different things.

[Fabian]

Well fair enough. I still don't see how this is useful at all, since Gold would be a D card at fulfilling Diamond's role, but is still clearly a better card than Walled Village, which might be a a C at fulfilling Village's role. Or any other of countless examples.

Like, I think your way of evaluating cards would make some sense if you made a "top 5 Warehouse-type cards", but since we're comparing $2 cards in this example, not Warehouse-type cards, I don't see any upside to it, only confusion and misevaluations.

And yes, Duchess and Cellar do indeed do very different things. Isn't that why we're trying to decide which is better instead of saying "these cards do the exact same things"?

[papaHav]

Diamond would definitely make gold underpowered.

To say "cellar is bad" doesnt mean anything on its own. It has to be worse than other cards available in the game.
The only way to compare the merit of a card is with other cards that exist.

[theory]

Well, Gold's not a "D" at its role.  Its role is to make money, and it does that really, really well.  So it's an A. 

Walled Village, what is it supposed to do?  Give you +Actions, help an engine.  How does it do that?  Well, C.  It's pretty average.  Doesn't hurt you, but doesn't help you like Worker's Village or Fishing Village or Border Village.  Is it bad enough to be on a Worst $4's?  Probably not, because cards like Thief are a D/D- at its role (to get money into your deck and hurt your opponent's deck; the first of which it's merely ineffective at, and the second of which it's actively counterproductive).

[Fabian]

papaHav, Gold would be underpowered in the games where Diamond appeared. Clearly it wouldn't make Gold underpowered in general though, right? That's my point.

And yes, clearly you need to compare cards to something.. like Duchess? And the other $2 cards? When making a "worst $2 cards" list?

And theory, I understand we disagree and all and that's cool, but calling Gold an A when Diamond is around is clearly very silly.

[DG]

Since it is pertinent, I'll repeat what I posted on the Dominion Strategy site.

"I’d say that cost is too important to ignore. Cheap deck components should be cheap. You can’t get cheaper than cellars. Walled villages and farming villages are often too expensive to pursue strategies that villages or native villages can support. Monuments and militias are not weak just because goons exist. I think the reason the cellar is in the worst 5 list is because by the nature of lists, ‘something has to be’. Used at the right time the cellar, moat, and secret chamber can be decisive game winning cards. Those times are just less frequent than for the hamlet."

The cellar is a balanced 2 cost card. Here's a surprising thought though. If it cost 3 then I bet people would only buy it when they needed it, and once they only bought it when they needed it the card would look a lot better!

[werothegreat]

Diamond would definitely make gold underpowered.

To say "cellar is bad" doesnt mean anything on its own. It has to be worse than other cards available in the game.
The only way to compare the merit of a card is with other cards that exist.

Available in the game that you're CURRENTLY PLAYING, which I think a lot of you are forgetting.  If there are cards that are "better" than Cellar in your current setup, then yes, Cellar is bad.  But if Cellar is the only thing that does what it does, it can be essential in certain setups.  For example, the first game setup.  It takes far too long to use Remodel and Mine to trash away useless garbage - it's much easier to Cellar it away - which also helps you draw more cards to coincide with your Smithy.  I was playing this setup the other day (admittedly with a first-time player) but with Market, Cellar, Village, Smithy and Mine, I had Gold pouring out of my ass and was eventually able to buy two or three Provinces every turn.  Without the Cellar, that would have dwindled down to one Province.  And again, since most of you play with a random selection of cards, your focus should be on what immediately is better than what, what you should be early game, what mid-game, what late game, and not "hey, is this card better or worse than some card that isn't even in this setup?"  From what I gather, no one ever picks and chooses what cards to play with on isotropic - in fact, this seems to be frowned upon.  So if you're not picking "best" cards, how can you value these cards merely based on something doing its job "better"?  Determine the worth of a card on its own, not compared to whatever else is out there.

[theory]

That's very true, but if you aggregate every possible random setup, you'd realize that Mountebank is a "good" card in X% of them, and Thief is a "good" card in Y% of them, where X >>>> Y.  Identifying the "best" and "worst" cards is an exercise in finding which cards have the highest X and lowest Y.

Alternatively, you can consider that if you buy Mountebank on a random board, on average, your win probability changes by X (where X is positive), whereas if you buy Thief on a random board, on average, your win probability changes by Y (where Y is negative).  Call this X or Y the "win probability delta".  So we try to identify the cards that on aggregate have the most positive and least positive win probability delta across all boards.


(this is such an rrenaud post)

[jotheonah]

On the other hand, such a metric could just as easily reflect a card's being hard to play well as its being bad. A good card that's often played poorly would rate as a weak card in that system, and only super-easy-win cards rate as strong.

[werothegreat]

That's very true, but if you aggregate every possible random setup, you'd realize that Mountebank is a "good" card in X% of them, and Thief is a "good" card in Y% of them, where X >>>> Y.  Identifying the "best" and "worst" cards is an exercise in finding which cards have the highest X and lowest Y.

Alternatively, you can consider that if you buy Mountebank on a random board, on average, your win probability changes by X (where X is positive), whereas if you buy Thief on a random board, on average, your win probability changes by Y (where Y is negative).  Call this X or Y the "win probability delta".  So we try to identify the cards that on aggregate have the most positive and least positive win probability delta across all boards.


(this is such an rrenaud post)

And that's totally fine.  I am for this.  This is a good way of rating cards.  Though Thief could still be effective if played correctly.  If your opponent has heard how sucky Thief is, they'll avoid it, but if you go for it (provided there are +2 Actions available), it will make them hesitant to get Treasures, diminishing their buying power, and forcing them to buy Actions with + Coins.  In the meantime, you're using their money to buy Victory cards.  It all depends, really.  Ooo I was playing with a rather silly person the other day - Pirate Ship was on the board, but he was buying Golds and Silvers... he was severely regretting that when my four Pirate Ships were at +$5 and I was getting multiple Provinces.

[werothegreat]

On the other hand, such a metric could just as easily reflect a card's being hard to play well as its being bad. A good card that's often played poorly would rate as a weak card in that system, and only super-easy-win cards rate as strong.

Also this.  Yes.

[HiveMindEmulator]

Alternatively, you can consider that if you buy Mountebank on a random board, on average, your win probability changes by X (where X is positive), whereas if you buy Thief on a random board, on average, your win probability changes by Y (where Y is negative).  Call this X or Y the "win probability delta".  So we try to identify the cards that on aggregate have the most positive and least positive win probability delta across all boards.


(this is such an rrenaud post)

I kind of get the argument about "there exist better cards at different prices" argument. The existence of venture doesn't *make* adventurer bad, it's just *evidence* that it's bad (because as a game designer you can make one at a similar cost without it being overpowered)...
However, I don't agree with the definition of "good" and "bad" in this post. You don't buy cards on a "random board". A card should be "good" if you *want* to buy it more, not if it helps if you randomly buy it. Similarly "bad" cards should be cards you can most easily get away without buying, not cards that hurt you more if you randomly buy them. If you only buy random cards, you lose out on the effectiveness of cards in combos, because by executing these combos, you are more often purchasing cards with other cards they go well with, rather than with "random" other cards.

[timchen]

I think combining Fabian and Theory makes a pretty solid point. A card is bad is just because it is intrinsically bad (at its price). Nevertheless, comparing similar cards can give us hints on whether a card is bad or not.

On the other hand, I actually have never realized how bad cellar is in a BM game. In the simulator it seems BM+cellar cannot win against BMU. Does the play rule have something to do with that?

From another thread now we know that in a BM-ish setup Fool's gold is actually better than silver and gold. How does that change the place of it on the list?

[tlloyd]

you're looking for the weakest cards (defined as the cards worst at fulfilling their roles)

Theory, I take it you agree with my critique then? You are actually evaluating the relative value of cards with similar functions in a list that purports to be evaluating the relative value of cards with identical cost.

So I think of "The Five Worst $X Cards" as looking for the cards that cost X and that are F or D in its role.

Obviously you can think whatever you want, but your thinking here is ... flawed. A crappy trasher can be stronger than an equally-priced and moderately effective deck inspector if trashing is much more valuable than deck inspection. Let's put it this way: each card's value is a product of the value of the role it plays (draw, trashing, cycling, sifting, actions, etc.) and how effectively it fills that role (fishing village provides +actions better than other cards that serve that same purpose). If you want to rank the $2 cards, you need to calculate this for each card and then rank them. What you are describing is actually a two-step process: rank the cards by function, then group them by price while trying to keep each card near the same rank it had in the by-function list. That is just nonsense.

If you decide that Cellar has some rank relative to the other $2 cards based on how often you want to sift/cycle and how well Cellar does that, and then Warehouse comes along and serves the same purpose but does it better, how could that possibly affect the ranking of Cellar vis-a-vis the other $2 cards? Compare: on average, Witch is a better card than Counting House, using whatever metric we're using for these comparisons. Now I invent a card called Cheap-Witch which does exactly the same thing but only costs $4. Given that Cheap-Witch is strictly superior to Witch, you will never buy a Witch, right? So now Counting House is better than Witch, right? Of course not.

The only way I can think to save your argument here is to say what we're really talking about is how frequently you will want to buy a card. Cellar is better than Moat, but since Warehouse fills the same role as Cellar and is better than Cellar, you'll actually buy Moat more often than Cellar. But that's pretty weak reasoning given how rarely Cellar, Warehouse and Moat will show up in the same kingdom.

[Fabian]

fwiw, I think theory's last post describes a much more reasonable way of ranking cards (though it would still have problems since you don't buy cards at random in any/all setups, especially for engine type decks, as HiveMind points out, etc), but it has nothing to do with the original theory method of grading all trashers relative to each other, then grading all village type cards relative to each other, then grading all XYZ etc, and whichever cards get the lowest grade in their respective category goes on the "worst $2 list", which I and many others have pointed out is basically ... flawed.

Edit: Or put another way, Cellar's win probability delta isn't negative/crap because Warehouse does Cellar's job better, its win probability delta is negative/crap because the card is crap (most of the time etc)

Or put yet another way, Warehouse being printed did (close to) nothing to affect Cellar's win probability delta.

[theory]

If you decide that Cellar has some rank relative to the other $2 cards based on how often you want to sift/cycle and how well Cellar does that, and then Warehouse comes along and serves the same purpose but does it better, how could that possibly affect the ranking of Cellar vis-a-vis the other $2 cards? Compare: on average, Witch is a better card than Counting House, using whatever metric we're using for these comparisons. Now I invent a card called Cheap-Witch which does exactly the same thing but only costs $4. Given that Cheap-Witch is strictly superior to Witch, you will never buy a Witch, right? So now Counting House is better than Witch, right? Of course not.

Real-world example: I want to buy a 22" monitor.  I see one on sale for $200.  I don't really know if that's a good deal or not, but I have a gut feeling that it isn't.  I search Google and find 23" monitors that are otherwise identical and on sale for $150.  That is pretty good evidence to me that the 22" for $200 is a bad deal -- even if the 23" monitors are temporarily out of stock.  It could be that the 23" monitor people are insane and wildly underpricing their stuff, but I assume market pressures keep them in line.  (Market pressures in the analogy being Donald X.'s sense of design.)

On the flip side, if I see 23" monitors on sale for $100, that's pretty awesome.  When I then see a 24" monitor for $50, which I instinctively know is an unbelievable deal, that isn't proof that the 23" for $100 a bad deal.


So back to Cellar/Warehouse.  Warehouse is not causing Cellar to be "The Worst", any more than the 23" for $150 "causes" the 22" for $200 to be a bad deal, because 22" for $200 really is just inherently bad even if all the 23" are sold out right now.  But to an uninformed consumer, it's good evidence that Cellar can't be very good, because even when Cellar is a lot better, it's still not overpowered. 

[theory]

As for the win prob delta vs role grades as far as choosing cards: I don't know what I'm actually doing in my mind.  I have a suspicion that in the end, all of these metrics come out with the same answers.

[DStu]

But what is strange with all your example is how

In base Dominion, this was an absolutely critical card for Laboratory and Village/Smithy chains

can be inherently bad.

I think we should not take this too seriously. There are situations where the Cellar is a great help, but there are more situations where it is not. Pearl Diver usually does not harm, and has gotten some support by the expensions, and get's some bonus for being #1 last time. Cellar has not improved, and it's just the $2 list, there are not so many $2s, and it's just the #4. That does not mean that it's always an error to buy it. What's left? PD, Duchess, are above, Duchess being a special case. Haven, Pawn you can place above Cellar. XR is the better Cellar in many scenarios where you would really want a Cellar. FG was just elected overpowered by the simulators. BM-Courtyard is strong. Lighthouse is the strongest defendence and has some combos. I don't have to defend Hamlet here. Native Village has some really strong megaturns, Embargo ok maybe.

[Fabian]

Why are you assuming any of us are uninformed consumers? Your monitor example works a lot better if all of us are computer monitor experts. Put a bunch of beginners in charge of ranking the worst $2 cards, and sure... kinda... maybe.

After your very scientific ("such an rreanud post") last reply, it was surprising to get back to something like "I dunno I guess both work and are kinda the same", which seems very much like the opposite.

[tlloyd]

Warehouse is not causing Cellar to be "The Worst", any more than the 23" for $150 "causes" the 22" for $200 to be a bad deal, because 22" for $200 really is just inherently bad even if all the 23" are sold out right now.  But to an uninformed consumer, it's good evidence that Cellar can't be very good, because even when Cellar is a lot better, it's still not overpowered.

What this seems to ignore is the fact that you are not actually assigning each card some grade on the scale of Terrible to Amazing. You are ranking the cards relative to each other! (i.e., "better/worse" not "good/bad"). If Cellar is better than Moat, and then Warehouse is better than Cellar, that cannot possibly change the fact that Cellar is better than Moat. It certainly means that Cellar is weaker relative to Warehouse. It may also lead you to believe that Cellar is a weaker card in absolute terms. But since you already know that Cellar is better than Moat, then Moat must be weaker in absolute terms as well.

Maybe a counter-example will help illustrate: assume Minion is better than Lab. Now Horse Traders comes out. Horse Traders actually changes the value of Minion in absolute terms. Even though we won't often have Minion and Horse Trader in the same kingdom, it at least makes sense that the existence of cards like Horse Traders would change our valuation of Minion, not relative only to Horse Traders but relative to everything. It is thus a plausible argument that the creation of Horse Traders (a $4 card) could change the ranking of Minion and Lab ($5 cards) relative to each other.

But Warehouse doesn't change what Cellar does or how effectively it does it. Warehouse is simply better than Cellar (usually) at doing pretty much the same thing Cellar does. That may suggest Cellar is underpowered, but cannot logically change the valuation of Cellar relative to the other already-existing $2 cards.

[rinkworks]

I wanna know why Estate didn't make the cut.  Surely Estate is one of the 5 Worst $2 cards!  :-)