     |
 |
 |
 | | | | |
 |
|||
 |
||||||
          |
 |
 |
||||
|
Postmortem: What Went Wrong Play Testing. On May 1, 2004, a little over one month before being sent to manufacturing, The Political Machine sucked. It wasn't any fun. There was a cloud of despair hanging around the offices. We didn't know what we were going to do. So we started brainstorming, and lots of good ideas came out. What's particularly startling is the fact that these ideas I mention below all were devised, implemented and bug tested in 30 days:
So, I kid you not, all of these features were added in about 30 days. But why is this in the "what went wrong" category? Most of it nearly didn't get play tested. The guys got so enamored playing each other at multiplayer that no one had, for example, tested to see if you could actually win the campaign. Fortunately, in one of the last release candidates, I sat down and started to play the campaign for fun. And as the guy who coded the computer AI, if I can't win the campaign, nobody could. When I got up to George Washington to play, he won every state. I didn't even win a single state. He smoked me. So over the course of that weekend we began massive play testing to make sure the game played how we wanted. We then submitted to Ubisoft a new release candidate that Monday (someone from Ubisoft is probably reading this thinking: "Oh god, I didn't want to know that!") Insufficient compatibility testing. When we decided to use DirectX 9, we assumed that was fine. We had no idea that the game would have compatibility problems with some hardware configurations. Our view was that, as long as we followed the rules, we would be fine. We were wrong. Firstly, the main reason we used DirectX 9 was to get access to easy TrueType font support. But it turns out that the older integrated Intel video cards don't support that. So essentially, we have a game that doesn't work on those machines. There's no "workaround" for it, other than us scrapping TrueType fonts and going back to encoding fonts again, which we don't have the budget for (especially since this only affects one type of video card). But on a game that targets "casual gamers", you have a high number (in absolute terms) of people who have these low-end video cards, and who don't understand why a game that seems to have relatively mild requirements wouldn't work on their system. This results in very mad users. Unfortunately, mad users don't just return games, some of them go to review sites and take out their anger there. It only takes one angry guy giving a game a rating of 1 out of 10 on a game reviews website to really mess up the overall rating. Even though this is technically Intel's fault, the user doesn't care. If I had to do it over again, I would probably have required us to put in some sort of "fallback" mode with basic font support, so that the game would work on those Intel machines to a certain degree. Alternatively, we could have pushed Intel to fix their drivers before the game's release (the newer integrated Intel chipsets--or at least, those that are less than 2 years old - work fine).
Horrible Documentation. Our last game, Galactic Civilizations, had mediocre documentation. The Political Machine's documentation is even worse. First, it has gross errors in it, such as referring to local TV interviews and political debates. We took these features out of the game because they weren't fun at all and made things tedious. But we did that after the manual had been sent to the printers. Needless to say, these are big changes to be doing at the last second. But users probably could have forgiven that if the manual was better. I wrote the manual, and as the Gamasutra editors can tell you, my writing is pretty deplorable when there are no professionals such as themselves to convert my writing into something readable. The Political Machine had little editing, since there just wasn't enough time. Next time I'm going to find someone to work with me early, in order to create a good manual out of my terribly written design documents. Excessive Initiative. The game succeeded because we had good personal initiative by the development team. But sometimes, developers would sneak in a feature without passing it by the rest of us. A windowed mode was added in Release Candidate 2 without anyone knowing. 4X anti-aliasing was also added without telling anyone. These were all good things, but imagine if there'd been some terrible bug as a result? This also led to people having their own styles of coding. Stardock is growing quickly, and our programming guidelines aren't that specific. So we had internal battles on whether to use CString classes or just plain oh char szMyString[255];. We also had issues resolving just how functions should work. (Typically, it is against Stardock coding guidelines for functions to modify a parameter. They should always return a new value at which point a variable can be updated.) Multiplayer. The biggest regret I have, in hindsight, was the decision to have a multiplayer mode in The Political Machine. The game features a full-blown matchmaking service, in order to make it relatively easy to play multiplayer. I love playing games multiplayer, and I've played a lot of games online with people. That said, based on the sales statistics, and based on the server stats, less than 1% of players are playing the game multiplayer. As a gamer, I demand multiplayer in my games. It affects my purchasing decisions. But as a game developer, I recognize that people like me are an extreme minority. Outside of a handful of games, most games don't reach critical mass in online players to make a successful multiplayer community. We would have been much better off putting that time into enhancing the single player game. We had several other features in mind that we had to cut because of lack of time, time we would have had if we hadn't done multiplayer. For instance, we couldn't include multiple maps, and I'd love to have had the ability to play in Canada, or the UK, or the whole of Europe, or even a worldwide map. That's not even counting made-up maps we could have created. I think the inclusion of different maps to play on would have increased the replayability of the game a great deal. I would have also liked to have given different candidates "special" powers, to make each one more unique apart from statistical differences. But the time in development and testing on multiplayer eliminated those kinds of things. When gamers demand multiplayer, what they don't realize is that something has to be sacrificed for it. You have a finite budget and a finite time to use it, and so, to quote Spock: "The needs of the many outweigh the needs of the few." With that in mind, I wish we hadn't done multiplayer in The Political Machine. The Politics In a postmortem, the section dealing with what went right or wrong is usually the main part. But there's more in this particular game. In a political strategy game, especially in a hotly contested year such as this one with Bush vs. Kerry, we had to put a lot of effort into making sure the game was fair to both sides. People would be looking for bias in the game, and probing for any hidden agendas. We wanted the game to be accurate enough to the real world that political junkies wouldn't be turned off, but we wanted to also make sure it was a fun game. This is a strategy game, not a simulator. We think that the game is reasonably fair to both sides (or equally unfair, if you prefer). But to make it fair, we had to do a lot of research. We had to look at census data, polling data, exit polls, and watch a lot of cable TV news. I can tell you the pet topics of pretty much any cable TV host, even those not in the game. I can tell you that Fox News isn't particularly conservative, and I can also tell you that there isn't a partisan bias in any of the cable TV shows. They focus on the issues, and some issues favor Democrats and some issues favor Republicans. But the host's preference to discussing a given issue doesn't make them partisan on its own. Overall,
we're very happy with how the game turned out. But we'll never ever
do a game under these kinds of deadlines again. That was insane.
Now, if you'll excuse me, I have to get back to working on Santa
Saves Christmas which has to be done by… ah, never mind!
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Features
|
|
|