WotC_BenFinkel

WotC_BenFinkel



25 Apr

Comment

Originally posted by mooseman3

This bug is easily reproducible:

  • Play a nonhuman creature and give it a menace counter (in this case [[Fertilid]] and [[Blood Curdle]])
  • Mutate [[Cavern Whisperer]] onto it, so it gains the menace ability a second time.
  • Cast [[Heartless Act]] and remove the menace counter. The creature will now be able to blocked by a single creature.

Note that I've only tested this when the opponent casts Heartless Act during the attackers step, just in case that's the only time this happens. It's also somewhat likely that this logic applies to other keywords as well. I did check with the MTG Judge Chat and they agreed this is a bug, not a rule I'm misunderstanding.

edit: This bug was formally reported here if you want to vote it up to hopefully get this fixed soon.

To clarify, you have to Mutate a menace creature on top? Having menace naturally doesn't seem to result in this bug, I'll try again with mutating. #wotc_staff

EDIT: Yes, mutating seems to do the trick. Let's figure out what's going wrong.

EDIT 2: Alright, got it! Here's what's going on. Whenever we detect that you've gained a new ability, we usually kick off that new ability's continuous effects. However, multiple different places in the code can attempt to start those continuous effects (creating attachment relationships and redestining zone transfers are some somewhat esoteric examples), and we don't want to double-dip starting those continuous effects, so we see if there are any already from this ability.

Menace makes a BlockedByMinCountQualification, which is a continuous effect. When you add menace (say, Menace2, the one from Cavern Whisperer) to something that already has it (say, Menace1, from the Menace counter) we were erroneously thinking we'd be d...

Read more

24 Apr

Comment

The rule for predicting cost adjustments accidentally removed all checks for zone constraints (instead of just the ones about "this spell" - it's not a spell until it's on the stack!). So basically Mystical Dispute is predicting it cost 2 less if there was a blue card anywhere in the game. Oops! I have a fix, but it probably won't be live for a while, especially as it doesn't affect gameplay much (just gives you stops if you can afford U and your opponent controls a spell...) #wotc_staff

Comment

Originally posted by mokomi

For "A mutated Cavalier of Thorns should no longer be able to target "itself" if you chose to exile it when it dies."

Wait, I thought that each card has that ability. E.G. A Cavalier of thorns mutated with a greathorn. Shouldn't both the Greathorn and the Cavalier choose to exile cards? Can the Greathorn choose the Thorns?

Or am I thinking wrong as in the bug the Greathorn can choose the Greathorn.

According to my chat with MTG rules manager Eli Shiffrin, "another" on Cavalier of Thorns means "other than any card that constituted this permanent". I suppose a Melded creature that became a copy of Cavalier of Thorns would have had the same issue - y'know, if Eldritch Moon were in Arena. #wotc_staff


22 Apr

Comment

Originally posted by freestorageaccount

Re: there being two checks for legality, kind of like with targetting, would a counterpart companion called Big Lurrus that instead read "converted mana cost 3 or higher" not let me cast a 3-point [[stonecoil serpent]]? Because the spell would satisfy that but as a card it would not, so that stage is never reached?

That would probably need some new semantics in the prompting-time version of the noun to mean "or has X in its cost", as that could POTENTIALLY become a 3+ cmc spell. And we'd need to ensure the X-bound logic enforced a MINIMUM X choice. Let alone the autotap concerns... #wotc_staff

Comment

Originally posted by darkslide3000

I'm somewhat surprised they wrote it, actually, considering that Magic has a very well-defined set of rules. I bet what happened here is that their Card object has a GetCmc() method, and they just implemented that to always count X as zero... which I guess could easily happen to most of us if we tried to write this code from memory. But if you look at the comprehensive rules, there is a whole block of rules about how to calculate CMC for every eventuality, and it very clearly says

202.3e When calculating the converted mana cost of an object with an {X} in its mana cost, X is treated as 0 while the object is not on the stack, and X is treated as the number chosen for it while the object is on the stack.

You would think for a product this big that has to really get every edge case right they would basically just translate the comprehensive rules into code 1-to-1 for every game mechanic thing they implement (and ideally have an L4 judge with ...

Read more

Consider that Lurrus is the first card on MTG:Arena that both allows you cast a spell constrained by its CMC and doesn't provide an alternative cost for doing so (unlike cards like [[Electodominance]]). So right off the bat we're talking about something new, but that looked really familiar - I'm not surprised that our testing strategy missed this case.

In order to even begin proposing casting the spell, some rule or ability needs to grant you permission to do so. Lurrus' code sees a card in your graveyard with CMC <= 2, and so it proposes it, thus you can start casting it. We need an entirely different step in the process, and thus entirely different code, to then step in at the choose-X step to enforce that the resulting CMC is STILL <=2. That doesn't come "free" from just "implement the comprehensive rules 1:1" - Lurrus' "you have permission to cast a CMC <= 2 spell" code needs to check in again there. Fortunately, the fix makes it so all CMC-capped casting permissions w...

Read more

21 Apr

Comment

Originally posted by Dall0o

As a dev myself, I have a love/hate relationship with this kind of bug. Stressful but fun too. Thanks for people for breaking software and posting issue after.

Nailed it. Release bugs are the worst from the stress perspective - we have such limited time to fix in time for adequate testing before the "fast follow" release! The puzzle-solving aspect is one of my favorite components of this job. #wotc_staff

Comment

Originally posted by into_lexicons

thank y'all for your hard work on stuff like this. these are thankless tasks but i wanted you to know it's appreciated.

Thank you! Fixing these things is fun (of course, not as fun as getting them right in the first place!). #wotc_staff


19 Apr

Comment

Cards you see in your opponent's hand aren't "real" cards - internally, we call the object a RevealedCard. They get created when a known card goes to a hidden zone, and are removed when a card with the same appearance leaves (or the contents of that zone become unknown in some other way, like a Brainstorm effect). RevealedCards should have no impact on actual gameplay, such as effects that count the number of cards in a hand, or selecting a card to discard.

We didn't properly handle making RevealedCards for tokens that have been mutated; they make RevealedCards because the object-on-the-battlefield was a "real card". Clearly, that should be examined on a component-by-component basis. We should fix it this week, I hope, but know that it's only a visual problem. #wotc_staff

Comment

Thanks for the report. I have some ideas about what's happening here, I'll look into then soon. #wotc_staff


18 Apr

Comment

This is fixed in the upcoming release hopefully coming this week. The parse tree got flipped around, and now it basically thinks it means "search your library for a black creature card in your graveyard with cmc less than the number of creature cards (anywhere in the game)." A card can't be in your library and graveyard! #wotc_staff

Comment

This is fixed for the next release, hopefully coming this week. In the meantime I recommend manually placing stops when you want to cast it - the game will allow you to do so, it just mistakenly thinks you can't afford it. #wotc_staff

Comment

This is fixed in the next release, which I'm hoping is later this week.

Funnily enough, it is not more expensive to target the older Skybonder, just the newer one. #wotc_staff

Comment

Originally posted by wotc_kale

Thanks, ShueiHS!

We've been trying to get some QOL improvements in every patch. They're not always something everyone notices, but I'm glad you enjoyed this one.

I take a lot of notes from what people say in threads like these and try to improve things where I can. Thanks for the feedback!

Kale is my hero! #wotc_staff

Comment

Thanks for the report, this was an interaction I hadn't thought of. I checked with our paper-MTG rules folks and they confirmed that indeed, a mutated creature only earns you one land with Settle the Wreckage, so I'm fixing it! We hope to have an engine update sometime next week, so I hope the fix will be live then. #wotc_staff


14 Apr

Comment

Originally posted by mooseman3

Just to be clear, is this a specific issue with the card or the issue where a disconnect/concede won't notify the user in any way?

It seems to have been an issue with how the client understands the messaging about how cards are attached to other ones. I didn't work on this bug, so take this with a grain of salt, but I think Journey to Eternity transforming and thus stopping being attached to something caused the next time we made a new attachment to be handled incorrectly. #wotc_staff

Comment

Looking at our issue tracker, it seems this bug has been fixed for the Ikoria release. #wotc_staff


13 Apr

Comment

The event should be live again now! #wotc_staff


01 Apr

Comment

Originally posted by coffeemiteofglut

Tnx for replies. Wasn't aware of hyphen bug.

Fortunately, it should be fixed with the Ikoria release. #wotc_staff