MMOG Developers Still Don't Get It

I was treated this morning to an interesting MMOG post-mortem courtesy of The subject of the post-mortem was the failed NCSoft/NetDevil car warsy game Auto Assualt, a game I wouldn't play for free, much less pay for because it was truly a stinker. The concept of the game was brilliant, a game of dueling, Road Warrior-esque vehicles in instanced arenas with destructible environments. How could anyone screw up a concept with such promise? NetDevil, makers of the failure Jumpgate proved they could screw up a wet dream. And the lessons they impart in their post-mortem shows a startling level of incompetence that properly explains why both of their games are so terrible.

Scott Brown, the game's producer, starts off the idiot's parade with this stunner:

"Want to make a great game?" asked Scott Brown, "Don't make a deal with
milestones. There are better ways to do a game deal."

There it is, right there. The incompetent's first best excuse, it wasn't our fault, the publisher just wanted milestones. A milestone, for those unfamiliar with the term, is merely a deadline, a date by which some part of the development process must be completed. In terms of the developer, it means that if the milestone is completed on time, the publisher gives the developer a nice fat check with which to continue work on the game. Developers who are not massive successes like Blizzard generally live off those milestone checks like they were food stamps. I'd wager that 80% of the game development houses out there are one missed milestone check away from closing their doors for good.

But what Mr. Brown seems to be advocating here is similar to the douchebaggery exhibited by Brad McQuaid in his recent interview about the failures of Vanguard. Both seem to be saying that publishers should just give them one gigantic barrel of money to start with and then leave them alone until the game is done. Milestones, benchmarks, deadlines are BAD for MMOG development because they unnecessarily hamper the development process. This is horseshit, of course. All software development needs milestones. It needs clearly quantifiable processes with concrete deliverables on a schedule, with plenty of time built into the schedule for testing for bugs AND for usability/playability. McQuaid decided one QA person was enough for Vanguard, and Scott Brown seems to think milestone schedules are anathema to MMOG development because no one knows what they'll be working on a year from now. That's what project management is about, you ninny, figuring out who should be working on what and when they should be working on it. If you cannot do that, you are in the wrong business. There is nothing special about MMOG development that makes milestones useless, and the more you believe milestones are useless, the more you need them.

They did get a few things right.

They admitted that Auto Assault went into beta too early. Beta too early = bad press and players trashing your game. "Getting everything crappy and calling it alpha, a little bit better and calling it beta is just plain wrong," Scott said. "It's great for stress testing, but it's not for testing functionality. Like it or not, beta is marketing. It's when the public is playing your game and you want to put your best foot forward."

Just calling something a beta does not a beta make. Beta is the first real test of the hype machine, not of the game engine. There should be no real testing for the players in beta. That's when you are stress testing the server, the client and everything else. Any real changes to the game's underlying mechanics in beta is sure to result in crap. If it wasn't fun in alpha, nothing you do in beta is going to add the fun. And as he says in the post-mortem, if the game's developers aren't playing it in alpha and beta, your game WILL suck.

But the most irritating thing about this post-mortem, and the ass-covering simpering done by McQuaid since Vanguard's release is that both of these teams should have known better. It isn't like these two were fresh out of college with no game experience beyond developing a level for Counter-Strike. These two teams had both shipped working MMOG projects, with varying degrees of success. These lessons should have been learned long ago. NetDevil has even less reason to make these mistakes. McQuaid was insulated from his own failings by the success of Everquest, a success his ego wrongly attributed to his creative vision. NetDevil's first project was nothing like a success, and fell into the same traps. By most accounts, it just wasn't fun to play. This team should have learned from their first failure, and shouldn't have made some of the same mistakes. Being able to talk about them in a post-mortem does not mean you've actually learned anything unless you can take those lessons and not make them the second time around. And now NCSoft has given the boys more money to produce another game. Why? What have they done to prove they should be given anything more than skeleton crew duty guarding a brick shithouse?

MMOG publishers and investors keep throwing good money over bad to the same sets of assholes that screwed the pooch their first time at the plate. As a result, the only innovative products are unmitigated trainwrecks and the only decent products are derivative yet polished clones like World of Warcraft. Auto Assault should have been fantastic, or at least playable. The fact that it was an atrocious mess means the lessons of failure weren't learned.

Labels: ,


At 12:01 PM , Blogger Baldrake said...

I enjoy dropping by your blog from time to time, Haemish!

You missed the point re the milestones. They aren't arguing there should be no reporting requirements. They're arguing that the milestones should be flexible. As in, developer and publisher should be able to get together periodically and work out a revised schedule as things progress. For example, if scalability becomes a big issue, let the developers work on that rather than mindlessly plugging on user interface because that's what the schedule says.

Any sane development contract allows ongoing review and revision, because you can't possibly plan how things are going to go over a three year development period.


Post a Comment

Subscribe to Post Comments [Atom]

<< Home