This article is written by HiveMindEmulator, originally posted on the forum.
Dominion: Cornucopia
Fortune Teller is one of the more underrated attacks in the game. The problem is that it’s a deck top attack, which is subtle if you’re not paying much attention (particularly online where you don’t actually have to physically discard cards or click anything). You don’t have to discard cards from your hand, and you don’t get a big purple card in your deck to catch your attention. You may just feel like you’re getting worse draws than usual. But this isn’t a matter of draw luck, it’s the attack at work.
Attacks are very often the most important cards in a given kingdom. You have to be aware of the type of deck that the attacks thrive against. If your opponent’s deck doesn’t have much need for treasure, then treasure attacks like Thief, Pirate Ship, and Noble Brigand are not going to be useful, but if they’re going for a high density of big treasures, then they will be. The big impact of Noble Brigand is not that it is a must-buy attack every time, but rather that it disables big money strategies, which means you have to consider building a deck to avoid it when you’re planning out your strategy.
So what does Fortune Teller do? It skips some of their cards and leaves a (usually) useless one on top of the deck. There are two key parts to the attack: the cards skipped, and the junk left. Leaving junk means that their next hand (or this one if they draw) will have at least one dead card, which is slightly less painful than having them have only a 4-card hand. This doesn’t seem like much, but for decks that rely heavily on Silver and Copper, it makes it really hard to hit $8, since you need the average coin value among the 4 cards to be $2. The other part of the attack, the skipping cards, is a little more complicated to understand and causes a lot of confusion. If you’re scared of math, maybe you should skip the next paragraph…
Consider a deck with N cards that sees an average of M cards per turn. Then the probability of seeing any given card on a given turn is M/N. When attacked by Fortune Teller (assuming there is a victory card or Curse to leave on top), the probability of seeing a given card is something like (M-1)/(N-1). Since M<=N, this second quantity is smaller than the first, and the percentage difference is much greater for small ratios M/N. Consider a deck with no drawing (M=5) and N=20 cards. The probability of seeing any given (non-victory or curse) card is 5/20 = 0.25. When attacked by Fortune Teller, it is 4/19 = 0.21. This means you see your good cards (0.04/0.25=)16% less often! On the other hand, if you draw M=10 of your N=15 cards every turn, you go from 10/15 = 0.67 to 9/14 = 0.64, which is a decrease of less than 5%. The impact could greater, since your drawing cards will show up less often, thus decreasing the numerator by more than 1, but this is still a long way from the 16% figure for the larger deck without drawing.
Of course just computing these average doesn’t tell the whole story. In reality, you must see each card an integer number of times, your turns are not independent and your deck changes over time. But the main takeaway still holds true. When you’re attacked by a Fortune Teller, even though you shuffling more, you seeing your good cards less often! The benefits of the cycling are generally more than counteracted by the skipping good cards. You also get a sense of the trend in terms of deck size and drawing. In the early game when your deck is small or in situations where you have a good draw engine set up, the attack is only mildly annoying, but it can become very powerful against a slog-type deck. Also note that this analysis doesn’t depend on the number of good or bad cards in the deck, just the total number of cards and the average number drawn per turn.
So where does that leave us in terms of strategy? The key idea is that you want to avoid the types of decks that Fortune Teller will be strong against. A heavy drawing strategy utilizing Fortune Teller should dominate a slog-type strategy or a strategy that relies heavily on Silver and Copper. If you can’t build an engine to consistently play Fortune Tellers, they can still be useful in slog vs slog. Now the interesting thing is that if you’re both going for the engine strategy, you may not want to bother using a terminal action on a Fortune Teller, since the impact will be small, but the presence of the card in the kingdom is still of importance since it made non-engine strategies less effective.
Counters
In addition to building decks which draw a good percentage of the deck, there are a few other ways to deal with the Fortune Teller attack. If you build a deck with no victory cards or Curses, the attack just discards your deck and can’t leave you with a junk card. So you can go for strategies which trash Estates and green late or focus on VP tokens. There are also direct counters to deck top attacks like Jack of all Trades, Farming Village, Scrying Pool, Golem, Sage, Oracle, Native Village, Lookout, Chancellor, Scavenger, Adventurer, Venture, and Scout. It is also important to note that unlike Rabble, the wording of Fortune Teller is such that it does not skip dual-type victory cards. Nobles and Harems can be drawn more often rather than less often since the Fortune Tellers seek them out. You can also go for cards that like to have VP cards in hand like Baron, Crossroads, or Tournament.
Combos
There are a few specific combos that can take advantage of the deck-top attack of Fortune Teller. You can combine it with Jester to give guaranteed Curses, though you lose the normal effect of the attack unless you play a second one. You can also use to to target down VP cards to trash with Saboteur. When followed by a Minion, it essentially forces a 3-card hand. It can also be used to mitigate the potential benefits your opponent could get from your plays of Margrave, Council Room, Vault, or Governor for cards by ensuring the card your opponent draws is junk. Then of course there’s the anti-combo with other deck top attacks. You can only make the top of their deck so bad. If your other attack already put a victory card on top, Fortune Teller won’t do anything.
Swindler can be used in conjunction with Fortune Teller to destroy Victory cards . Although it usually doesn’t work well with Provinces (since you would just give them another Province), it can be devastating against alt-VP like Dukes or Gardens.
Example games
In Obi Wan Bonogi’s engine vs my HP+Salvager:
http://councilroom.com/game?game_id=game-20120613-234441-e6aae5cf.html
Here I figure that with Hunting Party and Salvager, this game should go pretty quickly, but Obi Wan has other ideas. He picks up a Fortune Teller which he proceeds to play nearly every turn, slowing me enough that he’s able to hold me to 3 Provinces in 16 turns.
Against lespeutere’s Silk Roads
http://councilroom.com/game?game_id=game-20130123-075538-d2f6ab16.html
Lespeutere goes for Nomad Camps and Silk Roads, which is countered quite strongly by my Apprentice + Fortune Teller. The constant attack delays his collection of Silk Roads and then makes it difficult to afford Duchies.
In a Sea Hag slog with Rabid:
http://councilroom.com/game?game_id=game-20120731-101955-5cbeb2bc.html
With Sea Hag giving Curses and no handsize inceasing other than Moat, this game is bound to be a slog. Rabid opts for a second Hag to win the Curse split while I prefer to get an early Fortune Teller to provide economy while still allowing me to attack. (Note that the immediate impact of both attacks is nearly identical: skip card(s) and leave junk on top.) I do end up losing the Curse split, but my deck builds up much faster, and by the time I take my sixth Curse, it’s turn 17 and I already have a Platinum. You will also notice that even before that point, he draws just as many dead cards as I do. I have a little luck hitting one of his Hags with mine on turn 6, but I’m pretty sure the strategy is still better.
Against qmech’s Embassy big money:
http://councilroom.com/game?game_id=game-20120716-110717-71cd22f0.html
Nobles + Vineyards probably indicates engine here anyway, but qmech decides to go for Embassy big money. I counter by adding a Fortune Teller which helps to make sure he can’t end it too quickly. This one ends up not being close enough that the Fortune Teller really won the game, but if there are Gardens instead of Vineyards or something it might matter.
Good with:
– lack of drawing (at least in opponents deck)
– opponents large deck size
– draw engines that allow repeated play
Not as good with:
– opponents heavy draw engines
– opponents decks with no victory cards/Curses
– dual-type victory cards
– cards that want victory cards in hand (Baron, Crossroads, Tournament)
Dominion Strategy 
This card is one of the few cases where you’d want to buy more than one Chapel….. I was smacking myself for working the deck incorrectly… maybe a Chapel was accidentally discarded without getting played when I realized that by my 5th turn, it still hadn’t come up yet. Then I remembered that I was playing a digital version of Dominion, so that shouldn’t be possible… the computer does all of that for you and WOULD do it correctly. Then I noticed that yeah, my single Chapel got “Fortune Teller”-ed away twice.
Only plus side with Shelters is that it bypasses Hovels.
The math makes no sense over there. And the fact you’re seeing your cards less is a plain corollary to the fact that your next hand/draw will be junk. The point which is true is : it increases the variance of your draw. As ackmondual said in the previous post, it might happen that you get your key card discarded and miss it for a turn or two, although you are generally guaranteed to draw it by turn 5. But, inversely, you might have had your key card in your 3rd hand (say with 2 estates and 2 coppers, trash all of them), get attacked, revealing your last estate at the bottom and draw chapel again in your 4th hand, trashing your last victory/curse card of your deck. Then, fortune teller is utter crap.
When you feel you’re less brilliant than your opponent, it might be an optimal strategy to rely on positive variance, hence luck. But generally, that’s not what we seek, because we want to prove we are the best.
I personally don’t think that Fortune Teller is an underrated card. It may be a trap to newest players that did not realize that it’s not a silver with advantages, since you need actions to play it. That’s all about the play it should get if you ask me, and your article tends to prove me the same : it’s effective only against a certain type of strategy, in a kingdom without a dozen of cards and only if there is no better card that does the same wth a better bonus (approximately, all the +2 coins are better).
I disagree, and your comment exactly shows why it’s underrated: “It’s usually bad, therefore just don’t care about it!”
Quoting myself from the forums:
“I think it is pretty common that FT is not a good card, but all these complains is why people consider it a weak card.
– In Big Money, you don’t want to buy it in the beginning, in the middle it’s hard to get, and you don’t see it that often anymore anyways.
– In an Engine, there is usually trashing, so I will have no targets.
That is what makes FT bad, and this all still holds. But as with every card, there are situations where they are better, maybe even useful cards, even if it’s the worst card of all. And knowing this is important, and exactly what this article tells:
What if the situation is different? What is if this is an engine, but there is no Estate-trashing? What is in Slogs? What is in engines that can’t buy out in one or two turns (Maybe because of Alt-vp and higher total victory count)?
Especially pure Cantrip engines really don’t like to find stopper cards, and having one of them on top of the deck is exactly what you don’t want. Especially engines also don’t have the problem of “I can’t pick that up because I Gold is more important”, and they can play cards often. Of course it’s pointless if the Estates are trashed and you only expect one turn where there are Provinces in the decks. But what if not? Look out for this situations, and there FT is a valuable attack, especially if the draw of the engine is not that strong that every card matters.”
If the only defense you have for this card is “it’s awful but it can be a bad evil sometimes”, well, ok. For instance, if the kingdom is made of nine +1-action cards and fortune teller, then the first silver you buy should be a fortune teller, since it’s strictly superior. Also, if you plan to trash your first silver as soon as possible with remodel/salvager, then you’d better open with it than with a silver.
It’s still a very narrow card, and I can’t see how it can possibly be underrated. Bad does not mean that I will never play it at all. Bad means that I will play it more often because of my opponent swindlering my swindlers/silvers than because I bought it myself. Also, note that it’s not that much low-rated in the “best 3$ cards”, ranked 27th on 32.
Anyway, my main point was not to say that the card is unplayable, but to recall that the virtual 4-card hand and the expectation computation of the math paragraph were exactly the same effects and thus shouldn’t be added like a double kiss-cool effect.
Even if it is the same effect, more often than not the “missing the key card” view is more important. Like missing a Workshop/Irownworks/Trader while going for Gardens/Slik Road is by far worse than getting hit by Militia, even if Militia makes you go below $4. Also, while building engines without strong trashing, there’s usually a few really important cards in the midgame (you have a single KC for instance, or miss a Wharf if you have 1 or 2) and missing them for a whole reshuffle will really slow you down. With strong trashing, missing your trasher hurts, although picking up an extra terminal instead of Silver can hurt the same, so this final case is probably more about variance. I may open Chapel/FT instead of Chapel/Silver if I am player 2 and there are good villages for later.
I think you missed my whole point. FT does not lower the expected value of hitting your key card, besides the fact that you’re forced to draw a useless card. It just increases the variance associated with this event.
In fact, in case of massive trashers such as chapel or mint (if you buy it turn 3), the EV of hitting it is even a bit favored by FT (still besides the fact you have virtually 1 card less in hand). Another case where FT can be counterproductive is the case where your opponent buys a key card on turn 3 (a gold for instance). He then has accessed to it earlier than expected thanks to the milling FT operated.
And once again, still there is this virtual -1 card in hand that will make it better than a silver 99% of the time if it’s your only terminal action.
Now, I realize also that each FT play per shuffling has an effect of shrinking your deck by 2.25 active cards as an expected value (if you consider that you have 9 active cards and 3 estates at this point of the game), which is equivalent of adding virtually and temporally 1 junk card in the deck. Cursing the opponent is good because it’s permanent and additive. FTing the opponent is bad because it’s temporal and non-additive.
“FT does not lower the expected value of hitting your key card, besides the fact that you’re forced to draw a useless card.”
But that’s the whole point of the card… Of course if you ignore the main thing a card does, it’s not going to be good…
“FTing the opponent is bad because it’s temporal and non-additive.”
So is Ghost Ship. But Ghost Ship isn’t bad, it just needs to be played frequently to get the best effect. Fortune Teller is a similar story.
“But that’s the whole point of the card…”
==> Then we agree, the only effect is to force your opponent to (maybe) draw a useless card and there is absolutely no other EV decreasing of drawing your “good” cards, contrarily to what the article states. That’s my main revendication!
Where does it say that? I don’t really like the math part either, probably your explanation is clearer, but I don’t see where it states an additional effect. It just says that the “but you shuffle more often, this counters the possibility of discarding good cards” is not valid. As I understand it…
That’s not what I understood from the article, and I agree with what you say.
Well, there are still 2 ways the attack affects you, both obviously related to the same mechanic, since there is only really one action the card does:
1. You have less (useful) cards on your next turn. This makes it hard to do whatever you want on that turn
2. Because you are forced to draw your VP cards and Curses more often, you are drawing your other (good) cards less often. This may seem like an obvious consequence of (1) to some people, but to others it isn’t. This differs from Militia, which actually only affects the next turn, but not cycling at all.
Another case where FT can be counterproductive is the case where your opponent buys a key card on turn 3 (a gold for instance). He then has accessed to it earlier than expected thanks to the milling FT operated.
Unless, of course, the FT puts it right back in his discard. It depends on timing and luck, but you’re almost certain to hit things other than copper at least once during the game, and often several times. Sometimes, even the cards that could have been counters to FT will be hit by it instead. The fact that it’s somewhat luck dependent doesn’t necessarily make it a bad card, especially from a player 2 perspective.
I do think that Chapel may deserve to be listed among the counters though — unless the Chapel itself is hit, having that bad card in your hand isn’t so bad, in fact, it’s what Chapel is there for. After a few rounds of this there is nothing left to hit, unless a curser is on the board. (Although the FT will come back in the endgame, if you still have it. Cursers have usually become worthless by then.) Ditto Salvager, Remodel, Bishop, etc. Cards that make your bad cards less bad, like Oasis and Warehouse, too. Although any of those *can* also be discarded by the FT, so maybe they’re not as good against it as they look? If you could count on always drawing your counter card at the right time, then even the mighty Mountebank would be nothing but an expensive terminal silver on a Moat board, and we all know that isn’t true.
Anyway, FT is not non-additive, it only looks that way. Systematically having worse hands has a very important cumulative effect. A repeated FT (say, FT/Scheme) adds up to a lot of settling for inferior cards; once you fall short of buying that turn 5 gold, you’re going to spend the whole rest of the game with one less gold in your deck *and* without all the other things you could have bought every time that gold would have come up (and all the times those things would have come up, etc. — this rapidly becomes too complicated to track in your head, especially when it’s actions that are being missed, but it definitely does add up.)
Unless by “non-additive” you meant it in the same sense as Militia or Margrave, that hitting someone with multiple FTs on the *same* turn has little added value, and therefore it’s not really worth planning to TR/KC it. That is true, and should perhaps be noted, but it’s equally true of some quite powerful and deservedly feared (and more expensive) attacks, such as Ghost Ship and Minion.
This does deserve a special note in 3+ players though — if your opponents are hitting each other (and you) with FT consistently enough, you may be able to skip buying it yourself and still get most of the benefit. This site is so 2P-centric that things like that will often get overlooked.
Nobody is saying it’s Mountebank, and for $3, it damn well shouldn’t be. And there are certainly boards where any $3 terminal is a questionable idea because you’re going to trip over it down the line once you start picking up $4 and $5 terminals. But the attack is easy to underrate and its beneficial effect is almost as good as non-attacks of the same price.
I’m just replying to explain the part of my post you quoted :
[i]Another case where FT can be counterproductive is the case where your opponent buys a key card on turn 3 (a gold for instance). He then has accessed to it earlier than expected thanks to the milling FT operated.[/i]
I mean that, by milling the opponent, FT speeds up his access to the cards he newly bought.
You are still exaggerating and still showing that you underrate it when you say that it’s useful only when you can use it as a Silver+.
I remember giving more examples in this post which you basically ignore.
The situation where you play an engine and the opponent can’t trash Estates are common enough, and in this situation it is a decent card which is a valuable addition to the deck. Playing against a slog is common enough, forcing it to a 4 card hand is also usually strong. Of course there are stronger things around, nobody wants to put FT on the #1$3 cards. “Fortune Teller is one of the more underrated attacks in the game” is also not a very high bar, there are not that many attacks that are underrated, usually it’s quite the opposite…
Am I the one who is exaggerating? I tell that I don’t deny that there are obvious cases where it can be played as a silver with benefits and you interpret that it’s the only case I would play it?
For FT to be useful in a game, you need :
– little trashing in the kingdom
– no better attack (basically equivalent to “no other attack”)
– you are on engine
– your opponent is on slog
– no card that negate its effect in the kingdom (golem, farming village, loan, …)
Or FT be the best terminal action, which may happen from time to time.
If we don’t agree it’s underrated, it can mean 2 things :
– we do not agree on its commonly agreed rating, but the 3$-cards ranking should avoid this
– we do not agree on its value, then may I ask you at which place would you rank the card? I could see it better than Smuggler, considering the luck factor of this latter. But besides this card, I really don’t consider it surpassing any of the 25 other 3$ cards.
The point was that the case you presented was the tautological one.
@rating: I don’t think the answer will say anything because of different methods how to get the cards into a one-dimensional rating, but my impression is that the cases where FT is a decent attack is often not recognized.
“Against lespeutere’s Silk Roads
http://councilroom.com/game?game_id=game-20130123-075538-d2f6ab16.html”
Well payed… I liked how the fortune teller won this one…. good job 🙂
“For FT to be useful in a game, you need :
– little trashing in the kingdom
– no better attack (basically equivalent to “no other attack”)
– you are on engine
– your opponent is on slog
– no card that negate its effect in the kingdom (golem, farming village, loan, …)”
Well, so no, not all of this is true.
1. Little trashing in the kingdom is fairly common. Most trashers are “weak.” And unless you opponent is going to score with VP tokens, he’ll have to pick up Victory cards at some point anyway, which makes him vulnerable to FT.
2. A better attack will take precedence over FT when there are few actions to spare, but they are not necessarily mutually exclusive. Some engines will have a surplus of +action.
3. FT is weak against engines but rather strong against not only slogs, but money strategies as well.
4. An opponent will not always have an FT counter in hand, especially if playing a money-oriented strategy.
When you say it’s not as good against opponents’ big draw engines, shouldn’t youdraw a distinction between reliable and unreliable engines? According to your math analysis FT doesn’t hurt an engine very much IF you assume the engine is reliable enough to go off despite being attacked. If the engine’s reliability is shaky though, then FT could really hurt.
Fortune Teller is also one of the few attacks that can be combined with discard attacks for a slightly stronger effect. It may not seem like much, but discard attacks are already strong so making them incrementally a little stronger can be a big deal.
Yes, if the opponent draws a dead card, they can just discard it, but they are still harmed by having at most 4 good cards to choose 3 from, instead of possibly 5 good cards. Also, if the opponent’s deck contains cantrips or drawing cards, a discard attack + top-decking a junk card makes it much less likely that they will have enough in their hand to start off an engine, or draw up to a reasonable hand.
Also consider Ghost Ship–play Ghost Ship, then Fortune Teller every turn and you will set them up so that they can only top-deck one good card for next turn and can only do it if they top-deck a dead card with it. This also minimizes the deck-cycling effect of Fortune Teller.
Nice point! In fact I think we could add that FT goes well with Cursers too. If your opponent only has two junk cards and you topdeck one, the odds of finding the second among the other four cards are greatly reduced. If they have ten junk cards and you topdeck one, the odds of finding more junk in the next four cards don’t really change that much. Or to look at it another way (which is perhaps more rigorous), look at this quote from the article:
“Consider a deck with N cards that sees an average of M cards per turn. Then the probability of seeing any given card on a given turn is M/N. When attacked by Fortune Teller (assuming there is a victory card or Curse to leave on top), the probability of seeing a given card is something like (M-1)/(N-1). Since M<=N, this second quantity is smaller than the first, AND THE PERCENTAGE DIFFERENCE IS MUCH GREATER FOR SMALL RATIOS M/N." [Emphasis added.]
In fact the same could go for trashing attacks since they, too, raise the ratio of junk to good cards. (Of course FT also works with trashing attacks that want to hit Victory cards but that's a separate matter.)
So it looks as though FT works well with almost all attacks except other deck inspectors. Very interesting.