Monday, May 23, 2011

Rules Fixes: The Legend Rule

Honestly, sometimes I wonder if Mark Rosewater's trolling his own column. In case you missed it, a couple weeks ago MaRo effectively threw down the gauntlet and challenged the community to design a better legend rule. Even if he didn't quite do it on purpose, his post started 27 pages (and counting) of forum chatter and managed to elicit a response from Tom LaPille of Magic R&D and "Latest Developments" fame. As a result, I've grotesquely over-invested my time on this troll-foolery, and here's what I've come up with:

Current Misconceptions about Legends

First, whatever MaRo might say, the legendary supertype isn't the same kind of downside mechanic as the other examples he cites in his column. For reference, his definition of a downside mechanic is one that allows a player to get a more powerful spell at a cheaper cost but at the price of something else. But this isn't the case with most legends. By playing a legendary card, you're actually getting something that is less mechanically powerful than another card at a comparable rarity. For example, look at the current mythic legends - they simply are not as good as the non-legendary mythics that exist at comparable costs. The only card where the legendary supertype is being used as a mechanical limitation is on Mox Opal (which in this case is the exception that proves the rule). The real trade-off in playing a legend isn't that you get something more powerful, it's that you get a card that makes you feel better about playing it, since it gives you a greater personal connection to the story of the game.

Second, there's a deep undercurrent of resentment about the price of chase mythic rares, which is manifesting in strange ways on this current topic (for example, the idea that legends should be restricted). But regardless of how you feel about the price of mythic singles, this doesn't actually have anything to do with the legend rule. In fact, legendary cards are relatively cheap by historical standards (especially when you consider how much demand there is for them for EDH). Legendary cards simply don't make up much of the cost of most competitive decks, which should be obvious with a little reflection — if there's pressure to limit the number of card played in a deck, then there's less demand for that card, and the price will be lower as a result.

And finally, even though the legend rule and the planeswalker uniqueness rule are pretty much the same, the two card types get significantly different handling during their creative production process. So even if there's a change to one, it doesn't necessarily mean there need be a corresponding change to the other. Again, people want to use the legend rule as an excuse to nerd-rage about Jace the Mind Sculptor (who, by the way is an outlier even by the standard of high priced mythics), which is probably irrelevant to the current discussion.

The Problem with the Problems

As is, the current legend rule is very brief. In its entirety:

704.5k. If two or more legendary permanents with the same name are on the battlefield, all are put into their owners’ graveyards. This is called the “legend rule.” If only one of those permanents is legendary, this rule doesn’t apply.

While this has very broad effects on gameplay and deck building, the rule itself is clear, concise, and requires no other prior knowledge of the game. By contrast, the rules for cycling (not even including typecycling) are twice as many words, and assume you already understand the nuances of activated abilities. So the problem with fixing the problems created by the legend rule is that there's no satisfactory way to fix it without creating more complication than the existing rule. Not that this has prevented anyone from trying. Here's a run-down on the various fixes that I've seen proposed in the last two weeks. I don't think any of them are right, but feel free to chime in as you see fit:

Choice A: Get Rid of the Legend Rule

This simply does not fit the criteria set by MaRo. If there's no legend rule, then there's no reason for printing the legendary supertype, and without it, there's no way to represent unique, non-planeswalker characters within the game. One of the key elements of Magic as a game is that it has a definite storyline, and with that story comes certain unique characters, just like in a novel, video game, or movie. To make this uniqueness apparent in the game, it must have some kind of mechanical representation, which means the legend rule could either be updated or replaced, but it couldn't be completely cut.

Choice B: Restrict Legends

The argument is if you restrict legends to a 1- or 2-of in a deck list, then it will minimize both the problem of drawing multiple legends and the problem of what to do when your opponent has a copy of the same legend on the table by limiting the chances of seeing legendary permanents. But this is a response bordering on non sequitur. Our goal is to make it easier and more rewarding to play legendary cards. Restricting them does the opposite: making it harder to draw cards makes players less likely to be able to play the cards they want to play.

It also doesn't make sense from a business perspective. Wizards' goal is to get customers to buy as much as they're reasonably able. Legendary cards are already problematic in this sense, because currently there's no game-based reason to fill out a set of any one legend (maybe there are collectors who buy as many of an individual card as they can?). I realize this line of argument makes some people angry, but it's worth repeating: Magic as a game survives because it's a profitable brand for a major game manufacturer, and if it didn't keep evolving, it would lose the essence of what makes it a unique game. Also, if corporations didn't strive for continuous improvement, Magic's R&D team wouldn't be interested in improving the rules of the game, and we wouldn't be having this discussion.

Choice C: Make Legends Better

Since the legendary type is a downside mechanic,  some argue that printing better legendary cards provides more compensation for their rules awkwardness. As with restricting legends, there's a sort of logic here, but it doesn't lead to a helpful answer. If all legendary cards (which are already rare or mythic) were so powerful that they became competitive-critical cards, it would certainly encourage players to play more legends, but that doesn't necessarily mean they'll have fun playing them. (It's an effect like Cranial Extraction based decks — just because they were a good choice to play didn't actually make them fun.) If more legendary cards start seeing regular play, then there will be a higher incident of legend-induced tensions, since you'd be more likely to draw multiples of your legendary permanent and more likely to face a mirror match. Therefore, this exacerbates the problems with the legend rule, rather than actually fixing them.

Choice D: Enchant World

Another popular answer was to simply have legendary permanents destroy all older copies when they enter the battlefield. What most proponents of this idea missed was that there's already been a class of cards that worked this way, enchant world, and that its rule is still hangin' out in the book:

704.5m. If two or more permanents have the supertype world, all except the one that has been a permanent with the world supertype on the battlefield for the shortest amount of time are put into their owners’ graveyards. In the event of a tie for the shortest amount of time, all are put into their owners’ graveyards. This is called the 'world rule.'
Many of the original world enchantments were miserable lock-down style cards, so being able to run out a world of your own was a way to break the opposing lock (I wasn't playing Magic in that decade, so I don't know how practical that actually was). However, for the purpose of refining legends, I think using something like the world rule would be too strong an answer for opposing legends. It would create a disincentive for players to act first, since neither player wants to run their legend out to be destroyed, which in turn inhibits the momentum of the game. Neither player will be willing to commit their best creatures to the battlefield, the game will begin to drag un-funly (yes, this is a word) as players are forced to clutch their best cards, and in the worst case scenario, creates a draw-go stalemate that would take twenty minutes to resolve. Or it would be like a Umezawa's Jitte-wars redux: players would run legendary cards just to destroy other players legends, and in the odd chance you didn't see the mirror card, you'd get to punish that opponent with your awesome legend. 

Choice E: Cycling/Grandeur

This got quite a bit of discussion on the mothership forum, and our very own Jay Treat and Jonathan Woodward had an extended conversation on twitter as well. The basic idea is that instead of directly making better legendary cards, give them some sort of extra ability that can be used if there's already another copy of that legend in play. The ensuing legendary subtype might have a rule along the lines of: "As long as there is a permanent in play that has the same name as this card, this card has Cycling 2." The other idea in this vein was some sort of activated ability unique to legendary cards along the lines of the Grandeur legends from Future Sight.

The first problem here is that even if you stick with a simple cycling (instead of fooling around with a more complicated ability), this creates a lot of hidden text. Hidden text apparently isn't a deal-breaker for the Magic R&D crew (after all, planeswalkers have several paragraphs contained just in their card formatting), but honestly, I hate it. I hadn't realized it until I was playing my beast deck with Garruk Wildspeaker against a friend who had been away from the game for a few years, and playing the planeswalker was like playing Calvinball. It felt like cheating. My personal preferences aside, we're trying for an elegant solution, and this doesn't seem to fit the bill. The other problem with this fix is that it only answers half the problems raised by the legendary type in the first place. But it doesn't help me work around my opponent's copy of that legendary card.

For example, if I already have an Akroma, Angel of Wrath in play and then draw a second one, having another ability for the card besides putting into play is obviously a bonus. But if my opponent has an Akroma, and I have one in my hand, then I need to accomplish two things. First, I've got to get my opponent's win condition off the table, and then I need to get a win condition of my own into play. So if I cycle my Akroma, maybe I draw into a Wrath of God, or perhaps she has some sort of Grandeur-style ability that can destroy a permanent, which I can use on the enemy angel. But neither of these options give me something that I don't already have under the current rules, since I can already use my legend as an answer to another legend. In any event, I'll still need to draw a new win condition, which again, is the same as in the current rule.

Despite these problems, I'd still consider this as a possibility. Unlike the first four answers discussed, it does actually provide an answer to some of legend's problems, I'm just not convinced that the answer is worth the added hassle.

Choice F: Oblivion Ring, with a side of Phasing

The last answer that saw a lot of forum discussion was to treat all legendary permanents like they were Oblivion Ring. Practically speaking, this would play a lot like the world supertype effect, but instead of sending older copies of the legendary card to the graveyard, it would remove them from play. For example if my opponent had an Isamaru, Hound of Konda on the battlefield, and I played my own Isamaru, my opponent's copy would be exiled (or otherwise sent out of play — there are some zoning issues to come). Then, if my opponent Shocks my Isamaru, my copy would go to the graveyard and my opponent would get his Isamaru back. Here's where the zoning issue applies: if my legendary creatures are bouncing in and out of play, they're going to trip assorted enters-the-battlefield/leaves-the-battlefield triggers, which may be an undesirable outcome. (For reference, the number of non-land legends that have these kinds of abilities is really small, less than 50 cards out of 500+ legends. And in practical terms, the only ones of these that actually see play are usual EDH suspects Godo, Bandit Warlord and Sharuum the Hegemon, and the ever noxious Vendilion Clique.)

Using phasing as an answer is doubly clever, even if some of the discussion of it was unintentionally so. First, having a permanent phase out also means that the permanent will phase back in at the beginning of its owner's turn, so effectively players could take turns using their legends. It also dodges the auto-draw problem that would arise if the new legend rule used O-ring as a base, where a player who puts a legendary card onto the battlefield with another copy already in play and a third copy in exile creates an infinite cycle of constantly exiling legends. (Since the exile loop would occur as a state-based action, there wouldn't even be a way for a play to break the loop with a removal spell.)

The resulting legend rule might look something like this:
If two or more legendary permanents with the same name are on the battlefield, all except the one that has been a permanent on the battlefield for the shortest amount of time phase out. In the event of a tie for the shortest amount of time, all are put into their owners’ graveyards. If only one of those permanents is legendary, this rule doesn't apply. When a legendary permanent phases in, all other legendary permanents with the same name phase out (refer to the appropriately modified rule for phasing).
This fix has the opposite conundrum as adding a cycling effect. It's a fair and fun way for dealing with my opponent's copy of a legend, but it doesn't give me much utility when I already have a legendary card in play and draw an extra copy. More significantly: holy cow Batman, the complications! I do intend to try play testing this variant. I'm not sure when I'll get a chance, but I'm going to print up some proxy decks of the Ravnica-era Boros Deck Wins and Ux control decks featuring Meloku / Keiga / Akroma and see if I can teach the "new" legend rule. But I'm not confident that the concept of phasing will go over well (after all, there's a reason it stopped seeing print).

Choice G: None of the Above

My preferred approach is to do nothing. In this instance, I don't think either Mark Rosewater or Tom LaPille grasped the entirety of the issue. I disagree with Tom's premise that, "it may be true that they preclude interaction between cards, but they do force players to interact with each other." The rules of the game provide the vehicle for player interaction, and if there are rules that prevent cards from interacting, then you're necessarily preventing some degree of player interaction. So in this respect, MaRo is correct — the current legend rule is a bit obnoxious. But I also disagree with Mark's notion that this is something that even needs "fixing." While E or F could at least be plausible fixes for some of the problems stemming from the legend rule, I don't think the costs of their added complications outweigh the benefits of modifying the existing legend rule. For the time being, I'm content to accept that not every facet of Magic is perfect (and never will be).


  1. Great review of the major solutions, Daniel.

    I have to say that even as it was fun to take MaRo's ideas at face value so that we could explore his challenge, there was some doubt around how big the problems he stated were. It was nice to see Tom's counterpoint.

    Similarly, it's been impressive to see all the ideas everyone has come up with as well as the reasoning behind what makes each one interesting and/or flawed.

    I think you're right that the best solution for the foreseeable future is to leave the rules as they are. I do see some design opportunities to smooth out legends on a case by case basis with abilities like grandeur, cycling and the like.

  2. "Un-funly" is not actually a word on any dictionary I could find. Neither is the term "funly".

  3. Only once did the problem of having multiple legends on the battlefield come up in a game I've been in. My opponent played his legend, then a few turns later played another copy of it. Turns out he had never even heard of the legend rule! (Unrelated note: that legend hosed my deck, so I ran a playset in my sideboard both as cheap removal and as a hose for the mirror match. Only was useful once, in a mirror.)

  4. Exactly two years after you guys post this, the Legend rules is changed. Never deny cool things?

    I personally like the change, it mitigates a lot of the feel bad situations the old one caused and does wonders for Commander.