A game without words: The making of Ellipsis

A game without words: The making of Ellipsis
James Batchelor

By James Batchelor

August 26th 2016 at 3:39PM

We speak to Salmi Games about crafting a mobile game that requires as little localisation as possible

Ensuring players understand the mechanics of your mobile game quickly is crucial to perfecting that first time user experience and keeping them engaged with the game.

Unfortunately, some developers have such high ambitions for their mobile games that conveying these mechanics require lengthy tutorials, often with in-game prompts or a friendly character spewing reams of text in order to explain and reiterate every function the title has.

This in turn can increase the amount of work that goes into localisation. Tutorials – and indeed any form of in-game text – needs to be availble in the native language of every market you target. Depending on your budget, this can seriously limit your chances of having a worldwide hit.

Indie studio Salmi Games took a different approach. Its neon-lit action puzzle game Ellipsis has little to no text, confident that its accessible gameplay can easily be picked up by players the world over. 

We spoke to the studio's co-founder Yacine Salmi (pictured below right) about the delicate balance involved in creating a game without words.

Why did you decide to build a game that avoids localisation? What are the advantages when it comes to launching your game worldwide?
We avoided text in our game for a couple reasons. The core gameplay is so intuitive that we were initially curious if we could make it pick-up-and-play without text. And to be honest, laziness was a driving factor. We didn't want to deal with localisation issues or font rendering, topics I really don't enjoy working on. We thought it would save us a lot of time.

The advantages are clear. You can reach markets around the world, even small ones you would never consider localising for. If you do things right, your game should be intuitive to pick up and play, making it more accessible to more players. And launching worldwide and making your game accessible to a larger audience also increases your chances of getting feature.

In the end it didn't save us any time. It took a lot of effort to make the game completely intuitive without text. Worse yet, we still had to localise the App Store descriptions to maximize our chances of getting featured.

To be honest, laziness was a driving factor. We didn't want to deal with localisation issues or font rendering, topics I really don't enjoy working on. We thought it would save us a lot of time.

How did you ensure your concept was simple/accessible enough that it wouldn’t require any text? Did the lack of text limit the type of game you could create?
Through an incredible amount of playtesting and iteration. We took advantage of every showcase we could participate in to show the game to new players and draw valuable feedback from their first time experiences. Getting the introductory elements of the game to feel just right was key to make it accessible and intuitive to the player.

It was key for us get access to a fresh stream of new players. Once players 'get it', you can no longer playtest the early part of the games. 

The lack of text did limit the type of game we could create. It would be difficult to create an RPG without text. Sticking to no text only became a driving force after our prototype took shape. We wanted to explore how we could make use of direct touch controls, instead of more traditional tap and swipe control schemes.

However, not using text led us to create an even better game. Whenever we encountered challenges teaching the player, we couldn't just take a short cut and throw in text. We iterated until we found a good solution. This really helped us create a better game. By the time we released I was confident I could hand a tablet to any player on the street and they would 'get it' within a few minutes without me ever saying a word. 

What was the biggest challenge of explaining the game mechanics without text-based tutorials?
The biggest challenge was conveying the distinction between the game itself and the overworld map: a series of connected blue bubbles the player can select, each representing a level. Each system has a somewhat different control system – touch and hold in-game, but swipe to navigate and touch to select in the map.

We spent a lot of time on this, trying out many approaches. What element should we introduce first? The game or the map? We went with the game first even though it can be jarring for some players to be dropped right into the action. But learning how to play the game was more important than navigating the map.

The map and game look similar, which led to occasional confusion – even though it also helps tie the two together later on in the experience.

Camera control in the map adds another layer of complexity, so we disabled this until the fourth level was unlocked since everything up to that point could fit on-screen. 

There are tons of other little tricks we did to make it a smooth learning experience. There are subtle hints that players slowly pick up over time, even when they are 20, 30 or 40 levels in. 

Players die often, but there are no lifes in Ellipsis. We never punished the player for dying.  You die, you can restart immediately.

How did you avoid oversimplifying gameplay so that Ellipsis still presents a challenge and engages players?
We broke it down into manageable chunks. We introduced each element carefully and individually whenever possible. It was also important to give the player room to experiment. Some of the early levels are almost little sandboxes.

Death plays an important learning role in Ellipsis. Players die often, but there are no lifes in Ellipsis. We never punished the player for dying.  You die, you can restart immediately. And each level is bite size so that you never feel you have lost a lot of progress when you die.

We didn't shy away from introducing complex behavior and gameplay, but we always made sure we introduced the building blocks beforehand. Each concept is carefully taught to the player and then re-inforced.

Finally, we used the entire range of 135 levels to continuously explore new concepts. This kept the game fresh from start to finish and present a wide range of complex ideas.

How well has Ellipsis performed around the world? Has it reached markets you previously haven’t been able to?
At this point we've sold copies in over 120 different countries so it's done pretty well for a small indie title. The game's done well in our traditional Western markets like North America and Europe. We've had really strong results in Asia, particularly in China and Japan but also in the Middle East and Russia. China represents about 20 per cent of our sales – and 90 per cent of our piracy. Our lack of text and approachable art style helped make this happen.

At the risk of going on a brief tangent: on Android we have 30,000 pirated installs in China, and zero sales. We only belatedly learned that the Google Play Store is not available in China. If you’re going to publish internationally on Android, it’s worthwhile investigating publisher partnerships for China. There are many alternative Android stores available, but they’re hard to access as an non-Chinese developer.

What advice do you have for other devs looking to minimise the amount of localisation their game requires?
Make full use of the range of tools at your disposal to teach the player how to play. This is not limited to text. You have visuals, sounds, animations, color – all can provide non-verbal cues and feedback to the player. For example, in Ellipsis we used consistent colors for similar themes: red equals bad, blue is good or the player, green represents speed. 

I wouldn't advise 'no text' as a core strategy, but it helps to keep in mind something like 'is there a way I can teach this mechanic in less words'. Sometimes there is, sometimes there isn't.

If you want to minimise the amount of text needed, it’s important to introduce new elements one at a time. When they do something wrong, you can nudge them in the right direction. When they do something right, you can give them positive feedback.

You don't always have to explain – you can show the player how to play. On mobile players tend not to read text anyways, or they skim through it, so keep that in mind. Explore, experiment, and most importantly, playtest.

I would encourage players to look at some of the great games out there who do a great job teaching players with minimal use of text. Current examples like The Witness on the PC or Leap Day on mobile. Or classics like Limbo, Flower, Flow and many more.

Are there limits to this strategy, or could more game types/genres be designed in a similar way that requires little to no localisation?
I think it really depends on the game genre. It worked for us but I don't think it's a magical panacea. I wouldn't put 'no text' as a core strategy, but it helps to keep in mind something like 'is there a way I can teach this mechanic in less words'. Sometimes there is, sometimes there isn't. I think Threes is a brilliant example of a game that uses a minimal amount of text in a delightful way that adds to the experience.