Solving interdimensional crimes

Designing the investigation system for Double Cross

One of my favourite moments in Double Cross’s pre-production was when we agreed to try the interdimensional police as the game's setting. Worlds with different mechanics, characters with unique looks, and a huge amount of story ideas that could fit there. Oh, the potential!

It wasn't until a few months later that Alex, our CEO and Double Cross's creative director, asked me if I had an investigation system ready for a pitch due a few days later. I was confused at first, but he explained to me that an investigation system was the piece that we were missing. After all, we were already planning a system of non-linear platforming levels to represent a variety of locations in the multiverse, and were already designing a combat system that'd let players confront and capture criminals and evildoers in a highly customizable way. But we had nothing to showcase the detective work and mystery-solving aspect of our main character's job, RIFT agent Zahra Sinclair. And having a cool investigation system would add something unique to the action game we were developing.

I quickly designed the first iteration of the investigation system. To do that, I took the detective board trope and thought of a mechanic that could fit.

In this first iteration, as you beat the game's levels and talk to certain NPCs you would find items or notes that worked as clues, like you would in Ace Attorney. These were hidden in the levels, like the upgradium is. While in the game's hub level at the RIFT HQ, you could access Zahra's detective board and place clues alongside pictures of characters you'd met and locations you'd visited. You had a limited amount of red strings that you could use to connect clues, profiles and locations.

If an NPC's profile was connected to the right clues or locations with a red string, new dialogue options would be available when talking to them, and you'd be able to unlock further clues, areas, or lore. That successful dialogue changed the red string into a green one, meaning that part of they mystery was solved. Eventually, getting a green string between the masked villain and one of the other RIFT agents would reveal the traitor and unlock the game's final level.

The board would be a nice visual representation of the player's progress, as well as a way of helping understand the complex, non-linear, multi-layered plot we were planning. Furthermore, it felt like a great way of having players fulfill the detective fantasy.

The first iteration worked well for the pitch, but as we went deeper into its details we realized that the system was excessively complicated and didn't really streamline towards the goal of figuring out who the traitor was. So Design and Narrative started to work together in designing a new investigation system. For the next three weeks one of the designers, Justin, and myself turned our computers off, sat on a couch, and started crafting a paper prototype that included the characters, locations and clues for the game.

In the new system we kept the clue gathering throughout the game's levels, but changed what you do with them. Instead of physically connecting stuff on a board, you would make a 'deduction' by connecting a clue and a character profile with a phrase, a system inspired by Detective Grimoire's.

A further simplification we implemented in this system was that each character only had one clue tied to them. Once you had used a clue properly, you would be able to get some evidence that served as an alibi to the character, removing them from the list. You repeated this process until you were only left with three suspects. At that point, the final deduction was a more complicated version of the same system, having to tie three clues to one character through multiple phrases. Doing it successfully would unlock the final part of the game.

As soon as we had the paper prototype ready, we started bringing our co-workers one by one to play through it. We observed where they got stuck, what they found too easy or difficult, listened to their feedback, and slowly improved the prototype after each playtest session.

Eventually, the investigation system underwent several changes, such as dividing the characters in 'profiles' for RIFT agents and 'case file' for the villains. This is a summary of how it worked:

It wasn't perfect, but it was fun and it involved some great character interactions. Our schedule was already pretty tight without spending more weeks working on this system, so we decided to move forward with this.

After leaving it be for some weeks, we came back to it once it was actually time for the coders to implement it. Reviewing it again, Design realized that it was too complicated and many players might not like the difficulty of having to figure out the clues in the current manner, and getting negative feedback such as having to spend time going through menus again after failing could get frustrating. So we decided to start removing some of the intricacies and simplifying the system.

First of all, after some discussion we tried removing the Detective Grimoire-style phrases, which were considered an unnecessary step to verify if players actually understood the connection between the clue and the character. This didn’t sound as a good idea to me at first, but to be fair the changes resulted on positive feedback on playtesting, so we decided to keep the phrases out.

We also decided to have Zahra show the clues directly to the characters, instead of connecting them to profiles in a menu. Besides removing the unnecessary step of having to talk to a character after making the right deduction, this allowed us to write some funny or insightful reactions to the wrong clues, which ended up being a great addition.

One more change that had happened was that clues were no longer hidden in the levels. Instead, they were now found in chokepoints or at the end, and you couldn't miss them. This makes it clearer that solving the investigation is a mandatory part of the game to unlock levels, and not something optional.

Finally, several weeks later, we turned the step of assigning evidence to case files into automatic, which removed an entire tab from the menu and greatly simplified the process. This is how the simplified investigation system works:

And that's pretty much the system you will find in the final product!

In retrospective, some of the simplifications we applied to the investigation system ended up hurting rather than helping.

Some of the changes really work—for instance, merging the Evidence and Case file tabs, or adding more interactions with the NPCs. But others removed the challenge and agency from the investigation. The detective work involved in having to find clues and understand them had been simplified to a fetch quest. This doesn’t sit well with the interdimensional agent fantasy we have aimed for and it’s not surprising that many players miss a deeper system.

At the same time, some other players feel like the investigation is an unnecessary addition to a platforming game that only serves to lengthen the game. I believe this would have felt better if the mechanics had more to do with each other, such as when clues were optional to find. As you can tell from this blog post, the investigation mechanic was developed as an add-on to an existing idea and wasn’t properly integrated—the design decisions of the platforming system didn’t affect the investigation, and vice versa.

The end result isn’t as robust as it could’ve been. We fell in a middle-point between “let’s embrace the mix of mechanics that represent the interdimensional detective fantasy”, where the system would have been more complex and better integrated, and “let’s just make a platforming game”, where this mechanic should’ve been scrapped. And so, it can be unsatisfactory to players expecting either of those.

However, I think that the end result is still interesting and entertaining. There are definitely some lessons we’ve learned here, and I hope we get a chance to revisit some of the discarded ideas in future project and apply those lessons.

(This article was originally published on 13AM's website)