There have been many threads in many places recently that have pointed out that it's perfectly legal under policy to partially answer questions about objects, so long as all of the things said are true.

For instance:
A: What's Vampire Nighthawk again?
N: A 2/3 flyer

A: what colour's that token? (From beckon apparition)
N: white

A: What card types are in your graveyard?
N: Land, sorcery. (There's also an instant.)

If A phrases the questions very slightly differently, these answers are illegal. This means that players have to know the magical incantation for getting good answers out of their opponent, leading to lots of feel bad situations, and also to a lot of angle shooting. (In the case of creatures, the magical incantation is “What is the oracle text of that?”)

Why do we allow this?

I realise that policy that makes this situation illegal, without putting too much obligation on the answerer is harder than the current policy to write. However, it seems worth the effort.

Can wiser heads, or ones that have a fuller perspective on the history of things, enlighten me?

A “full” answer is rarely required in real games of Magic. If I have a
Lightning Bolt in hand and ask my opponent, “What's that Nighthawk?” I am
actually just looking for “A 2/3.” I don't need him reciting the full
Oracle text, and it would be totally insane for him to have committed an
infraction by giving me the answer I wanted.

Or if I have Debtor's Knell on the board and I ask, “What's in your
graveyard?” and he just tells me the creature cards, why should that be an
infraction?

So you can't require “complete” answers without completely bogging down
communication. And you certainly can't require a player to read his
opponent's mind as to the intent of a question and then answer that. So, we
are stuck with “only say true things.”

Especially given that the usual question is actually “So, how big is
Tarmogoyf?” and then the players cooperate to figure it out, I don't think
this is worth the effort to fix, assuming such a fix is even possible.

This means that players may (rarely) be burned by incomplete answers, but
the alternative is totally unworkable.

I have talked to L4 Jared Sylva about this at GP Strasbourg and as far as I understand the reasoning behind this is that those feel bad moments are better than punishing players for unintentionally forgetting to indicate derived information. Players are furthermore expected to acknowledge derived information on their own if their opponent is willing to help or not.
While I personally agree that it would be more fair to require full answers or no answers in regard to derived information and I dislike the verbal acrobatics that allow people to misrepresent derived information it is still the lesser of two evils.

So with situations like this assuming this is Competitive REL (otherwise these responses are in fact illegal). All of the information you have requested falls under derived information. The Oracle of a card, in your first case Vampire Nighthawk is derived, as is the colors of a token and the types of cards in your yard. So it takes some skill, in this case reading, to get all of the information. I am still responsible for the decisions that I make in game even if I base those decisions on information obtained from my opponent, no matter how complete that information. In most cases I could obtain the same information simply by picking up the card(s) in question. Now you may be thinking, wait! My opponent has a card in a foreign language which I cannot read, well we allow for that as well! Players may request Oracle from a judge, or they may use their smartphone to look up Oracle text as long as both players can see the device.
This allows players with experience to potentially Jedi mind trick their opponent into some play that is possibly bad for them. Now all of this being said they cannot let their opponent do something illegal, such as Terror the Beckon token. Knowing the cards is an advantage and potentially a disadvantage (if you forget some detail) at the same time, and we want players to play their game, if they want/need more information they can always ask for it, but they do have to weight the source of that information. Their opponent is not necessarily incentivised to provide complete information on everything in the game, why should I help my opponent win the game? Now there are certain cases where we do make that distinction, and that is why we have free information, such as life totals and other game state related information. Anyway, I hope that this helps you out! Thanks!

I delve into Communication Policy more than I delve into any other aspect of the rules.

In the end, although it's an old article, the words from This 2009 Player Communication Guide by Toby Elliot still holds true and basically answers everything.

The reason why we make players have make all answers be the “complete” truth rather than “incomplete” truth is that we don't (want to) punish for not telling things they deem uninteresting. The example of Lightning Bolt vs Vampire Nighthawk was already made, another example would be “what is Terravore?” -> “Power/toughness equal to number of lands in all graveyards” ommitting the Trample part (when the answer was made on an empty board).

If policy demands player to answer complete if they answer, I see how these two problems immediatly arise:
1) Players will not communicate at all and instead ask a judge whenever they want to know what a certain card does. Less communication between players is generally bad and it does provide much more work for judges :)
2) Regardless of what the policy is, there's always angle-shooting. Assuming policychange: I ask my opponent what his Nighthawk is and he answers 2/3 lifelink. I block with my Emrakul and he tells me it dies: do I now have a right to rewind because my opponent answered incomplete? Right now players are allowed to give incomplete answers, with other policies players might try to ask questions just to get incomplete answers and abuse that.

As a competitive/semi-professional player, I think the current Communication Policy is fair enough for the players to accept, easy enough for the players to understand in most cases and not abuseable enough to require change.

Because it best represents how:
1) Players actually play
2) We want players to play.

How players actually play: They shortcut things, only focus on things that are relevant. My raging goblin, might have haste, but after the first turn in play, Im probably going to describe him as “just a 1/1 dork”. He is going to leave out the parts that he doesnt feel are relevant.

Lets consider the options of how we want players to play. We really have 3 options: I am required to provide all information, some information or no information when a player asks me for derived information

If I am required to provide ‘No’ information..how difficult would it be to play. Every answer to every question would be ‘I cant tell you’, and what if a player did provide some information? Remember, we are saying they are *required* to provide no information. Would we enforce penalties for being helpful? How miserible would games be with players scared to talk to each other.

Now lets look at requiring all information. How do we enforce that? We have to describe all characteristincs of my Raging Goblin when asked what it is? Including the fact that it is red? If I leave something out, am I cheating? Did I commit an infraction? Again, we dont want players to be scared to talk to each other.

If Im not required to give information, but im allowed to give some…this is the most flexible approach. I dont have to answer if Im hoping they cant figure it out. Im allowed to be helpful and expidite the game by providing an answer to “Cards in Hand?” The idea that a player can game the system by giving partial information and leaving out something important is an unfortunate consequence but, there are ways players can mitigate that. Ask to see the card, call a judge for oracle text, be vigilant. And the amount of good gained, outweighs the potential bad, especially when compared to the alternatives.

“JUUUUUUDGE! He didn't read the flavour text! I demand you disqualify him for cheating!”

To take this to the absurd, if incomplete information was punishable, an opponent could continue asking, “What's that do? And that? And… What was the first one again?” hoping the player makes a mistake. If he does, he calls a judge.

It's both players' responsibility to understand and maintain the game state. If a player gets burned once for incomplete information from an opponent, you can be sure he'll always read it (or look it up) himself from then on.

Hang on. I think we may be getting bogged down by a red herring here.

If requiring complete information is so problematic, then why do we require exactly that at Regular REL?

Edited Jeff Morrow (April 18, 2013 06:22:41 PM)

Figuring out derived information is a skill and something we want to test at competitive. Its part of being a better player. We dont feel the need to test that skill as much at Regular, which has a different focus. However, even at Regular, players often leave stuff out, choosing not to get bogged down in the details, but they leave out the stuff they percieve as irrelevant. Its only when that ‘left out stuff’ becomes relevant do we even get involved. And then we have a documented way to approach the solution and hug it out.

Originally posted by Jeff Morrow:

Hang on. I think we may be getting bogged down by a red herring here.

If requiring complete information is so problematic, then why do we require exactly that at Regular REL?

The difference for me is, at regular there is no penalty for not given complete information. So as long as a player is giving enough information, I would be fine with it. What is enough information would differ from situation to situation and from judge to judge. But a player doesn't get penalized in regular if he tries as good as he can.

In competitive however we are penalizing players for not providing the information they are required to. So in this case it would either introduce incosistency that one judge will give warning for an answer while another won't give a warning for the same answer. Or it would require to make it clear what information a player is required to answer (can we actualy define this completly?). So at competitive we punish players with a warning for trying as helpfull as they can if the fail by some definition. So in the end, it will encourage people not to answer because this is the only way they can avoid a warning for sure. And that is something I never want.

To me, there is a difference between incomplete and partial answers. An incomplete answer is one, where some items from a list of otherwise identical items are left out. For example, when asked “What types of cards are in your graveyard?” and anwering “Land and Sorcery”, omitting “Creature”. With partial answers, some information is left out, but there is no expectation of an complete answer. (“What does that creature do?” - “Attack you for 2.” not mentioning the fact it can also ping creatures).

With incomplete answers, there is an expectation that a complete answer is given, otherwise why bother asking? I don't think it benefits the game at all that we allow incomplete answers. The questions to those answers are often asked to speed up play, and are usually answered truthfully. By allowing incomplete answers, we basically always force players to check for themselves, even if a player expects that the opponent knows the answer. This seems similar to hiding a card in hand when asked how many cards you have in hand, or hiding a land when asked about that. (After all, the opponent could always demand to physically count the cards himself.)

This is also not about bluffing or players being observant. In the example above, the player knows exactly what he wants to know and also knows how to come to that conclusion. Another example is a white creature that is also black due to an effect. Why should a player be punished by asking “What color is that creature?”, instead of “Is that creature black?” The former can be answered incompletely, the latter must be answered correctly if at all. But in both cases the player obviously had the right idea if he wants to attack with his black intimidater, but didn't know the intricate technicalities of wording the question correctly. Remember, that the opponent always has the option not to answer a question about derived information.

Edit: What is comes down to: By allowing incomplete answers, we promote an atmosphere of dishonesty and mistrust, even in a match where both players intend to be friendly and helpful.

(As an aside: Personally I think the Tarmogoyf case discussed in another thread is clear cut lying. Answers should always be taken in context of the question. The answer “I have a Sorcery and a Land in my graveyard.” means something different when asked “How big is your Tarmogoyf?” and “What cards are in your graveyard?”)

Edited Sebastian Rittau (April 18, 2013 10:39:08 PM)

Why should we not expect players to check for themselves, especially at competitive? I trust most (all?) of the players I play with, yet I still always ask, “May I read that card?”

Even from a strategic perspective, asking your opponent might be a tell on what your next play. I remember playing in an RTR draft where my opponent looked over and asked, “Is your Angel of Serenity a monocoloured creature?” *facepalm*

We should not expect players to check for themselves because they don't check for themselves. That's how people play Magic. Even at Comp, it is a social, interactive game, which goes much smoother when players communicate with each other.

I think there's a large grey area here depending on the card in question. If someone drops a Lightning Mauler and you ask, “Is that the first strike guy or the haste guy?”, that probably makes sense. If someone drops a Liege of the Tangle and you say, “What does that do?” it's probably more appropriate to say, “You should just read it.”

I think there's a few different angles to the conversation that lead to where the “line” currently exists…

(1) Cards don't always say what they actually do.
(2) Players don't necessarily want to help their opponents.
(3) There is a perceived benefit to know the rules (and thus, certain interactions) better than your opponent.
(4) Questions are sometimes articulated in a way that isn't clear or even correct.
(5) Magic is a social game, but people do play it competitively.
(6) Understanding of the rules isn't complete.
(7) Cards don't always say what they actually do.
(8) Players want to know precisely where the line is.

…and derived information at Competitive REL represents the “lowest common denominator” when it comes to that “line” for communication.

I agree with Shawn and Sebastian, that in many cases it is far easier to just answer the question you've been asked, rather than try to dance around the answer. And, socially, most people will cooperate. “Cards in hand?” *player fans out* “Two.” “Okay, thanks.” Even something complicated like a Tarmogoyf's P/T is something that most people will answer. And will do their best to try to help out.

The situations where players will go the opposite direction is usually in a situation that is very close, and the decision is very relevant. Sometimes one that decides the match. And if the player believes the opponent could make a mistake, then yes, they stop being as helpful as they may have been earlier in the game. Is that reasonable behavior? Should we require the player to be as helpful as they were earlier in the game in this situation, thus making it seem like they “have to help” their opponent?

Further, it can be problematic to state in policy the specific kind of question that can be asked. Should we require people to answer “What does that do?”-type questions? Or, should we require people to answer simple “Does that have *insert ability*?”-type questions? How do you draw that line?

I'm not saying I'm thrilled with the policy, but as a bare minimum behavior, I'll accept that we allow “incomplete” or “partially correct” answers, which aren't exhaustive. Whether the decision to not provide an exhaustive answer is because of ignorance at the question, for strategic reasons, or perhaps even language barriers, I think that's a reasonable expectation. Do most players go above the line on this behavior? Yes. But, I can see the reason why the bar is lower as being sometimes necessary.