|
9 months ago ::
Sep 13, 2012 - 2:46AM
#1
|
Date Joined:
Mar 16, 2004
|
A year or two ago, I created a combo that proves Magic to be Turing complete. I put up a little website about it. Now it's doing the rounds of the internet ( Reddit, BoingBoing, Kotaku, Metafilter, Slashdot, Mark Rosewater's Tumblr), so I thought I really ought to post it on the official Wizards boards. The basic idea is to assemble a universal Turing machine out of Magic cards, so that a massive cascading sequence of triggered abilities occurs which simulates computation of a Turing machine, without needing any interaction from the players at all. It's mainly a load of ETB and LTB triggers, with the core being six copies of Teysa, Orzhov Scion (coexisting via Mirror Gallery ), all extensively hacked with Artificial Evolution and Mind Bend . The questions I had to answer include: Q: How do you make a tape extending arbitrarily far in two directions with a "step left" and "step right" operation? A: The tape to the left of the head is a series of Ally tokens, one 1 toughness away from dying, one 2 toughness away, etc; and the tape to the right is a similar series of Zombie tokens. To move left, we give all Allies +1/+1 and all Zombies -1/-1 (creature types adjusted as appropriate ). This causes the smallest Ally to die, and a different ability triggers depending on what colour it was. Q: How do you simulate the vital Turing machine concept of multiple states? A: Phasing! All the Teysas are enchanted with Teferi's Curse or similar, and we arrange for the machine to cast Time and Tide whenever it needs to change state. Q: How the heck do you cast an instant as a triggered ability? A: Chancellor of the Spires , with help from Aether Flash , Skirk Drill Sergeant , Wheel of Sun and Moon , False Dawn , Artificial Evolution , and lots of Engineered Plague . Simple, right? Q: Can you really prevent players needing to do anything? A: ...Almost. The key operation "Give all Allies +1/+1" is accomplished via Kazuul Warlord , which unfortunately has a "you may" in his rules text. So the cascade of triggered abilities isn't completely requiring no input from the players: one player has to effectively keep saying "yes" every time Kazuul Warlord or Chancellor of the Spires asks him "do you want to do this?" Another player has to similarly keep saying yes every time Skirk Drill Sergeant asks him the same question. But other than those "always say yes" operations, there's no interaction by the players required at all. Q: Does this mean we can implement Magic Online inside another game of Magic? A: Theoretically, yes.  But it might take quite a long time to get past the loading screen. Full details at www.toothycat.net/~hologram/Turing/
|
|
|
|
9 months ago ::
Sep 13, 2012 - 3:18AM
#2
|
Date Joined:
Sep 23, 2011
|
I actually thought about this a week ago. I was thinking "Magic cards interact with each other, perhaps you could build a computer out of a combination of them." But I almost instantly abandoned the idea because I remembered that Magic was able to emulate sentience with a single card .
|
|
|
|
9 months ago ::
Sep 13, 2012 - 6:11AM
#3
|
Date Joined:
Dec 15, 2011
|
ALEXTFISH:
So what do you plan to use it for?
will parts of it be able to work better than a classical turing machine?
Or is it just something you wanted to figure out and did?
|
|
|
|
9 months ago ::
Sep 13, 2012 - 7:30AM
#4
|
Date Joined:
Mar 16, 2004
|
ALEXTFISH:
So what do you plan to use it for?
will parts of it be able to work better than a classical turing machine?
Or is it just something you wanted to figure out and did?
Ha! No, I don't think it's useful for anything much. Turing machines are only really useful as theoretical ideas anyway. It's incredibly cool for people who know about what Turing machines and Magic are.
In particular, I think it's the first demonstration of Turing completeness in the rules of any tabletop game. I don't think there are any other board games or card games whose rules accommodate the complexity required to make a Turing machine (and I've played a few hundred).
I created it basically because someone asked me "Is Magic Turing complete?" I figured the answer must be yes, and started to try to prove it, and eventually (with help from a few friends) came up with that intricate combo.
I suppose one possible way to use it is that there are certain questions about Magic that it proves, or proves to be unanswerable. For example, the question "is this an infinite loop?" is impossible to tell if the loop in question is calculating, say, whether every even integer greater than 2 can be expressed as the sum of two primes (or any similar unanswered problem). So we've proved there are some game states in Magic where it's impossible to tell whether the game is a draw or not.
|
|
|
|
9 months ago ::
Sep 13, 2012 - 8:06AM
#5
|
|
|
You are my favourite Magic player for doing this.
L1 Judge
|
|
|
|
9 months ago ::
Sep 13, 2012 - 8:19AM
#6
|
Date Joined:
Mar 23, 2006
|
So we've proved there are some game states in Magic where it's impossible to tell whether the game is a draw or not.
oh, you.
120.6. Some effects replace card draws.
|
|
|
|
9 months ago ::
Sep 13, 2012 - 8:26AM
#7
|
Date Joined:
Mar 23, 2006
|
also if you let yourself use un cards, magical hacker on noxious ghoul lets you play the role of warlord without involving a "may". doesn't eliminate the player choices but reduces them. or, in real magic world, put out a second noxious ghoul of each type and a goldnight commander . then you have each creature getting +1/+1 and each non-X creature getting -2/-2, resulting in each non-X creature getting -1/-1. with no choices made.
120.6. Some effects replace card draws.
|
|
|
|
9 months ago ::
Sep 13, 2012 - 8:32AM
#8
|
Date Joined:
Mar 23, 2006
|
also you can reduce them by finding a way to sacrifice/kill the chancellor and having mikaeus the unhallowed in play. (with a way to remove the counter, obviously, but that's trivial.) I don't know a way to remove the chancellor decision though, that one's tricky.
120.6. Some effects replace card draws.
|
|
|
|
9 months ago ::
Sep 13, 2012 - 9:33AM
#9
|
Date Joined:
Aug 12, 2012
|
Good stuff. The concept of the stack where each action resolves in sequence is the same as the stack an operating system's kernel works with, right?
Magic the Gathering Adventures Blog http://mtgadventures.blogspot.com/
|
|
|
|
9 months ago ::
Sep 13, 2012 - 10:02AM
#10
|
Date Joined:
Apr 28, 2008
|
or, in real magic world, put out a second noxious ghoul of each type and a goldnight commander . then you have each creature getting +1/+1 and each non-X creature getting -2/-2, resulting in each non-X creature getting -1/-1. with no choices made.

I haven't looked at all this stuff, but isn't the potential to stack the triggers relevant in this scenario?
Everything Mown does is elegant.

|
|
|