Lost & Found’s secrets

By Phil Stuart, How we did it

Following Disney Fairies: Lost & Found’s Christmas release, we thought we’d share how we went about developing it in Unity.

Creating a fun game for young teenaged girls can be serious business. It’s hard not to fall into the usual tropes especially when working with a big brand, to a specific set of criteria. The hidden-object genre seemed really applicable to the brief. It was challenging but not stressful, could be played cooperatively with friends and family and they makes use of rich environment artwork. For us, it was also a known technical challenge which felt very important given the short 3 month development timeline.

Disney liked the idea, and so we began our education in the Hidden Object genre.

The feature set

The feature set borrows from many of our favourites in the genre, whilst being careful to retain the uniqueness of the proposition. The final features that made it in were:

  • 7 unique levels
  • A 3-star grade system
  • A multi-tap combo bonus
  • 7 Intro and outro dialogue scenes
  • 6 Interactive segues between each level
  • A journal of unlockable Fairy facts
  • Interactive hint system
  • In-app purchases for future level packs

One of the most hotly disputed issues was “added value features vs simply making more levels”. On the one hand, added value features do just that – make the player feel like there’s lots to do and see, raising the perceived monetary value of the game. On the other hand, isn’t true value gained from the depth of gameplay delivered by more levels? Strategically we veered towards doing less better. Doing just seven levels meant a neat correlation between the seven Fairies, and the quality of the core would allow future seasonal updates to increase the depth of the proposition.

Level design

It turns out it’s a challenge balancing the difficulty of a hidden-object game when you’re the one hiding the objects. This problem is endemic in all puzzle game genres and our somewhat brute-force solution was a steady, constant flow of testers throughout the project, mainly recruited from friends and family, bolstered by formal testers towards the end of the project.

Our system for level creation was to hide a varying percentage of easy, medium and hard objects accross the seven levels. Despite this, we found that the biggest factor in level difficulty was the background art used. For example, one background was frozen over – we froze the objects to match the background and they became virtually invisible!. Thanks to a relatively loose narrative structure we were able to grade levels based on difficulty and then rearrange them to suite.

Another variable we had to help us was the number of objects to find in each level. While this does not affect actual difficulty it lengthens the play session, increasing or decreasing perceived difficulty as required.

Every level houses at least double the number of objects needed for completion, with a “shopping list” of objects created at random for the player every time they play each level. This ensures that while not infinitely replayable, the game stays fresh if the player wants to go back and try for a higher star rating or unlock all the bonus content.

The game world

Whatever we created for Disney, we wanted it to be a ‘best of breed’ product and create a sense of wonder for the player. The idea of a kind of panoramic world had been floated in a few internal meetings, and before long we had a rough prototype to take to the client. Even using placeholder assets, it was clear that this was the direction to go in – the prototype had people instantly swiping around, eyes glued to the screen, even without any tasks to perform.

Here’s how it works:

  • A background image sits on a spherical segment to create the 3D panorama
  • Objects placed over the background matching the topology
  • Atmosphere particles add depth and subtle motion
  • Lighting particles enhance drama and add subtle motion
  • Camera operates on a fixed pivot point to complete the 3D panorama effect

It’s really just like the old model dioramas that can be found in places like the Natural History Museum, that use a curved background to create a false sense of perspective.

The system we use to create the effect is in fact very simple. The background image is mapped onto a segment of a perfect sphere, using the default sphere UV configuration. This creates pinching at the top and bottom of the image, but as the image is wider than it is tall, this is not a major issue.

To change the magnitude of the 3D effect one must change the proportion of the sphere being used. Unfortunately for us this meant going into a 3D program and creating a new sphere by hand. In the end, this was not a problem for the project but it was definitely a risk point – valuable project time could easily have been eaten up making small adjustments here.

Conversely, the strength of using geometry to create the effect as opposed to a fancy screen-based shader effect is that it’s universally compatible on any device that can run the Unity engine.

Level production pipeline

Every level we created using the panorama effect had 4 stages:

  1. The level is “designed” in Photoshop, that means selecting a Fairies background and integrating a selection of Fairy objects using lighting and masking.
  2. This Photoshop file is deconstructed, extracting all the objects and the background as separate PNG files.
  3. The level is reconstructed in Unity, with the background mapped onto the sphere segment and the objects mapped onto multiple, positioned planes.
  4. A beauty pass is performed, adding effects to simulate floaty pollen, haze, lights, running water etc.

Levels were often tweaked after testing to make objects easier or harder, or removed completely. Unity’s modular approach to project file structure made the amendment process easy to handle – designers could be hard at work updating PNG files while the devlopment team were debugging scripts related to the same levels without any fuss.

Audience testing

To ensure Lost & Found worked for the target audience we carried out user testing with at Oakdale Junior School in South Woodford, London.  Equipped with iPads and a beta version of the game our focus was around the core gameplay associated with the hidden object levels, whether object names were understandable, and whether the second by second gameplay was balanced in terms of fun and difficulty.

The game was tested with several groups of 8 year old girls (the game being aimed at 6-9 year old girls). Although not selected for their familiarity with the Disney Fairies brand, the groups were largely familiar with the product, and were genuinely excited for the chance to play the game.

In the first of two exercises we focused on a card-sorting activity. Three sets of ten objects had been selected from the 120+ that feature in the game. Each object set comprised ten cards showing object images and ten cards with object names on them. The names were introduced first, with the group invited to explain what each of the names meant to them. Anything not understood, or that was interpreted in an unexpected way was noted. After all ten names in a set had been introduced and laid out on the table, the ten associated images were introduced and the group asked to pair them up.

With card sorting complete the game was introduced and the tablets put in front of each group. They were immediately engaged by the game and both groups were seen to work together to find the objects as they played through the first three scenes of the game. As the groups progressed they were also pointed towards elements of the UI they didn’t appear to have picked up on.

The findings

In the card-sorting exercise the two test groups achieved a 95% success rate in pairing objects, despite not being familiar with all object names when they were introduced. Ultimately we used this information to improve clarity of several object names in the game.

The gameplay test went very well, with both groups instantly familiar with the controls and gameplay objectives, and completing each of the three levels in times of three to four minutes. Given that all the testers were towards the upper end of the target demographic and were working in teams of three, we were comfortable with this speed of completion.

It was observed that some elements of the UI were not quickly picked up (such as the hint system and journal button). Consequently we implemented improvements to ensure such UI elements were more visible. For instance we added occasional animations to the hint button if the player hasn’t found any objects in a given period of time.

What’s next

Building on the success of launching a #1 Kids game on both iPhone and iPad discussions are underway for an Android version, new seasonal updates and rolling it into new territories. Watch this space for new levels, new games modes and even more Fairies’ magic (sic).

Phil is the co-founder and Creative Director of Preloaded. A fanatical gamer and a champion for the power of games which do more than just entertain, he is responsible for the studio’s ‘Games with Purpose’ vision.

Previous post

Games for Change #9

Next post

Amplify & Preloaded