Welcome to 2048 usability evaluation

As part of a benchmark series, we will look at many 2048 games out there to see how they compare, and highlight best practices swell as common pitfalls for this type of games.

Next up, welcome to 2048.


welcometo2048-tutorial-swipe-betterThe game starts with the tutorial and leads the player really quickly to the play button. The user flow is very smooth and clear. The first screen teaches the player to swipe, which makes browsing the tutorial screen smooth. The tutorial itself is very visual and can be understood even without reading the textual information.

Notice how the explanation fo the first interaction is in the same direction as the movement needed for the tutorial: this avoids confusion and reinforces the lesson learned.

Player strategies

welcometo2048-merge-sequencePlayers make different types of decisions. When completing a sequence with a discrete goal, players will decide each step of their action, for example, when they’re near reaching a higher score, and make a conscious effort to reach it. This type of behavior is characterized by slower interactions, and pauses. The player makes a decision between each gesture to define the next step.

The human brain is made to simplify complex decisions, and limit the number of decisions made to reduce the cognitive cost of an activity. The more the activity is familiar, the more the player will make decisions about behavior patterns, rather than single movements.

This works pretty much the same way as walking : at first, a baby has to focus on each step and manage his balance. Once he becomes an expert at walking, he only needs to decide whether to start or stop, and doesn’t need to pay attention to the movement his body carries out to keep in balance and move forward. This leaves his mind free to make different types of decisions, such as compute the route to the destination.

welcometo2048-feedback-special-noticedIn 2048, one of the strategies is moving numbers in a single direction and switching only when that doesn’t work anymore, until it works again.

For example, the player will make a gesture UP repeatedly, until no more movements are available. He will then move the board LEFT or RIGHT to unlock the situation, then resume the always UP motion.

It will also usually take some time for the player to notice when the repeated UP motion doesn’t work. This is because the gestures are automatically repeated. The player’s attention isn’t fully required to do so, the cognitive load is less important, which leads to more input errors before the player realises he needs to make a decision again.

Inputs & interactions

welcometo2048-input-not-straightThe player tends to make oblique movements, rather than clean, straight up or down gestures. The player will make a gesture that will have no effect a couple of times before they actually get the issue and take the time to make a more precise gesture, which will work. In doing so, the game requires the user to pay attention to his interactions, which removes focus from the game’s problem solving challenge.

welcometo2048-input-not-workingIn welcome to 2048, sometimes even straight gestures are not taken in account, as seen on the following screenshot. The movement amplitude is very small here, so the player isn’t aware that the movement is oblique, again. Here, the player fixes the gesture by making a more ample movement.


The game shows animated, subtle feedback thorughout the screen. The score moves slightly forward and appears bigger when increased.

The exact change in the score isn’t highlighted, and the player has no aid to compute the amount of points he gets from each successfull movement.

Appearing tiles grow bigger from the center, which helps the player notice them more easily and understand their apearance.

The merging of tiles isn’t highlighted by an additionnal feedback, but the color and number change. This allows the player to interact quite fast with the game. The player might fail to notice some changes however. There is also no feedback on errors, when the player tries to move in a direction where nothing can happen.


Whereas the exact score increase isn’t highlighted, the game emphasizes progress in the game as achievements, when the player reaches a new type of tile. This appears on top of the screen, with an animation that makes the dialog very visible, without being intrusive with the playing session.

Menus and options


Leaving the game is pretty easy. A nice animation introduces the menu, which gives a sense of polish, and the use of icons allows users to understand the options without reading / translation needed.

Yet, the animations are brief and smooth, so the player is not frustrated by wasting a lot of time on useless content if he needs to close the app fast.

What did we learn in this game?

  • Tutorial slides are a quick and easy way to teach basics, even better when they are this visual, only have 3 slides, and they can be manipulated with the same interaction as wht the player will need to do in the game later
  • Player inputs are sloppy, specially when they’re experienced and use strategy decisions over single action decisions. Oblique inputs need to be dealt with to allow players to focus on the problem solving, rather than how to properly interact with the game
  • Short feedback animations help notice and understand changes in the interfaces: tiles that appear, that merge, input errors, achievements unlocked…
  • Score highlight without focus on the amount of points forces the player to do the math. It’s ok if scoring isn’t our main focus, but if it is, the player needs to know how he can score better, faster, stronger