Sometimes, Failure Is an Option

12:45 am Game Development

“I have not failed. I’ve just found 10,000 ways that won’t work.”

“I am not discouraged, because every wrong attempt discarded is another step forward.”

– Thomas Edison

Edison would have been a great game designer. He understood the importance of iteration, trying new things, observing the results and adjusting his techniques. One behavior I often see in our industry is little to no acceptance of failure. I’m not talking about a game sucking and being considered a failure, but instead of internal failures on game features, tech, levels, designs, and so on.

Here’s a few examples of failure I’ve seen first hand or have heard about across the industry and how they should have been handled:

Failure: Banana Team is given a chance to implement Slippery Banana Peel Tech (SBPT). Banana Team implements said peels, but after one iteration the bananas aren’t slippery enough. SBPT is considered a failure, so Banana Team is forced to move onto the next feature.

What should have happened: The problem with this situation is that Banana team should have had more chances to iterate on SBPT. Eventually, if things aren’t going in the right direction the team should decide to cut SBPT. However, one, two, three, even a hundred iterations might be necessary to make a feature work properly. That’s really up for the team to decide, but one or two iterations on anything generally isn’t enough. Sometimes, people need to fail multiple times in order to see the err of their ways.

Failure: Potato Team has made multiple potato based simulators. Potato Chips, Potato Soup, Mashed Potatoes, Sweet Potato Pie, you name it, they’ve simulated it. However, consumers really like mashed potatoes and keep buying them. Potato Team is tasked with making mashed potatoes new and exciting. They’ve already made a Garlic Mashed sim, a Super Buttery Mashed sim, even Jalapeno mashed sim. They’re running out of ideas and the their newest mashed, Honey Maple Mashed sim isn’t sitting well with consumers. The company CEO (Mr. Potato) isn’t happy with Potato Team’s failure and decides that Potato Team is out of touch. Mr Potato only allows Potato Team to cook recipes Mrs. Potato comes up with.

What should have happened: The big mistake here is expecting a team of people to hit home runs every time they’re up to bat. The other big one is messing with the formula too much. The second mistake here is punishing the home-run hitting team for finally making a mistake, basically guaranteeing that they’ll never hit a homer again by disenfranchising them. Potato team should have have been given more chances to fail and learn from their mistakes, especially due to their history of pleasing customers with their simulated starchy goodness.

Failure: The newest member of Team Popcorn has never popped a virtual kernel in his life. When anyone asks him if he knows how to make virtual popcorn, he says “Sure!” When he inevitably fails, and when the Sr. Lead Popcorn Production Designer tells him exactly how to pop the corn in overly technical terms, then asks the Jr. Popcorn designer if he understands, Jr. Popcorn designer says “Sure!” and fails again. This cycle repeats until the Jr. Popcorn designer is shifted to Potato Team to code up the butter system.

What should have happened: This one should be obvious: the new guy should have asked for help, or said “I don’t understand what you just said.” The problem here is being afraid to fail, which is almost always much worse than any failure. I’ve worked with people who have had this problem- they see asking for help as a sign of weakness. Generally, managers like answering questions because it keeps them on their toes, and it let them know what people are working on, and that people are engaged in solving problems. Sure, having someone ask too many questions can be annoying (at a certain point, it becomes clear that people are disengaging from problems and seek help without trying to take a stab at solving it themselves) though, so there’s a fine balance here.

The other problem is that the more Senior guy gave way to technical advice which was over the new guy’s head. It’s tough, but people who have been working at a game studio for a while need to think back on what it was like trying to learn that proprietary scripting language, that internal set of API shader calls, or that in house level editing tool for the first time.

Failure: Cabbage Group loves cabbage. I mean, hey, who doesn’t? They focus test a grilled cabbage controller (GCC) on one guy late in the project because Cabbage group is positive that GCC will be a smashing success. This one focus tester simply loves it, which of course makes Cabbage Group feel vindicated. The press comes by and even gets to see GCC (but don’t touch it of course, this is our demo cabbage). Cabbage team loves GCC, at least one other person in the world loves it, so they mass produce GCC and market it. A strange thing happens though, it turns out that not many people actually like GCC. Cabbage group doesn’t get it: they even focused tested GCC and had the press say such wonderful things! Why aren’t people in love with GCC?

What should have happened: GCC should have been focused tested on multiple people and iterated on early on. I’ve heard of game levels tested by one person or by a group of people internally who like a particular facet of the game, the game gets made based on this one data point, and when gamers get play the level or use the feature, they by and large hate it. Focus testing is one of the most important processes of game making- it can provide solid quantitative (and qualitative) data which people on your team can use to make their game better, so long as their mind is open. I know of multiple studios (Bungie and Relic, to name a few) that use heavy focus testing, iteration, and adjustment and it shows. The game needs to go through focus testing and potentially fail multiple times before it’s ready for public consumption.

There are more cases I’m leaving out or forgetting. What am I getting at here? Let your team fail so your game won’t. Just don’t fail too much or you’ll probably really fail.

(I wrote this over a year ago and have been sitting on it for a while. Posting it now because, hey… why not?)

Leave a Comment

Your comment

Please note: Comment moderation is enabled and may delay your comment. There is no need to resubmit your comment.