A Good Heist Requires a Good Plan

December 2, 2016

Earlier this year I was working on the prototype for my currently in development game Robyn HUD. At the time, I had this notion of creating a “sort of hacker, stealth, heist game with a strong story and character component.” Doesn’t exactly roll of the tongue does it? I didn’t know what else to call it at the time as I was envisioning a combination of gameplay that I’d never encountered in a game before. Developing the prototype helped me figure out what the game really is (or will be).

In Robyn HUD you take on the role of HUD, a hacker working in a safe house to help guide the cat burglar Robyn on the various heists the two partake in. As HUD, you gain access to the security cameras at your target and view Robyn’s progress through those cameras. Using your hacking abilities you help to disable alarms, redirect guard patrols, open security doors, and so on.

The most concerning issue that arose when I had a few people try out the prototype was that they were generally lost. In the early prototype you got full access to all of the security cameras in the test level. The goal of the test level was to sneak into a penthouse apartment, steal as much loot as possible, and sneak back out. Between having a view of the entire level all at once and having a very general goal of “get loot” there wasn’t a lot direction for players to help them figure out what to do or where to direct Robyn to.

Taking a break from Robyn HUD for a while, I took the opportunity to rewatch one of my favorite heist movies: Ocean’s Eleven with George Clooney and Brad Pitt. When the movie got to the briefing/planning scene something clicked inside my head. A large part of the appeal of a heist story is seeing the thieves meticulously planning for the heist. It donned on me that if I added some sort of pre-level planning system into the game it would directly address the issue of providing direction to players as well as incorporating a key aspect of more traditional heist narratives.

Way, way back in 1998 a new type of first person shooter game emerged in the form of Tom Clancy’s Rainbow Six. Unlike other FPS games of the time where a lone wolf soldier could storm and defeat an entire army of baddies with nary a scratch, Rainbow Six was a much more realistic tactical experience. One or two shots from an enemy could put the player down. The player had to work as part of a team and carefully plan their moves to achieve their objectives and get through each mission alive.

To organize each mission, Rainbow Six had a very complicated planning system in place. A very, very complicated planning system. Once you got used to how the system worked, you could formulate quite detailed plans, coordinating the moves and objectives of multiple squads of AI teammates working in a single mission. At the time, it was quite revolutionary. That didn’t stop it from having a very high learning curve though.

Fast forward nearly twenty years later. Most players don’t have the patience for complicated interfaces anymore. They want to get into a game and start blasting things right away. I knew that if I were to incorporate a planning system then it would need to be more streamlined than that of Rainbow Six. Even so, as soon as I hit on the idea of incorporating planning into my game I realized what my game really was: a strategic stealth game.

I spent a good day playing around with the Rainbow Six planning system figuring out what made it tick. I wasn’t looking so much to copy that system as I was to find inspiration that would lead to a more streamlined interface for my own game. One thing I realized early on was that a lot of the complexity in the Rainbow Six system was due to the player being able to plan the movements of multiple squads for a single mission. This requires laying out multiple paths through the mission, one for each team, and just being able to switch between teams. It also gave rise to very complicated control systems where a certain “go” command issued over the radio would send one or multiple squads into action.

With Robyn HUD, I don’t have multiple squads to coordinate. I just have Robyn herself to plan for. That alone strips out much of the complexity of the Rainbow Six system. My analysis of Rainbow Six also showed that a plan, when you distill it right down to its basic essence, is nothing more than a list of commands of the form “walk to a specific point and take some action.”

[The prototype blueprint didn't call attention to anything in particular.]
The prototype blueprint didn't call attention to anything in particular.

From there I started playing around with the Robyn HUD prototype. The prototype already incorporated a “blueprint” view, which showed the layout of a level. In the prototype, you initially used this view to choose an entry point into the level, but that was the extent of pre-planning that was done before starting the level itself.

Using the blueprint view, I played around with the idea of incorporating various “points of interest” including entrances into and exits out of the level. These points of interest are places in the map that Robyn can be directed to in order to perform some action, like steal an item at that location or connect a USB device to a computer that allows HUD (the player) to hack it. I quickly developed a context menu system for these points. With it, you click on a given point and can then choose from several actions that apply to that point. For example, if you click on a piece of loot you can issue instructions to steal that loot or to simply go to where it is.

[Creating a plan with points of interest.]
Creating a plan with points of interest.
Click to view larger.

From all of this emerged a system where you constructed a plan by choosing an entrance into the level, an exit from the level, and then any number of point of interest instructions in between. A basic plan is complete when such a path has been created that includes all the objectives of the level.

For example, if the objective is to steal a valuable statue a basic plan can be formed by selecting an entrance, selecting the “Steal” instruction on the statue’s point of interest on the blueprint, and then selecting an exit. Of course, making a simple run into the level straight for the loot without accounting for guard patrols or locked doors is likely to get Robyn caught. However, I wanted to allow players who like to jump in and “wing it” to do so as well as for players who want to take a more considered approach to do that too.

I still have more tweaks to do to the system and I want to get some playtest feedback on it in the not too distant future. If the system is too complicated then players will get annoyed by it and skip over it. At the same time, if the system doesn’t provide enough options for customizing the approach to a level then it will ultimately end up being useless. Given that it was the realization I needed a planning system that told me what my game was, I really want to make sure I get this right. Watch for more updates on the planning system in the future.