Last week—May 17^th through 20^th, 2012, to be precise—Jonathan Blow, Daniel Benmergui, Marc ten Bosch, and I spent four days in a rented house in Stinson Beach, California, at an event we called the Depth Jam. During these four days, we talked about, playtested, designed, and programmed our video games in alternating timeslots, went to sleep, and then woke up and did it all over again the next day.

As the name suggests, our goals for the Depth Jam were to go deep into the design of a specific part of each of our games, answer a specific design question, or solve a specific design problem. It was fun, educational, inspirational, and exhausting.

We also ate too much catered food. Daniel said we should have called it the Breadth Jam.

In this article, I'm going to discuss our experiences at this first Depth Jam, what motivated us to do this, what we hoped to accomplish, and how we set it up—both logistically and conceptually—in the hopes that other game developers will be interested in running a Depth Jam for their games.

Why Depth Jam?

In a most excellent coincidence, 2012 is the 10^th anniversary of the Indie Game Jam, one of the first "get a bunch of people together in a room for a long weekend and make some games" events. Over the past decade, game jams have become more and more popular, and a wide variety of folks put on game jams in every setting, from schools to large corporations, basically anywhere game developers want to be inspired and work on something wacky and short-term.

As game jams grew in popularity, Jonathan and I started having some reservations about them and their influence on the industry, on game development culture, on games, and on the art form itself. These thoughts and discussions culminated in my 2010 Game Developers Conference rant, Please Finish Your Game.

In some important ways, the Depth Jam is a reaction to the idea and execution of game jams in the game industry today. Game jams are shallow by design. You come in, you make some games, you're done. Maybe you polish your jam game a bit afterwards and throw it up on the web, but it's enough of a challenge just to get code executing in four days, and trying to do real thoughtful exploration in that amount of time in a new codebase is almost impossible. Sometimes you luck out and get something amazing from a game jam, but in a lot of ways the structure is directly working against that.

Let me point out that a lot of people like to do game jams simply because they like to hang out with other people making games. That's great, and I think the more people making games, the better, but it's become clear to me game jams are not really pushing game design forward anymore.

So, we started talking about deeper alternatives during the summer of 2010, and a few things became clear:

• First, your focus has to be on a game you're already working on, where you know the code inside and out and can change it to test new ideas quickly, where you have solved the easy stuff already, and where the game is already interesting and deep, or has the clear potential to be deep with some good design decisions. Of course, the game shouldn't be so far along that you can't make potentially large changes to it.
• Second, it has to be kept to a small number of people. Game jams these days can get really huge, involving tens of people in the same area working...that is not a recipe for thoughtfulness.
• Finally, it has to be a retreat, away from employees, bills, family, and the other distractions of normal life.

After talking a bit more about it at IndieCade in October of 2010, Jonathan and I started to make a concrete plan. We decided the first Depth Jam would include SpyParty and The Witness obviously, and we'd invite Miegakure and Journey. At the time, all of these games were early enough in their development that change was still possible and welcome, and we knew all the designers and thought the mix would work well.

Everybody was interested and excited by the idea, so naturally we got busy and completely dropped it for a year.

When we finally picked the idea back up at the 2011 IndieCade, it was too late in its development for Journey to benefit from the Depth Jam, so we started thinking about an appropriate replacement game. We thought about just doing three games, but four games in four days seemed like the right number for the first jam. Around the same time, Jonathan and I were talking to Daniel Benmergui about his game Storyteller, and I invited him to come to Oakland from Buenos Aires for an extended work trip, where he'd hang out with us and develop his game. Once he decided to do this, and we scheduled it for the three months after GDC 2012, it was clear the Depth Jam should take place in the middle of his trip, and that Storyteller should be the 4^th game.

How to Depth Jam

Since this was the first Depth Jam, we basically had to make it all up ourselves, so there may be better ways to do it, and we'd be delighted to hear if you try a Depth Jam yourself and change some things for the better.

There were a bunch of different aspects of the Depth Jam to consider, and since the jam itself was about depth, we figured it was worth thinking somewhat deeply about these issues as well:

• The Location    We knew it had to be a place away-from-it-all, but there are about a billion of these places in California, so we needed to narrow it down. First, we decided to do it in a vacation rental house, as opposed to a hotel or a conference center. Everyone being together in the same space seemed important, so a hotel was out, and since we were only four people, any conference center or retreat destination would almost certainly have other groups staying at the same time as us. However, hotels and conference centers serve food as part of their packages, so now we needed to handle that. Cooking is fun, but it takes time and energy, so we decided to have professional caterers bring lunch and dinner for the entire stay, and just have some simple snacks and breakfast cereal around for the mornings. Finally, we had to decide where to look for the rental house, because there are lots of those too. We knew we wanted it somewhat isolated and beautiful, and differentiated from our daily routine. To narrow things down, we used "woods versus beach" as a razor, and decided beach was better because it allowed quicker escapes—a hike in the woods takes more time than a quick walk to the beach to think about a problem. We rented the house for five nights so we had four full days, meaning we came in on a Wednesday evening and left on a Monday morning.

• The Format    Four games, four days; simple, right? Well, not really. There are a lot of ways to split up a day, and after thinking about it for a bit, it seemed unwise to give one day to each game, since people get tired of thinking about the same thing, plus that would prevent people from making requested programming or tuning changes and then retesting their games. We eventually decided on doing all four games each day, with two hour slots, then free time after dinner to program or just stare at a blank wall. These two hour slots quickly hit the wall of reality on the first day, when the scheduled 15 minute breaks in the afternoon weren't long enough to recharge after an intense session, so we rejiggered it a bit more on site. We settled on a two hour slot each morning between breakfast and lunch and shorter 1.5 or 1.75 hour slots in the afternoon with longer breaks in between, then dinner. The morning slot tended to be the freshest since people were rested, and each timeslot had a slightly different feel to it, so we made sure the games got one of each. We also allowed the last slot to happen after dinner if a good discussion started in the afternoon, or people felt like they needed more of a break. However, the quality of the conversation definitely dipped as the day wore on, and if the sessions leaked into the after-dinner period then it meant less time to implement the day's ideas. We also found it important to take real breaks from the games, and not let the the discussions bleed into breaks...Daniel and I played some intense frisbee on the beach every afternoon, Jonathan did t'ai chi and kung fu on the patio, and Marc read.

• The Questions    The biggest thing to decide is what exactly each of us were going to think deeply about at the Depth Jam. We all started thinking about ideas a couple months before the jam, making sure we each played all the games during this period, and then finally decided on and introduced the ideas to the others at a pre-jam meeting in Oakland. We hammered on the ideas a bit during discussion, pointing out problems with the questions and alternatives, but in the end each person set his own agenda for his game. The questions we each posed were quite different, and we'll all write in more detail about questions on our individual game blogs at a later date. Here are short descriptions of our questions as concrete examples to give an idea of the kinds of topics we investigated:
  • SpyParty    My question was a very specific one about advanced Sniper play and deductivity with the highlight/lowlight mechanic. I currently have two levels of highlight and two levels of lowlight, but in the beta I've observed some very good Snipers using the levels to count different actions, to the point where instead of using the highlight for suspicion and lowlight for lack of suspicion, they were using all the levels as a 5-counter for different activities in the world, like number of times a character would go to the statues. This was an interesting discovery, but felt like it was getting too deductive for the gameplay aesthetic I wanted, so I was considering nerfing the mechanic down to just one level each of highlight and lowlight. There was some concern during the pre-jam meeting about this question being too specific, or the Depth Jam attendees not having a high enough skill level at SpyParty to help answer the question, but I was convinced I'd get some useful data from the jam. On-site, we tested it pretty thoroughly, including going the complete other direction and having the game automatically put counters for activities over every character's head, so we could see what the other end of the deductivity spectrum felt like. In the end, the nerf feels right, so I'm going to put it in the next beta build. As a bonus, we redesigned the Inspect Statues mission and hopefully fixed some of its flaws. I'll post more detail on these questions and decisions later on the SpyParty blog.
  • Storyteller    Daniel has been making huge changes to the internals of his game since winning the Nuovo Award at the IGF. He was faced with a hard decision about which direction to take the game, whether towards a more freeform interpretive plot based game, or a more tightly constrained character puzzle game, and so he decided figuring out which way to go would be his question. He implemented multiple puzzles and new mechanics for us to test on the run-up to the jam, and then modified the game at the jam in response to feedback. He'll elucidate his chosen direction on his Storyteller blog.
  • Miegakure    Marc led discussions about whether leaving the outside of the levels up to the imagination of the player was in fact the best way to visually represent the compact levels the gameplay requires. We also playtested and discussed two series of levels involving specific late-game mechanics. Watch the Miegakure blog for more information.
  • The Witness    (undisclosed)

Again, your decisions regarding some of these issues might be different, but they all matter in more and less subtle ways, so be thoughtful about them as you're setting up your Depth Jam.

Outcome and Thoughts

This article has gone from a high concept about deeply exploring game designs to advice on how to pick a caterer, which is kind of like the Depth Jam itself. At the end of the day, it was about making concrete forward progress on our games, using the deep conceptual thinking and discussions in the process, but always ending up in low level code and level design. That's as it should be.

We haven't done the official Depth Jam post mortem meeting yet, because we wanted time for it all to sink in, but I think the vibe afterwards was generally positive, but not just a bit exhausted. I would say I had very high expectations and it met them or slightly exceeded them, but it did not completely blow my mind, possibly because my expectations were already so high, or maybe because the takeaway from the Depth Jam is a complex and subtle addition or modification to a large body of thoughts I already had about my game, rather than something completely new and different. I would definitely do it again, and it was valuable both for SpyParty, and for me as a game designer. The caliber of conversation was at a level that's hard to get on a regular basis, so having four straight days of that was awesome.

The big problems with the Depth Jam, as I see them, are:

• Cost - Because of the various decisions we made regarding location and food, the Depth Jam was expensive: it cost about $5000 USD, around $1500 for the catering and $3000 for the house rental. You could obviously save money in a lot of ways, like cooking yourself or only catering dinner and making PB&J for lunch, renting or borrowing a cheaper place, etc., but we wanted to make the probability of success as high as possible this first time, so we tried to make choices that eliminated distractions and energy sinks. So, you could definitely do it cheaper than we did, but I would be wary of skimping on things that end up causing more non-design cognitive load.
• Scalability - It's a shame, but there's no way the Depth Jam idea is scalable in the way a game jam is. I think you could do a three or a four person Depth Jam, but that's about the range. My hunch is having other designers there who aren't asking a question about their own games but who are just engaged in the discussions and playtests with the three or four chosen games wouldn't work very well, but we didn't try this so I can't say for sure. The Depth Jam also obviously works best when the attendees can make any required changes to their game, which means they have to be programmers. I changed the network protocol for SpyParty to test a theory during jam...that is not something you can easily put in a tuning file and make accessible to non-programmers.
• Endurance - It may sound awesome to hang out with three of your friends on a beach talking about video games for four days, but you are really working hard and smart the whole time, and the schedule becomes oppressive. Even the food was oppressive, since it came at a specific time and we had to eat it even if we weren't hungry yet. It's much harder to Depth Jam than it is to game jam, or work four regular work days. Also, there is much more preparation before the jam, since everyone needs to play all the games, and think deeply about all the questions to get the most out of it.

So, there you have it, the report so far. Will Depth Jams catch on like game jams did? I hope so, but I don't know, and that's largely up to you. We'll almost certainly do another one at some point in the future, but it seems Depth Jams are most valuable at specific times in a game's development cycle, so it doesn't feel like an every-year type thing for us. Time will tell, and keep me posted on your Depth Jams if you do them. Also, feel free to ask questions and I'll try to update this article with answers.

Daniel wrote about the jam here.

---

Photo Gallery

• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
•