Iterative Testing in GameCity’s Open Arcade

We’ve just returned from the eleventh GameCity Festival in Nottingham, with many a device battery recharging and our larynxes still recovering from the traditional “Marioke” afterparty. We were there, of course, to show Dragon Queens – enjoying its first ever public showcase as part of this year’s fringe Open Arcade.

pa280023

GameCity has long sought to bring games to the general public, shedding particular light upon the act of game development. It’s that focus which gave us room to not only prove our concept amongst Dragon Queens‘ target audiences, but also to ‘jam’ the event and iterate rapidly upon the game’s UX, level design and game balance. In short: we arrived at the festival with a rough prototype, and found we were able to polish it through 5 builds over the course of three days.

Proofing

The version we tested was stripped down to its bare essentials, incorporating some aspects of Dragon Queens‘ intended UX and UI along with ‘hotseated’ multiplayer and a focus on the game’s core mechanics:

  • 2D, top-down perspective (as opposed to a 3D world and a relatively fixed camera)
  • Single-device, ‘hotseat’ multiplayer (as opposed to options for single-player, or networked play across two nearby devices)
  • No ‘dragon flight’ mode (which in the final game will serve a variety of optional functions)
  • First pass at a standard, visually clear and uncluttered UI (as opposed to a broadly customisable one)
  • A single, straightforward level map, intended for sessions of no more than 20 minutes

For every point save the available play modes, we were able to test our initial build, discuss its execution with visiting players, and then either work on a fix in parallel to that or incorporate those changes into a new, nightly build. In practice this often meant Gemma would chaperone games on one device whilst Delia would prepare a build for another identical one. We would then swap devices between games – ready to identify new and old bugs, and condense what is already quite common exhibition practice into cycles of an hour or even less. For such an early stage of development as ours this was invaluable and far more efficient than we had first expected.

Most of our changes were made to the UI, ensuring that players were being presented with enough information to play without the need for an interactive tutorial. This often required the use of text labels - something we intend to reduce again in future iterations.
Most of our changes were made to the UI, ensuring that players were being presented with enough information to play without the need for an interactive tutorial. This often required the use of text labels – something we intend to reduce again in future iterations.

It’s worth saying that this was only ever likely to work with either very early proofs-of-concept (as ours is), or a comparatively simple game. Although we were able to create or tweak a handful of assets ‘on the fly’, anything more than these surface-level edits would have required days between builds. It was also to our advantage to find that – due to the manner in which GameCity’s Open Arcade was arranged – public footfall was comparatively small, certainly when compared with the likes of an exposition.

Iterations

With Delia working on new builds for about two-thirds of the total time we spend exhibiting (as well as a couple of nightly builds), we managed:

  • a complete re-positioning of the game’s settlement information panels
  • a re-drawn demo level, significantly reducing both the session length and the amount of time players spent in isolation of each other
  • a rough but effective, new UI for displaying settlements’ needs and produce – further reducing session length as well as cognitive load on the players
  • partial implementation of a new level feature, to help provide players with a safety net against an aggressive opponent
  • as well as the identification and fixing of a dozen or more straightforward and ‘edge case’ bugs

The end result was a Saturday prototype which was leaps and bounds ahead of what we presented to players on the Thursday. We were fortunate enough to be able to test this version of the game amongst new and returning players alike, further validating our choices and enriching feedback on the gameplay, UI and the concept itself.

Although this proof-of-concept will stand at some remove from the final version of Dragon Queens, it’s been gratifying to test so many of its ‘soft’ features – and the underlying market potential – in a niche space like GameCity. We can now look forward to putting our next build out to a wider public, having already incorporated valuable feedback from everyone who played in Nottingham – to whom we are of course extremely grateful.