What We Learned—Magic Offline
Magic Online came to a screeching halt this week with the indefinite suspension of all Daily, Premier, MOCS, and PTQ events. The suspensions came after both the MOCS scheduled for Saturday the 9th and the PTQ scheduled for Sunday the 10th crashed. What seems to have fueled the drastic response by Wizards of the Coast was the experience of Pro Tour Hall of Fame member Brian Kibler, who was 7-0 in the MOCS when it crashed. Kibler took to his blog where he expressed his opinion that if Wizards could not deliver on tournaments with non-digital prize support, they had no business running those events to begin with. Today, we take a look at Magic Online’s bumpy past, dismal present, and hopeful future.
Disclosure: When I’m not writing for Hipsters of the Coast I am a full-time software developer for an international corporation that you’ve likely heard of. I have been writing server-side computer systems professionally for over 7 years. I have a degree in Computer Engineering from Lehigh University. These are my qualifications for analyzing the technical problems in Magic Online.
Magic Offline – A Cautionary Tale
2002-2003: Rough Beginnings – the Birth of MTGO
MTG Online began as Magic Online with Digital Objects (MODO) when a small software company, Leaping Lizard, approached Wizards about creating an online version of their successful CCG. The game went live to the public on June 24th, 2002. I had just turned 19 and was working at a summer camp that year, so I wouldn’t start using MTGO until September when I returned to university. This original version was very short-lived. After less than a year, Wizards decided to no longer use Leaping Lizard to manage the game, and brought the development of Magic Online in-house.
This presents the first technical challenge of MTGO that perhaps still echoes today. The original platform for the game was written by one group of developers, and then maintained by a separate group of developers, who also had the responsibility of continuing to update it. The problems that can arise in this situation are manifold. Wizards was enamored with the initial success of the product. They wanted to build off of it and announced a fancy new Version 2.0 of the software which would add enhancements, including the 8th Edition rules change, to coincide with the release of 8th Edition on paper.
Early on in the history of MTGO we are seeing a precedent set for two things that will come back to haunt us. First, we are building a platform in-house instead of through a professional software developer. Second, we are trying to push out software changes in-line with the release of paper-magic changes. The result at the time was predictable. The V2 release was disastrous. Servers constantly crashed, the rules engine didn’t function properly, and numerous bugs made the playing experience miserable. Wizards decided to turn off the servers, putting them into no-pay beta mode, while they worked out the kinks.
I’m sure this is starting to sound very familiar. Ten years later, not much has changed about Magic Online. The servers still have stability issues, the rules don’t always work properly, and numerous bugs still prevent the experience from being an enjoyable one. And, as the cherry on top, many features have now been disabled to allow Wizards to address these issues. We’ll take a look at the next ten years of MTGO’s life after V2, leading up to the Kibler incident this past week, to see how the game progressed. Most importantly, we are going to see what went right, what went wrong, and what should have been handled better. In the end of our journey down memory lane, I’ll tell you what I believe the future has in store for Magic Online.
2003: Crashing MTGO, Part I – Chuck’s Virtual Party
Eventually, normal service resumed and things in MTGO were mostly back to normal. In October of 2003, as a way to thank the players for putting up with the shortcomings and difficulties the past year had given them, Wizards announced that they would have a weekend of free sealed deck tournaments. This weekend was known as Chuck’s Virtual Party, named after then-CEO Chuck Huebner. Needless to say the event was a wreck. I won’t go into too many details, but suffice to say Wizards had not solved all of the server instability problems. Thanks to the free events, more people than ever had logged into MTGO, and the new V2 simply couldn’t handle the load.
What came next was the first of many apologies from Wizards about the state of Magic Online. Randy Buehler, who was the lead developer of Magic at the time (currently Aaron Forsythe’s position) wrote a State of the Game column addressing the problems with MTGO. Please note that a lot of the things Randy Buehler said on October 29th, 2003, echo the words of Worth Wollpert’s message from last week. Here are some of the highlights from that address which are pertinent to what is happening with MTGO today:
Magic Online is seen as a big success at Wizards of the Coast and Wizards of the Coast is going to do whatever it takes to make sure that Magic Online remains healthy and vibrant.
Right off the bat we’re talking about how Wizards is dedicated to Magic. The cynic in all of us has no choice but to ask, if Wizards was so committed to Magic Online in 2003, why are we still facing the exact same issues in 2013? Really there are two simple reasons this could be the case. Perhaps Wizards was never really committed to Magic Online. This seems unlikely. The other option is that Wizards didn’t get the right people to handle Magic Online. This is the option I want to explore further.
You might think that we could add more servers to deal with this problem, but that’s just not how Magic Online works. We can add more game servers to handle as many games as people want to play, but there is only one master server that handles everything else that goes on (chat, trading, ratings, etc.). Every time any user does anything outside of a “duel,” Magic Online has to spend some time thinking about that user. As we add more cool new features to the game, the amount of memory that needs to be allocated to each user keeps going up. At some point, when enough users are logged in doing enough things, the whole master server comes crashing down.
This is the first time we ever hear about how MTGO actually works. Buehler is describing the V2 system, and we’ll see later whether or not this still applies to the V3 system we all know today. Simply put, Magic Online is designed with a two-tiered server architecture. The top layer does everything that let’s users interact with the game. The lower layer handles actual games of Magic the Gathering. This is a design decision that, I believe, still plagues Magic Online today.
In order to have a permanent solution to this problem, we are currently putting together specifications for an overhaul of the system that will allow us to handle more users simply by adding more master servers. This is a major project that will take a lot of time and effort, but our number one priority is to make sure that Magic Online is fully scalable so that it can handle all the users and all the new features that we could ever hope to add to it.
That clearly didn’t work out so well, did it? This is another design decision that will come back to haunt all of us. In essence, instead of creating a scalable system that could actually handle “all the users and all the new features” that Wizards would ever want to add to Magic Online, they created a system that would require them to add more servers to support more users. This is actually the definition of an unscalable solution because every server you add complicates an already intricate system. What this says to me is that while Wizards was committed to making Magic Online great, the right people weren’t making the decisions to get them there. The last line of Buehler’s article is damning:
In a weird way, this is actually a good problem to have. We are truly sorry that we put the Magic Online community through this, but we are optimistic that ten years from now we’ll look back on this as the Magic Online-equivalent of the Homelands set: Mistakes were made and we shouldn’t have done it that way, but we learned from it and the game survived just fine.
Ten years later, if V2 looks like Homelands, then V3 may look like Fallen Empires ten years from now. While we have plenty more features than we had in 2003, the core problems remain. The online version of Magic does not scale with the success of the paper version of Magic. The decisions made at the time of Chuck’s Virtual Party would not provide the long-term solutions Wizards wanted. But that’s looking a bit too far ahead. In all fairness to Wizards of the Coast, the majority of the next ten years of MTGO would be a rousing success.
2004-2007: Adolescence – MTGO Version 2.0
For most of MTGO V2’s life, it would be defined by the impending V3. First though, Wizards needed to clean house after the debacle of Chuck’s Virtual Party. In February of 2004, Dan Myers was made the communications manager for MTGO. This was an internal move, since Myers had spent a lot of time already working for Magic Online. This was around the time that the MTG website was being completely overhauled, and improving the communications of MTGO was a part of that initiative. A week later, Myers made the official announcement that V3 was coming.
There’s a lot of very important information to glean here. The first is that Wizards fully expects it to take at least 18 months to write V3 if it is a complete rebuild. That’s a bit confusing. At this point, in February of 2004, Wizards has not yet decided if they want to completely rebuild the system, but they know it would take at least 18 months to do so. Another important piece of information Myers provides is that there is a small team maintaining V2, and another team working on the potential V3.
But the big problem at the time was the absence of Premier Events, which had been offline for a while. Extra programmers were brought in to clean up the Premier Events, which means working on the Master Server code. Let’s recap real quick then shall we? In February of 2004 there are three teams working on MTGO’s code. Team A is maintaining the day-to-day operations, adding new sets, and fixing bugs. Team B is working on the Master Server in an attempt to cleanup the stability issues that prevent Premier Events from running. Finally, Team C is hard at work planning V3, which may or may not be a complete rebuild, but will take at least 18 months.
Does this sound like a logistics nightmare to anyone else? Perhaps one that could have long-lasting ramifications that may suddenly and unexpectedly cause the entire system to collapse in no more than ten years? I thought it might.
Another big hire came later that year with Justin Ziran becoming the MTGO Brand Manager, taking over from Linda Cox. Ziran was also an internal move, coming across from having worked on D&D, Chainmail, Pokemon, and Magic. In April he announces the return of Premier Events which Wizards delivers on, as they return in June.
However, he also announces that V3 is still 12-18 months away. We now know that it is a complete rewrite of the MTGO system. It will have two phases, the first replicating everything they have in V2 but with increased stability. The latter phase would be for “nice to haves” which in this line of work translates to “things that are never actually getting done but developers put them on a list called nice to haves which is a nice way of saying it’s never gonna happen.” League play apparently ended up in Phase 2.
Ziran closes his opening announcement by repeating the company line that we previously heard from Dan Meyers, Linda Cox, and Randy Buehler: Wizards is committed to supporting MTGO. This is starting to become a bit of a joke. Clearly Wizards is committed to making it look like they’re committed to MTGO, but what have they really done to show it? They’ve kept the same people internally working on the same project and they keep having the same problems. I believe they think they’re doing the right thing, but they clearly are not.
2007-2013: Masters and Champions – MTGO Version 3.0
If you’ve been playing Magic Online for a while then you should have scoffed a bit when you read the 18 month estimate that Meyers and Ziran gave for V3. We now know that it actually took the better part of four years to develop MTGO V3. As the release of V3 kept being pushed further back, it’s important to recall that MTGO development had been split into three branches. For several years now, not several months, one team was updating rules, another the Master Server, and a third team was rewriting everything that the other two teams had done. Not the greatest business model but the game kept chugging along.
The transition would really begin in 2007. Randy Buehler was promoted to VP of Digital Gaming in January, and then Worth Wollpert replaced Justin Ziran as Brand Manager in April. V3’s release date keeps getting pushed back, but it finally goes into public Beta in March. It’s a big year for MTGO. The game continues to grow with the popularity of paper Magic. Future Sight, Masters Edition, and Lorwyn highlight a big year. 2008 is even bigger with the actual release of V3 in April.
Things are finally looking up. V3 has its problems, as expected, but after the initial surface shock of the changes to the interface, it’s a much better system than V2. This is a bit of a golden age for MTGO. Over the next few years, the system powers through the bugs and stability issues. They introduce the Magic Online Championship Series in 2009. This is a massive circuit and culminates in cash prizes and an invitation to the Magic Online Championships. In 2010, Pro Tour Qualifiers are introduced to Magic Online and are a rousing success.
More people are playing MTGO than ever before. More games. More tournaments. More trading. More transactions. The limits of the V3 system are being tested and eventually, as they did once before, they would fail.
2013: Crashing MTGO, Part II – Kibler’s Virtual Party
All of the problems that MTGO V3 experienced over the past few years can be summed up in Brian Kibler’s blog post about how the MOCS should not exist. Kibler’s main gripe is valid. The event compensation was an invitation to the Magic Online Championship. The consolation rewarded to players who were still in the event when it crashed was a number of digital packs, promo cards, and an opportunity to play in a make-up event. To Kibler this is unacceptable for a tournament that was meant to pay out a non-digital reward.
MTGO had been experiencing these problems for a while leading up to the final weekend. However, the opinion of the widely beloved Brian Kibler forced them to take action. On the Wednesday following Kibler’s weekend rant, Worth Wollpert announced that Wizards would be taking down daily, premier, MOCS, and PTQ events. He reassured us that the team is committed to bringing us the best experience possible. He also did an interview with Evan Erwin for The Magic Show.
But we should not be reassured. In 2003 Randy Buehler told us that Wizards was committed to MTGO after Chuck’s Virtual Party. In 2004 Justin Ziran told us that Wizards was committed to MTGO when V3 was announced. In 2008 Worth Wollpert told us that Wizards was committed to MTGO when V3 was released full of bugs and missing features. Now, in 2013 Worth Wollpert tells us that Wizards is committed to MTGO during the recent stability issues that have brought it to its knees.
Insanity: Doing the same thing over and over again and expecting different results.
The Future of Magic Online
Ten years ago, Randy Buehler sincerely hoped that we would look back on V1 of MTGO as the Homelands of Magic Online. While he’s right that V1 is essentially the worst piece of digital trash that Wizards ever produced, MTGO unfortunately hasn’t made the same progress as its paper component. V2 may best be compared to Mirage, the first block ever produced, not well-designed but with hints of a greater thing on the horizon. V3 could aptly be compared to Urza’s Block, which had some amazing moments, provided for a great experience sometimes, but was also a huge mess and caused more grief than it should have.
What will V4 be? Ten years from now, will it be remembered as the Ravnica block of MTGO? Or will it fall short and be remembered more like Kamigawa. We’ve now discovered that Wizards has very likely mismanaged MTGO from the day they assumed control from Leaping Lizard. Between the heaps of code being built upon, the multiple teams developing the same project, and the continued use of an in-house solution instead of outside consulting, the mistakes have been numerous and repetitive.
No one has come out with an estimate for how soon MTGO will return to its full glory. If the past is bound to repeat itself, then expect them to announce V4 shortly, as a complete overhaul of V3, with some ridiculously short estimate like 12-18 months. The reality is that it will take much longer for them to, once again, redesign the system. We could be looking at a V2-to-V3 transition all over again.
But, perhaps there is hope that Wizards will finally, truly commit to making MTGO great. They will treat it as a software problem, not a gaming problem, and bring in a software company or consultants to work on the program. They will actually address the stability and scalability issues that have plagued it for a decade. They will provide the MTGO we all deserve at this point.
It will be at least a few months before we know the answer to this question, but if it runs the same course as last time, the next step will be cleaning house and bringing in new people to run MTGO. Worth Wollpert may want to start dusting off his LinkedIn account.
Pro Tour Update
Grand Prix Washington, D.C.
After a dozen Grand Prix Top 8s and having formerly been Player of the Year, Owen Turtenwald finally won a Grand Prix championship. He won past the nearly 1,700 participants on day one, and then the 235 of them who made day two, and finally battled through Andrew Cuneo, (10) Sam Black, and Jared Boettcher to take home the trophy (shown above).
Overall, eleven ranked pros showed up this weekend to try their hand at the Legacy format. An impressive three of them, Turtenwald, Black, and (20) Craig Wescoe finished in the Top 8. (25) Christian Calcano and (4) Shahar Shenhar also finished in the top 64 to pick up some Pro Points. Turtenwald piloted Delver, which was a popular choice on the day, but the format remains incredibly diverse with eight different archetypes represented in the Top 8.
Top 25 Update
I expect GP DC to have a big impact on the Top 25, which has been relatively stagnant lately. Perhaps we’re discovering a flaw in the system that Wizards uses, or there just aren’t enough events with high-level players in them. Last week saw Raph Levy move up past Paul Rietzl. Levy finished in 9th place at Grand Prix Valencia, finishing in 9th for the fourth time in his career. It’s a tough break, but it got him pushed back up to 17th in the world.
The Quick Hits
- LSV announces a CFB collaboration with Cascade Games for the 2014 season [Channel Fireball]
- Riki Hayashi expertly breaks down what unsportsmanlike conduct means [StarCity Games]
- Return to Ravnica battles Innistrad this week on Battle of the Blocks [StarCity Games]
- MJ Scott presents the first runner-up in her Commander fiction contest [Gathering Magic]
- Darwin Kastle talks about the darker side of seemingly nice people cheating [Gathering Magic]
- Heather Lafferty interviews gamer and talented artist Dave Lee [Gamer Boy, Gamer Girl]
- Mark Nestico talks about a few things that upset him, like MTGO [Quick Hits]
- Jason Alt reacts to reactions to Brian Kibler’s reaction to MTGO crashing [Jason's Article]
- Josh Silvestri lists the things he’d like fixed about MTGO [Silvestri Says]
- Jim Davis chimes in on the MTGO conundrum [StarCity Games]
- Melissa DeTora reminisces about her experiences with MTGO for over a decade [TCGPlayer]
- Mike Linneman interviews MTG artist John Avon [Gathering Magic]
- MJ Scott presents some rejected Commander 2013 flavor text [Gathering Magic]
Wallpaper of the Week
We’re back to Theros wallpapers after a (very) brief foray into Commander 2013. This week brings us the massive 10/10 Colossus of Akros, indestructible defender extraordinaire. My only gripe is that he’s not facing the opposite direction, defending my application icons on the left-hand side of my desktop. Instead he stands watch over my taskbar, but it’s an impressive watch nonetheless.
The Week Ahead
This coming Saturday is the 50th Anniversary special for Doctor Who. I highly recommend it. If you’re not into Doctor Who, then there will be two Grand Prix tournaments you can follow instead! The first will come from Kyoto, Japan and present the first major Team Limited event for Theros block. Kyoto is hosting its first Grand Prix since March of 2007. This will be the 5th Grand Prix ever held in Kyoto. A Pro Tour was also held there in the spring of 2009. If you’re a fan of Standard you won’t want to miss out on Grand Prix Albuquerque, coming to you live from New Mexico by SunMesa Events. This is the first ever Grand Prix to be held in New Mexico. Pro Level events most recently visited the Southwestern United States at Las Vegas in June this year.
What We Learned is a weekly feature here at Hipsters of the Coast written by former amateur Magic Player Rich Stein, who came really close to making day two of a Grand Prix on several occasions. Each week we will take a look at the past seven days of major events, big news items, and community happenings so that you can keep up-to-date on all the latest and greatest Magic: the Gathering community news.