Minecraft Dungeons: behind closed doors and Gamescom Accessibility Impressions


Aspects of the game discussed below may change before or after launch.


I've never played Minecraft. Most people would be surprised at this, but when you're a gamer without sight, up until recently (when Minecraft added full menu narration), the blocky worlds and intricate building was almost completely off-limits.

To be clear, I use the term Gamer Without Sight (GWS) as "legal blindness", often just shortened to being "blind" can and often does include usable and/or residual vision, which I've never had.

With the announcement of Minecraft dungeons (MD for short) at E3, I was very much interested to see just what might happen with the accessibility of this series.

An Unexpected Playtest

As it so happens, I recently had the opportunity to play a work in progress build of the game behind closed doors, alongside Nathan Rose (one of the individuals working on the title) and a fellow accessibility advocate and consultant, Mitch Muenster. This took place around the time of Microsoft's Gaming and Disability Boot Camp on Microsoft's Campus in Redmond.

After a set of demos where Nathan played the game and discussed audio cues, I was already impressed at the level of detail the team had put in, including cues for filtering item types, footsteps, enemy sounds and action cues. Text To Speech (TTS) of any kind wasn't in this build, but I could certainly see the potential for this game to be accessible as a single player experience, even in this state. In particular, an aspect I noticed was absent and that would be of use as a gamer without sight was basic pathfinding, where an audio cue can be used to guide the player towards an objective, similar to how Gears Of War 4's Fabricator Ping functions. Microsoft Research India (MSRI) have been constructing a Unity plugin that allows for such functionality to be implemented in any Unity project, and seeing tech like that as part of a full experience could, I feel, definitely change the landscape of accessibility for the better, even if the concepts have to be ported to a different engine.

The game, being played from a top-down, isometric view, has its issues when it comes to navigation. After Mitch, who is fully sighted, created a small representation of the view via a few Lego bricks lying around in the room and Nathan had some more feedback from both of us on some interesting challenges and possibilities, most of which would be easily resolved if given the time and resources to tackle them, Mitch proposed an interesting test: he would be, in essence, my CoPilot, without a controller.

Up to this point, nobody had played the game with a controller, all Nathan's actions came from the keyboard of the laptop running the game. Once I had a controller in hand though, the discovery process began.

As well as left stick being movement, right trigger was used to shoot an arrow or charged shot (if held), A was melee, X activated a shield and Y activated a fishing rod, which I didn't test during my time with the game. B also would've activated an item, though I never got the chance to use this either.

This build featured autotargeting when enemies are in range, though there were no sound cues for locking on or breaking a lock. This meant that Mitch had to call out that I was locked on to an enemy or when there was nothing to target when I attempted to aim. Not necessarily an ideal solution, but one that worked given the setup.

Interestingly, enemies occasionally made sounds when offscreen, but usually only a single time (possibly once they saw you, though this wasn't easy to confirm). As much as this could be frustrating for melee attacks, it also had the upside of allowing you to easily take down enemies with the bow, at least if the target was in lock-on range.

Mitch and I had to establish how best to communicate information (which can vary from player to player), as well as roughly how to move and get around some frustrating camera angle issues (including being led across a bridge which, in this build, you could fall from with the need to respawn which was only possible with console commands). These bridge segments were handled by Nathan for time's sake. I was able to navigate with relative ease around various geometric challenges, take down multiple groups of enemies and pick up items, all with Mitch's assistance in small amounts, though not as much as I'd have expected from a new title.

Even though I knew that this wasn't the final build of the game, I was definitely very much impressed with the speed, overall pacing and my ability to pick up the title, even with the lack of information as compared to the visuals (which I fed back to Nathan at numerous points during the playtest). I think, for me, the most interesting point came, not during my time at the helm, but when I watched Mitch take control.

Why, you might ask? Given my experience with the game (about 20 minutes or half an hour of play time)?

I was able to discern pretty much what Mitch was doing, even just from audio, calling out enemies off-screen, items that he was picking up, enemies he was defeating, etc, all with just the aid of audio cues demonstrating that even though the game wasn't fully accessible, it was an experience that could be said to be enjoyable and collaborative.

all this being said, with time and effort from the development team and with individuals like Nathan willing to listen, I had high hopes. I knew that once features like audio pathfinding, full menu and screen narrations (particularly important for inventory management and other crucial gameplay elements), as well as additional audio cues and haptic feedback to go along with them made it in, this might just be the most accessible Minecraft experience yet. I was looking forward to my next chance to try the game, though I didn't know when this would be at this point in time.

Gamescom 2019 build

At Gamescom 2019, I had a great opportunity not only to play MD once again, but to play it alongside one of the lead developers of the game. Other than the fact that I remembered several buttons (surprising even myself), the developer was impressed at how much I'd been able to pick up on, including enemies coming in from different angles and their distance from me.

This was facilitated by a small but substantial change I noticed, namely the addition of greater frequency in enemy cues and haptic feedback for being hit. These two changes in conjunction meant that I was able to survive all but the harshest environmental hazards, whilst being navigated by the developer using the same controller. This was comparatively similar to how myself and Mitch worked together in the first playtest, but armed with the recollections from my previous effort, I was arguably much more effective. Again, TTS of any kind wasn't in this build, but I was assured that it and other features were definitely being worked on.

Nathan, who was also present at the event, did clarify that the development team is also working on getting audio navigation cues into the game, even though they were not present in the demo on the show floor. Even this addition would allow me to smoothly play as part of a team and take the lead in scenarios where I would perform more effectively than the rest of my party, for example if I had a more suitable weapon, so I'm definitely interested to see how well that works out during any future opportunities to test the game.

Such passion and dedication from a development team is amazing to see and highlights an interesting and important point: Hearing from the development team themselves that accessibility features are being worked on for a game, as well as what said set of features includes, is, at times, just as encouraging and positive as seeing what's in said game already. Even having information on what's missing from a demo that already exists in newer studio-only builds is also useful, as it demonstrates that developers understand what's needed to make for a more fully-fledged experience in terms of accessibility and the game in general.

I'm very much looking forward to creating more content surrounding the game as it gets closer to release, including providing what information I can on the accessibility of this very interesting title.

Minecraft Dungeons is scheduled for release in Spring 2020 and I'd like to thank Mitch and Nathan for, respectively, participating in the first playtest and enabling me to take this great opportunity to test the game in the first place.

Back To The Main Reviews/Guides Page