Playtest Results

Thursday, 27 October 2016

The first wave of playtests has ended now, to promising feedback. Our initial playtest protocol was refined across the many playtests until we reached a smooth streamlined process, complete with local recording of matches for later review, a lengthy questionnaire covering various aspects of the game, and free-form comments from the playtesters.

Major points of feedback:
  • The Unit Watcher was not well understood. Pictured right, the watcher shows the same basic 'selected unit' info as any other conventional RTS, albeit in a more compressed format. Unit name, picture, health bar, defense, and attack damage against all other unit armor types is listed out vertically. The small blue dot indicates what type of armor this unit itself possesses. Sadly we are taxed for HUD space and cannot display this data anywhere else or in an expanded format without nearly a complete HUD overhaul, but I agree there is room for improvement.
  • Terrain defense and speed bonuses were also not well understood. Another problem area for playtesters was the effects and benefits of terrain on their units. Unique to SCS' gameplay, the type of terrain a unique is moving over widely affects its speed and defense. Continuing the Starcraft comparisons, a Stalker will always move the same speed anywhere on the map, and always deal the same damage any target it shoots. In SCS, moving through trees and mountains will buff up your defense with the cover they provide, while greatly diminishing their move speed. Inversely, roads make units move faster than they would otherwise, while providing them next to no defense against unexpected attacks.
    These two areas struck us as problem places that require attention, because these speed and defense buffs are integral to the core gameplay design, not something we are about to just toss out because playtesters didn't understand them. They need to be represented in such a way that its obvious and natural that a marine moves faster than a tank would in Woods or that a Tank cannot pass over Mountains the way a soldier can.
  • Some units were God Units. Part of playtesting is always the balancing, and while our initial balance was good, nothing shows its flaws like outside players given a chance to try to exploit it. Very rapidly we were shown which units were 'god units', capable of beating nearly any other or being too cost effective. Sidecar Scout Bikes rapidly proved themselves to be the single most powerful unit in almost all matchups, in large numbers even overpowering tanks meant to be their counter. Healing units such as Medics and Engineers also rated highly due to their ability to keep said Bikes and other units alive a bit too effectively, able to out-heal most incoming damage with ease. A quote from one of the playtesters:
    "Rush bikes, I win if my bike comes up before he can defend against it, I lose if he counters it."
  •  Without CO powers, it was easy to build game-winning momentum. In the current tested build, the key element meant to counteract steamrolling momentum is not yet finished so wasn't implemented, and as the players become more skillful with their unit composition, it showed. Rapidly capturing half the map/s points proved the income difference was usually enough to guarantee a win depending on player skill, as there were still some dramatic turnarounds when players out-positioned their opponents.
  • Improved unit control was the most requested feature by far, as basic commands such as Hold Position were barely implemented, and others such as Patrol completely nonexistent. Better player information also was highly rated, tooltips being second most requested as without previous knowledge of the game, there is very little to tell you what that an Attack Dog can't attack tanks, or what the tank with the long barrel is a Tank Destroyer.
Major planned changes consist of:
  • Locking the mouse cursor to the game screen. A common problem almost all the playtesters had was with the fact that the game doesn't lock your cursor into its own window. It runs in a browser but we as devs had become so used to using arrow keys to pan the camera that we didn't realize how much people relied on edge panning.
  • Tweaks to Attack-Move behaviour. Another oversight of ours was how ingrained the RTS standardized controls are to RTS players. For example, in Starcraft, you press 'A' to order an attack, then left click the location or target you wish to attack. in SCS, A is held down and then right click is used to issue the order, like any other targeted action. This confused several of our playtesters who were used to Starcraft's style, and while both have merits, it might be easier to follow their conventions.
  • Better Visual Feedback. Given that none of the playtesters in this circuit had experience with the game before, we found it very common for them to miss things that we thought were 'obvious' after having played the game for 1,000+ hours ourselves. Your headquarters being captured was the most prominent, as its loss spells your defeat. Playtesters wouldn't notice it being captured because the only 'announcement' of this was just a sinking bar next to the structure indicating capture progress.
  • Unit-building Structures beyond the HQ were universally ignored. This one is a bit tougher to rectify. In most matches, even when facing an opponent attempting to play a teaching match, playtesters wouldn't realize that structures other than their initial HQ could be selected and produce units, even when watching units appear from no where atop those structures for the enemy side. Even the experienced players in times of stress would default to only building at their HQ, instead of the multitude of Factories across maps.
  • Mousewheel to Zoom. This one is a bit tentative as it has to be done right or else the pixel art graphic style will get all blown up from mis-scaled pixels. We've squashed dozens of strange pixel glitching and artifacting in the past that scaling for a zoom is a can of worms to open up again.
Thanks to our rigorous quality controls, bugs were few and far between. High lag occurred when a match was run overseas from Canada to the UK, but beyond that there were next to no desynchronization issues. Occasional graphical glitches such as sprites overlapping that shouldn't popped up, but were easy rendering order fixes to be squashed,

As for moving ahead, the waters are still murky as we plan out when/how public releases will be done, the development schedule of features new and old, and evaluating this feedback and its benefits to further development.

No comments:

Post a Comment