MDY v. Blizzard Opinion

Legal commentators in the blogosphere (e.g. Nic Suzor, Technollama, Rebecca Tushnet, Venkat & Eric) have already offered some initial thoughts on the Ninth Circuit decision in MDY v. Blizzard. Since I talked about the district court opinion in this case in Chapter 9 of my book, I thought I'd post a few reactions too.

This post is going to be a bit on the long side, but that's only because the issues raised on this appeal are a bit tricky, meaning that I feel the need to lay a little doctrinal groundwork before getting to my thoughts about the case.

Though there was an interesting tortious interference decision in the appeal, I'm going to focus on the two copyright issues that were decided by the Ninth Circuit, one involving a claim that users of MDY's Glider program breached World of Warcraft (WoW)'s software license and the other claiming that users of MDY's Glider program violated the Digital Millennium Copyright Act (DMCA)'s prohibitions on circumventing technological protection measures that limit access to copyrighted works. This second claim focused on the operation of Blizzard's Warden program, which monitors a player's computer to see if it is running any unauthorized software.

The appellate court essentially found in favor of MDY on the licensing issue, reversing the lower court, and in favor of Blizzard on the DMCA-Warden issue, affirming the lower court. That adds up to a win for Blizzard. That win could be reversed, in theory, if MDY pursues further appeals.  An en banc review of the Ninth Circuit is possible and there's always the slim chance of getting the case reviewed by the United States Supreme Court.

I'll discuss the software licensing issue first.

1) Does violating a ban on botting make a user a copyright infringer? When a user plays WoW, some of the game's software code is copied into the random access memory of the user's computer. This "RAM copying" is treated as an exercise of the copyright owner's exclusive reproduction right pursuant to existing Ninth Circuit doctrine.

Whether the RAM copy doctrine makes sense has been subject to some debate. However, if we accept that premise, it follows that playing WoW entails making a copy of the WoW software. If you are making a copy and you are not licensed to make that copy, you're presumptively infringing copyright. It is worth noting that section 117 might seem to carve out an exception for software purchasers making RAM copies that are part of normal software use, but again, the Ninth Circuit's precedent (affirmed in this case) leads to the conclusion that section 117 doesn't apply in standard consumer software purchases.

So starting at that point, if there is a term in the software license that you fail to abide by, and you make a RAM copy while breaking the terms of the license, are you then engaged in copyright infringement? E.g., I grant you a license to use my spreadsheet program, but include in the license the condition that you must stand on one foot while doing so. You agree to my software license and then, while using the spreadsheet, you lose your balance. When your foot touches the earth, have you committed copyright infringement?

Or perhaps did you merely breach an independent contract with me that is unrelated to copyright, meaning I can only sue for breach of contract?

The difference in the two sorts of claims can be very significant. Breaching a contract gives rise to traditional remedies -- i.e., how were you damaged when I stood on two feet and used the software? Infringing a copyright, on the other hand, gives rise to statutory damages, which generally don't track the actual damages that a copyright holder suffered as a result of the infringement. Statutory damages are a much more powerful remedy.

The Ninth Circuit in MDY draws a distinction between licensing terms that give rise to copyright infringement liability and those terms that fall outside the scope of the copyright entitlement. This makes abundant sense. Software copyrights should not give their owners the power to essentially regulate all user behaviors during software use with the threat of statutory damages for any breach of the terms of use.

To give rise to the infringement remedy, it makes sense that a licensing condition should have some "copyrightishness" about it. By analogy, this is sort of like the touch & concern doctrine in the law of property servitudes, which requires real covenants (persistent social arrangements regarding land use) to have some "landishness" about them if they are going to be applied to downstream purchasers of the estate.

The precise term at issue in MDY is Blizzard's prohibition on unauthorized "botting" software. Glider is a botting program, so it pretty clearly breaches Blizzard's license. Essentially, a bot program turns your avatar into a robot. Using Glider means that you are not playing World of Warcraft -- instead you're letting the bot program control your avatar on autopilot, setting it to run around Azeroth and kill things and collect loot.

If you use Glider, it is clear that you are in violation of the software license of WoW. And MDY essentially conceded on appeal that if users are copyright infringers when they use Glider in this way, MDY is liable based on that conclusion as an accessory to the underlying infringement.

But, as it turns out, according to the Ninth Circuit's ruling, using bots in violation of the WoW software license actually does not make a user a copyright infringer. Why not? Because there's no nexus to the copyright law in the license prohibition against bots.

It's a little hard for me to neatly sum up the reasoning here, but the basic point seems to be that "how you play the game" does not really have much to do with making copies of the game software or recording the content, etc.

As an outcome, I'm fine with this. Still, I must confess I'm a little hazy about the rule. Supporting the court's view, it is true that the claim being made by Blizzard is not that Glider users make copies of its software or experience WoW without paying subscription fees. If this were a traditional video game, I think those facts would strongly support the lack of a copyright nexus. But WoW is not a conventional video game and the copyright at issue is not a conventional copyright. In fact, with regard to the DMCA claim (which I'll get to next), the Ninth Circuit opinion accepts the idea that Blizzard owns a discrete copyright interest in WoW's "dynamic non-literal elements."

So Blizzard is not only claiming protection of the game's literal software code or the various multi-media components (audio files, graphics files, animation files). The copyright asserted in this case extends to the user's experience of the game software when connected to the servers (which are in turn connected to the clients of the thousands of other players). The work accessed is "dynamic" because it is constantly evolving and "non-literal" because it is a fluid multimedia audiovisual experience rather than a conventional fixed text. Note that these dynamic non-literal elements are produced, at least to some degree, by the players.

In fact, what is particularly irksome to Blizzard (and to its subscribers) about Glider is not that Glider players get access to the world, but that that they fail to access the world. The Glider user doesn't actually want the "dynamic non-literal elements" of Azeroth. The paradigmatic Glider user is not even experiencing WoW when Glider is active.

So why is Blizzard complaining about Glider users who pay MDY not to experience Azeroth? Because this case, I think, has always been about gold-farming. Blizzard sued Glider, I think, because the software is useful for gold-farming and gold-farming disrupts the game economy, poses virtual property headaches for the company, and is deemed cheating by a substantial portion of WoW's user base.  If virtual gold is valuable and can be sold for profits, botting to collect it is a way to make money and Glider becomes a business tool.

But in the case, it seemed that neither side (for separate good reasons) stressed or deeply explored the connection between MDY's profits and the gold farming business. Given that the parties did not pursue it, the court certainly did not jump on it sua sponte.

Before moving on to the DMCA, I want to touch back on the condition/covenant distinction in the license.  Is it really true that the prohibition on botting has no nexus to Blizzard's copyright? Actually, I might argue that it does. The manner in which the Glider player acts (or more accurately, fails to act) actually does shapes the nature of the "dynamic non-literal" elements that are experienced by the other players. The Glider player is, essentially, a bad authorial collaborator, making WoW worse for other subscribers.

Is that a copyright nexus? Certainly not a traditional nexus, but I do think an argument could be made.

Still, as I said, I'm generally pleased with the outcome here. Whatever the coherence of the court's explanation on these facts, the public is certainly better off in a world where every minor violation of a clickwrap license won't put you on the hook for copyright infringement.

2) DMCA 1201(a)(2)

Although the DMCA is another copyright claim, it is actually a much different claim involving a much different law.

The software licensing claim described above follows this logic:

1) Blizzard says no botting or you can't make RAM copies,
2) players who use Glider make RAM copies,
3) these players infringe copyright,
4) MDY is liable for helping them do that.

As I said above, the Ninth Circuit finds for MDY (and Glider users) by failing to agree with Blizzard at step 3.

The DMCA claim concerns Blizzard's effort to control bots via technology. Specifically, Blizzard created a program called Warden that monitors the active memory of the user's computer. Warden scans the user's computer for signs of unauthorized software. When the user connects to Blizzard's servers and the game is running, Warden pokes around in the game memory. If it finds signs of something it doesn't like, it severs the player's connection to the servers.

Blizzard's DMCA claim follows a simpler legal logic:

1) Blizzard's Warden program is a technological measure that controls access to the "dynamic" WoW program,
2) MDY's Glider program is a tool that circumvents Warden.

Essentially, that's the whole claim. This is because the DMCA, at 17 U.S.C. 1201(a)(2) concisely prohibits trafficking in tools that "circumvent a technological measure that effectively controls access to a work..."

Because Blizzard relies on Warden's policing activities for this claim, it had to take the position (mentioned above) that Blizzard holds a discrete copyright interest in WoW's dynamic non-literal game play elements. This is because Warden controls access to those dynamic elements, but doesn't effectively control access to the rest of the client software. E.g., the user has access to the audiovisual elements of the game and to the literal software code even when the client is not connected to the servers. The user only gets to see the game be dynamic, though, when the client ties into the server.  This is when Warden does its police work.

So Blizzard's claim was that Warden is a technological measure that effectively controls access to WoW's dynamic non-literal work (the gameplay). The lower court accepted this claim. The Ninth Circuit accepted it as well, disagreeing with a Federal Circuit opinion that would have interpreted 1201 more narrowly, perhaps allowing MDY to prevail on this claim as well.

A large swath of the Ninth Circuit opinion is given to explaining what is fairly obvious if you read 17 U.S.C. 1201. Namely, section 1201(a) (at issue in this case) is directed at the circumvention of access-control measures whereas the separate provisions of 1201(b) outlaw trafficking in tools that enable the circumvention of rights protections. The court's descriptions of (a) and (b) are really quite nice, and I'll be sure to give them to my students in the future who get confused about the difference between 1201(a) and (b).

However, the Ninth Circuit uses the difference between 1201(a) and (b) to make a more radical, though not unprecedented, claim. Following what the Second Circuit did in the case of Universal Studios v. Corley, the Ninth Circuit determines that the 1201(a) "access" provisions are essentially a new right added to copyright law that are almost entirely separate from the traditional rights of copyright owners. Disagreeing with the Federal Circuit's decision in the Chamberlain (garage door opener) case, the Ninth Circuit rules that violations of 1201(a) can occur without the demonstration of any nexus (contra the discussion above re licensing) to copyright.

So, in essence, according to the court, it doesn't really matter that users are circumventing Warden with Glider solely to engage in a different sort of game play that is unrelated to copyright entitlements. The court itself stated that the botting prohibition lacks a copyright nexus. This doesn't help MDY evade 1201(a) we are essentially told, because that prohibition isn't about any that a user might do after evading Warden. The very act of evading Warden is the violation, even if subsequent actions have nothing to do with infringing copyright.

With that established, it is also established that MDY, by providing a tool that circumvents Warden, violates 1201(a)(2), even if the use of bots by users doesn't amount to copyright infringement.

Does that sound a bit odd? It certainly does to me.

Just like I disagreed with the Second Circuit when it decided Corley, I find myself in disagreement with the Ninth Circuit in this case. Whatever the legislative history of 1201 may say, it seems to me that 1201 is part of the copyright statute. Requiring that anti-circumvention violations bear some relation to copyright entitlements makes just as much sense here as it does with regard to the questions of software licensing. As many commentators have said over the past decade, treating 1201(a) as paracopyright entitlement essentially divorced from the logic of traditional copyright makes for both bad policy and bad copyright law.

Finally, I want to note that, for reasons that I explain in the book, I'm not unsympathetic to Blizzard's efforts with regard to prohibiting botting software. I'm actually okay, in this particular instance, with Blizzard winning this particular case.

My main concern is about the collateral consequences of this sort of DMCA decision. Copyright owners seeking to use the DMCA in this way are not all going to be like Blizzard going after botting programs. Few plaintiffs will be enforcing game rules with the general support of the user base.

Instead I predict that future cases that cite to this ruling will involve cloud computing companies and social network providers facing off against disruptive innovators -- and I'm pretty sure I'll be significantly less sympathetic to those efforts.  Increasingly, the software we are running is going to be connected to remote servers and policed by programs like Warden that want to keep tabs on what we're doing with our iPads and smartphones, etc.

If you think about it, the ruling in this case is essentially about technology replacing law.  Blizzard's contractual prohibition on bots has no copyright teeth.  Its technological effort to prevent botting, however, obtain very pointy copyright teeth from 1201(a).  Seems to me like an incentive for future cloud computing titans to build more comprehensive Wardens -- or even traitorware.

So there you have it. In sum, a mixed bag. (Lots of grays, a few greens, no purples.)

(Prior coverage here.)

Comments

[Humphrey Bogart]

1.

An interesting thought for the DMCA claim I don't think was ever brought up.

That is, Is MDY actually circumventing warden (as in, Are they modifying the warden process so that it operates incorrectly, much as a "crack" would allow access to a program without paying) or is MDY simply modifying its own process to be undetectable. Im not sure that the DMCA would cover that case. Is there DMCA grounds simply because warden was incompetent at detecting a advanced program? Imagine it in a game of cat and mouse. The DMCA says, you cant interfere with the cat at all. But is making the mouse super-camoflague and impossible to find grant the same protections?

Note, this discussion could be mute if MDY was in fact modifying warden but I have not heard any claims either way. I did hear that MDY developed a "shadow driver" which made the program hidden from the windows process list, but again- This is not 'modifying' warden.

I hope MDY's lawyer brought up this point as it seems he may be remiss if he didnt.

[greglas]

2.

Humphrey,

The court deals with this at footnotes 16-18 and the text accompanying. I think it would be a reasonable interpretation of the DMCA to require that a circumvention technology in some way "breaks" the protection of a work. But that isn't how courts have interpreted. If a program disguises itself to gain access, that has been ruled to be a form of circumvention -- sort of like forging papers to get access to a building. (I hate the spatial analogies, but I'm afraid they're deeply ingrained in the logic of 1201.)

What bothers me is not so much that the court finds liability because Glider evades Warden (that bridge has been crossed several times), but that the sort of access control involved here is actually post-facto.

It's sort of like a guard in a movie theater who walks around checking to see if you have a ticket stub -- does that guard effectively "control access" to the work? Well, yes, in one sense, but you're already through the door, so the access control occurs after access has been gained. This suggests that post-hoc policing of behaviors (notably, copyright-unrelated behaviors) can amount to a technological access control measure under 1201.

Ruling otherwise would have been a way to narrow the scope of 1201, but the court doesn't seem interested in pursuing that distinction.

[greglas]

3.

And if this wasn't enough legal commentary on MDY, Lionel Sobel provides several more links to peruse here.

[robusticus]

4.

Is there a reason 1201(f)(2) wouldn't apply? That exception places an infringement nexus and does not specify authorization, good faith or otherwise as 1201(g) does.

[andy]

5.

Anyway...in my opinion blizzard is the best.

[greglas]

6.

Hi robusticus,

Doesn't seem like MDY raised that exemption claim. I'll have to double check the briefs. While it looks plausible, I think the jurisprudence has interpreted that exemption narrowly, but I should do some digging around to refresh and update my knowledge on that...

Greg

[robusticus]

7.

Well, Lexmark, apparently. And an amicus brief in Chamberlain.

"Thus, even accepting the
validity of the concerns expressed in Chamberlain, those concerns
do not authorize us to override congressional intent and
add a non-textual element to the statute."

That sentiment is expressed throughout. My confusion is why the need to override congressional intent when the text is already there in 1201(f)(2)? Ignoring that part of the law seems to be a bigger override.

Further it seems the logic used to interpret 1201(a)(2) [and other sections] if applied to 1201(f)(2) in the same way in the same scope would have led to a different conclusion.

[greglas]

8.

Right -- I'm certainly with you on the curious way that the exemptions seem to mean less than the literal text reads, while the liability provisions are read more expansively than the literal text.

On 1201(f), the Davidson case (where Blizzard was a plaintiff in another case involving interoperation with an ongoing service) is the hurdle I was thinking of. If MDY decided not to go after 1201(f), that was probably the case they were looking at. The court there interpreted the exemption very narrowly, and that reading, if applied to Glider, would have ruled out 1201(f). But you're right that Lexmark's language is broader.

It's hard to second guess strategy, but it does seem like it would have been worth giving 1201(f) a go here. The court's approval of the Corley approach makes me dubious the argument would have persuaded it, though. I still haven't looked back at the briefs to see if it was ever raised by MDY, but a quick text search of my library suggests it wasn't.

[robusticus]

9.

Well strategerizing is much easier with a ruling in hand, I imagine.

Is BNetD the case you are thinking of? Both BNetD and Corley involved technologies that created the possibility (likely, one could argue) of infringement where no possibility existed prior. With Glider infringement is simply impossible. So even the narrowest interpretation of 1201(f), even a more narrow one than Lexmark - which involved copying software in its entirety - should include Glider. I can't imagine any case that would qualify under 1201(f) if not something like this.

Is the Supreme Court next? Or is there an interim step yet? Sorry for the basic questions, Holiday Inn Express lawyering has its limits.

[robusticus]

10.

Well strategerizing is much easier with a ruling in hand, I imagine.

Is BNetD the case you are thinking of? Both BNetD and Corley involved technologies that created the possibility (likely, one could argue) of infringement where no possibility existed prior. With Glider infringement is simply impossible. So even the narrowest interpretation of 1201(f), even a more narrow one than Lexmark - which involved copying software in its entirety - should include Glider. I can't imagine any case that would qualify under 1201(f) if not something like this.

Is the Supreme Court next? Or is there an interim step yet? Sorry for the basic questions, Holiday Inn Express lawyering has its limits.

[greglas]

11.

Yes, it's the BNetD case I'm thinking of, and you're right that Blizzard argued in that case that the avoidance of the access keys in the rogue Battle server meant that people could play with pirated copies of the game. But there's some interesting language in that decision about the way the TOS relates to 1201(f) -- I could try to parse it, but in short, I don't think that court would rule different w/r/t Glider.

In Corely as well, the defendants made the claim that DeCSS could be used to obtain access for non-infringing fair uses, but the court was not buying that, essentially stating, like the MDY court, that 1201(a) was paracopyright law, and it didn't matter what happened after the tool was used to circumvent.

To do this right, I'd down and re-read these cases interpreting 1201(f) closely before replying. Maybe I'd look at the secondary literature too. So I'm doing this wrong.

The interim step would be en banc review by the Ninth Circuit -- I think this would be a great case for that.

In theory, it could go to the Supreme Court (wouldn't that be neat!) but honestly, I'd much rather a more sympathetic DMCA 1201 case went up first instead. Like I said in the OP, I think this case if about gold farming and -- writ broadly -- virtual world governance. But the record doesn't lay that material out.

Also, there are plenty of other circuit splits out there and the Supreme Court only takes a handful of cases each year. So if MDY wants to petition for cert, it gets a lottery ticket. A valid lottery ticket, but still a lottery ticket.

[robusticus]

12.

Even if the court disagrees it would be nice to at least have it addressed so we could know exactly what is the issue. I'd like to think it's a far cry from dreaming up some non-infringing use when you get sued to absolutely no infringement being possible, no infringement having been committed. But I'd also like to think it should be illegal for a system to kill an application for arbitrary reasons.

As for sympathetic cases, well, the underlying concept here is third party applications driving innovation. Virtual world automation is a huge driver of innovation, way more so than most types of third party applications. Glider losing causes the pace of innovation to slow. Ironically that will delay Congress addressing virtual worlds in the way they've addressed other media.

So I would be more sympathetic to 1201(a) if the third parties were allowed in some context but not others. Like for example if you could pay more for an iPhone that runs Flash but is not supported, or pay less for the "pure" version. In that case I would think it would be ok to kill and ban in the "pure" version.

[Houston Simmons]

13.

I agree with Andy, in that, blizzard is the best!

It is so interesting to observe the painful developments of our growing virtual world on the real world; even possibly the Supreme Court.