This week I will comment on some of the feedback we received from the rest of the students during our alpha and beta playtesting sessions and how we approached the design of the game after that. This is a long one, so go get a cup of tea!
Alpha
For our Alpha presentation we found a table that was in one of the ends of a major flow of traffic, thanks to this we received a substantial amount of player feedback and attention from our fellow students. We ran the playtest for 2 hours, in which we got 28 responses in our Google Form.
We had a quite complete set of features we wanted to try out, chief among them were the firing mechanic and the level design implementation. Art-wise we had produced most assets to a testable level. We were missing all sprite sheet based animations and sounds, something that was complained about by a few testers, but was never one of our goals for this stage.
An average of the 28 responses showed that the general aesthetic of the game at this stage was graded 3.96/5 with most feedback focusing on the visual part with comments such as “Add more UI elements for health” and “The player character could be more detailed”. After filtering out un-helpful feedback, this helped us see that we were on the right track with our design choices but had some room for improvement.
To remedy specific flaws such as the controls for the shields feeling uncomfortable and counter-intuitive (they were mapped to the H, J, K, and L keys for the Alpha), we went back to one of the internally discussed ideas, putting them on the number keys 1 through 4. This way, even though we were still keeping a horizontal layout, the numbering of the shields would help the player understand the relationship between keys and gameplay elements much faster at any given point.
The cannon firing proved to be effective although feedback was received that it felt a bit slow to fire the charged version of the laser. An exploit was found by several testers in which you could turn the cannon whilst firing the powerful continuous laser, essentially being able to clear most of the screen with one attack. We considered turning this bug into a feature, but it was counter to our core loop (moving the cannon up and down and tilting to get the best firing angle) and would render the power-up shockwave obsolete.
The level design proved to be overwhelming to players who did not grasp the mechanics of the game, especially if they did not use the shields. The players who did grasp the mechanics on the other hand had much fewer difficulties, although ended up losing anyway as there was no end-state for the game at that time. This helped us notice two types of behaviours, the “panicked” players who would shoot as soon as they got the chance and would constantly wiggle the cannon but forget they could move it up and down and in some cases even forget about the shields, and the “strategic” players who would try to line up shots and consider whether to use the shield to block an attack rather than waste time shooting it. The second group was our target audience, but we still wanted the other group to have an enjoyable experience, so one of the major efforts going forward would be implementing visual and audible feedback as well as refining mechanics to make them feel more responsive.
Beta
For our Beta presentation we did not have such good luck when it came to positioning, we ended up in a bottlenecked area where most students would just try to squeeze through, not paying attention to the fact that there was a group there. We had to actively seek out testers this time. This tied with the fact that the game was longer made it so that we got less responses overall, for a total of 21 over a 2 hour open playtesting period.
Unfortunately, due to some issues with implementation and game-breaking bugs, we were not able to showcase our full progress during the Beta playtesting session, so a large amount of the feedback we got was not useful to us since those issues had already been corrected. For example, there were complaints about the visual cues of the cannon not being clear and suggestions of a charging animation and similar features (essentially being in the same state as in Alpha), this was already done and could not be implemented in time for players to experience but was solved for the formal Beta presentation.
Player cannon through stages: Early Mock-up, Pre-Alpha, Alpha and Beta.
Our main focus for this playtest, given the limited changes we were able to introduce, was mainly to test the new level design and the new reload mechanic. We also wanted to test the music and sound effects that were introduced with this build. The results were mixed, with some comments referring to old issues as explained before and some genuine critique for the new systems we were asking feedback for.
The aesthetic rating dropped from 3.96 to 3.85, but without a large enough response volume and follow through questions this data could not be used as a reliable metric, only as an approximation, and due to the fact that the question was not phrased exactly the same way in both questionnaires, we concluded that the opinion on the aesthetic remained largely unchanged, which was consistent with the fact that almost no visual changes were introduced between versions. Further testing would be useful, taking into account factors like the wording of the questions and computing for standard deviation whilst keeping in mind the potential errors due to a small and biased player base.
The opinion on the mechanics and controls improved, showing that the changes in timing and inputs were successful and helped the player accomplish the goals of the game with greater ease. Minor corrections to the shield activation timings were introduced after the playtesting, this was something not planned for by design that ended up in the Beta build due to a miscommunication error.
Level-design-wise the Beta proved to be much more successful, we had an impromptu tutorial in the start screen (which was wholeheartedly implemented by our lead coder 15 minutes before the test but unfortunately received a lot of negative feedback due to poor readability) and the enemy waves were designed for, as opposed to the random groups of enemies present in the Alpha.
Conclusion
Playtesting is a vital tool of game developer’s kit that should be used often, even if in small internal groups. This is something our team could have been better at and I regret not having had the time to playtest our own game more often.
I will also strive for better data-gathering when holding testing sessions, as asking the correct questions is vital for the understanding of the player’s needs and feelings towards the game.
Game Design & Stuff 