General CBS

Speaking of Cruel Ultimatum has anyone seen any wild games with it during Outlaws of Thunder Junction drafting?
I drafted it a couple times but the mana cost is really rough so it mostly ended up in the sideboard. On a similar note though, I played the qualifier play-in on the weekend and had a sweet dimir control deck with Rush of Dread. I had a grindy game against an Orzhov deck which I won by casting it with all modes turn 15 or so and flashed it back with Slickshot Lockpicker with all modes the following turn! Not the same as ultimatum but it felt a bit similar ;).
 
Are those specifically meme numbers?
no but yes. they're roughly the sizes i want for the two projects, so it's barely a fudge to get a fun spreadsheet gag out of it
I've only used it as a testing thing and one time I gave a Cruel Ultimatum to the guy who loves it.
makes sense, that latter reason especially is freakin' sweet
I think this has some merit, even if it's as simple as seeding a land into each pack or something, but it requires more sorting work, so I probably never will.
yeah, hence why i shied away from doing it with the large(r) cube... it's already large, adding layers of complexity like seeding seems like such a hassle
 
Speaking of Cruel Ultimatum has anyone seen any wild games with it during Outlaws of Thunder Junction drafting?
Images stolen from Discord posts elsewhere, since I didn't take any other screenshots at the time:

A tale of Cruel Ultimatum, the perfect crime, in two parts.

eKXKjZ4.png


qzp2byt.png


I have experienced many things, but I had not truly lived until I triggered Marchesa with Cruel Ultimatum, got a second Cruel Ultimatum... and almost lost anyway (to decking, obviously - not playing the Raven was a misplay even if I thought I had a few too few crimes for it)
 
Unfortunately there isn't an easy way of doing that short of manually going through all of the cards that cost 8+ mana (which isn't actually that bad - there are only a couple hundred of them and the cards that do what you want are pretty obvious). This is a side-effect of how the Scryfall team chose to store cards on the back-end, combined with there not really being useful tags for this kind of thing.

The closest I could get to cutting down the search was this hairy pile of nonsense:

mv>7 (fo:"may cast this" OR fo:"discard ~" OR fo:"discard this" OR fo:"rather than pay ~" OR is:split)

and that's probably still missing stuff.
 
The search I gave you actually does find morph cards like Akroma. Annoyingly, an earlier search string I tried did find suspend cards and prototype cards, but I broke that trying to be clever (I was trying to filter out cascade cards and Eldrazi with cast triggers...)

...

Just in case anyone is curious for an explanation of why this kind of search is annoyingly hard, one of the fun quirks of how Scryfall works is that it actually doesn't store that much game information (most of the information it keeps track of is stuff like "which image file should I grab when I show this printing of this card?", "what formats is this legal in?", or "how much does this printing cost in euros?"). Roughly speaking, the in-game information it stores boils down to...

  • The card's name
  • The card's full mana cost
  • The card's mana value (which scryfall still calls cmc internally, because it'd be really annoying to update it)
  • The card's typeline.
  • The card's colors and color identity, stored separately.
  • The keywords on the card.
  • The card's full oracle text, with all reminder text (a normal o: search appears to just ignore stuff between parentheses).
  • Type-specific extra values, like P/T or loyalty.
  • Any "related" cards (the faces of a DFC, any tokens the card might create, any cards that create a particular token...)

(I won't get into how it stores split/dfc cards right now because it's a bit weird, and includes some fields that are a complete mystery to me. And if you're wondering why I didn't mention rarity or the artist, that's because those are printing specific and the rest of this stuff isn't).

The key thing to notice here is that Scryfall hard-codes (card-codes, if you would) pretty much everything you'd want to search for, which makes sense when you consider that it has something like 90 thousand cards stored on there. But it does have some side effects:

  • You can't search for cards based off of their literal printed text. This normally doesn't matter, but it is annoying if you want to, say, only find printings that include reminder text.
  • The mana value that Scryfall stores is the card's mana value in zones other than the stack (which is the correct mana value to store). This makes split cards a pain in the ass. You want to know all of the split cards that have a half that you can cast for two mana? Search for every split card and manually filter the list.
  • Similarly, Scryfall doesn't store any costs other than the one in the top right corner of the card (split cards/mdfcs do get an exception here). Again, this is a sensible decision on Scryfall's part, but it means that conceptually simple searches like "I want to find every card with a cheaper flashback cost than its normal cost" or "I want to find every card that costs less than three mana to cycle" would be an absolute nightmare to actually do (I'm pretty sure that first one is effectively impossible unless you download the dataset and write your own code to search it and parse "converted mana costs", which sounds like too much work).
  • Thanks to how cards store the list of cards related to them, it'd be viable for Scryfall to let you do something like token:zombie to find every card that makes a Zombie token. The catch (and probably the reason why you can't do that) is that Scryfall only includes related cards that exist, meaning that a not-insignificant number of older cards don't make tokens as far as Scryfall is concerned.
 
But it does have some side effects:

  • You can't search for cards based off of their literal printed text. This normally doesn't matter, but it is annoying if you want to, say, only find printings that include reminder text.
  • The mana value that Scryfall stores is the card's mana value in zones other than the stack (which is the correct mana value to store). This makes split cards a pain in the ass. You want to know all of the split cards that have a half that you can cast for two mana? Search for every split card and manually filter the list.
  • Similarly, Scryfall doesn't store any costs other than the one in the top right corner of the card (split cards/mdfcs do get an exception here). Again, this is a sensible decision on Scryfall's part, but it means that conceptually simple searches like "I want to find every card with a cheaper flashback cost than its normal cost" or "I want to find every card that costs less than three mana to cycle" would be an absolute nightmare to actually do (I'm pretty sure that first one is effectively impossible unless you download the dataset and write your own code to search it and parse "converted mana costs", which sounds like too much work).
  • Thanks to how cards store the list of cards related to them, it'd be viable for Scryfall to let you do something like token:zombie to find every card that makes a Zombie token. The catch (and probably the reason why you can't do that) is that Scryfall only includes related cards that exist, meaning that a not-insignificant number of older cards don't make tokens as far as Scryfall is concerned.
I have such complaints about the first two on this list.

First one is a thing that WotC has already officially released in Gatherer! Better yet, it even has distinct "printed text" for each printing of the card. Why can't Scryfall scrape it literally one time, then once per main set going forward? It's not something that ever updates over time so it should require zero maintenance, just doubling their storage space and adding one search variable. Not trivial, but not exactly dificult, or so it seems to me. Perhaps they're legitimately constrained by storage space, maintainability, or coder time.

Second one is absolutely wild, since all of the following are true:
1) Scryfall has "is:split" as a search condition that already exists (returning the physical card, showing both halves, but not each half individually)
2) Individual halves of split cards correctly exist as two distinct cards in their database (see: Fire and Ice)
3) A split card's Oracle name is both of those cards, but if you click on Fire // Ice it will actually take you to a search that also finds Sword of Fire and Ice, and that also holds true for Fire/Ice and all other formatting things I swear I didn't mess it up in that exact way
4) Searching "fire / ice" in Gatherer returns exactly two results named "Fire // Ice (Fire)" and "Fire // Ice (Ice)" so this is also an already solved problem :(
 
Second one is absolutely wild, since all of the following are true:
1) Scryfall has "is:split" as a search condition that already exists (returning the physical card, showing both halves, but not each half individually)
That's because, to use Fire//Ice as an example, there's a "layout" value in the JSON that cards are stored as. "is:split" is actually just a search for cards with "layout":"split".

2) Individual halves of split cards correctly exist as two distinct cards in their database (see: Fire and Ice)
They don't, actually! If a card's layout says that a card has multiple "parts", there's a special "card_faces" value that stores an array of card parts representing each bit (this is the weirdness that I mentioned). Each of those card faces only stores the information that is different from the full card — since the different parts of a split card don't have different mana values anywhere other than the stack, Fire//Ice's halves just inherit the "cmc":4 from the overall card.

Fun weird fact: the second half of a split card seems to always have an extra "flavor_name" value, and I don't know why. It's always blank!

3) A split card's Oracle name is both of those cards, but if you click on Fire // Ice it will actually take you to a search that also finds Sword of Fire and Ice, and that also holds true for Fire/Ice and all other formatting things I swear I didn't mess it up in that exact way

The canonical name for the full card is "Fire // Ice" in Scryfall. Note the spaces and the quotes — if you don't put in the quotes, Scryfall strips out the slashes and interprets them as whitespace (this is pretty normal input sanitization stuff).

...

If you're wondering, yes these were all fun things to learn while trying to Scryfall's database while on a deadline.
 
Yes. This is exactly why I run 3 mv mana rocks. Also, because 3 mv mana rocks kind of suck, I designed a whole bunch of custom 3 mv rocks myself.

I love this approach and I’ve said it many times before. It’s such a clever way to have the rocks be super interesting with small upsides instead of hyper effective with pseudo downsides. They can even be build arounds.

However it will require a few things: Cube cannot be super high powered. And owner should be open to the idea of customs. I guess it can be executed without customs but the best ones are designed by Onde and not by Wizards lol.
 
@ravnic
Do you, or anyone else, feel that running Talisman/Signets takes away from green's ramp identity?

No. If certain conditions are met; most importantly, greens ramp must be better (which can be achieved by being cheaper or more synergistic or more explosive). Second, it needs to be more plentiful (you should be able to build a ramp deck without mana rocks but not without green).
 
Top