Our last game, Sol Survivor, was a very easy game to design levels for.  The process had relatively few gameplay significant variables, and the editor did most of the hard work by generating terrain for us.  Retrovirus, though, is a whole different animal.

To start, Retrovirus is being developed in a wholly new editor, Paragon.  In Paragon, we create spaces by spawning in basic shapes, and then manipulating them in chunks.  The end effect is that working in our editor is a lot like playing Minecraft, where we push and place blocks to create or restrict space.  Unfortunately, with each block measuring 4m by 4m, individual blocks were too large to be used for fine level detailing.  Any level where we used blocks like one might in Minecraft to create structures or level details ended up so large that the player felt slow and small by comparison.  Obviously, this was not good design!

Correcting this meant turning to a more scientific approach to level design.  We created a version of the classic Counter-Strike map de_dust in Paragon.  Dust was a setting that all of our potential testers would instinctively know.  Critically, it was also one where where our playtesters would be more likely to notice problems by comparing our version to their decade-long experience within that space.

Shortly after a test flight through dust, it became clear that specific parts of the level were out of scale.  Ledges in Counter-Strike that come up to thigh height were upwards of 12 feet high in Retrovirus, despite the player moving at a speed similar to that of CS.  The hall connecting the Terrorist spawn with Bomb Site A was too wide and too long because we’d used a 4m by 4m by 8m column to simulate the crates that normally provide cover in that space.  It took an examination of Retrovirus through the lens of prior game experiences to properly expose the scaling issue to the whole team.

Experimentation with de_dust revealed the problem, but we still needed a practical way to ground our level designs while we started principal production on the environments for Retrovirus.  We created a scale test level, built to put slices of various environments side-by-side as a sort of level design primer.  With our scale test level, we soon had a clear basis for comparison between environments, helping the player to feel appropriately nimble in the world.  We’ve since used that experience to set the tone for all of our future levels, and it feels like we’re moving toward something more like the game we all want to play.

Only now do we have more levels we intend to keep than levels we’ve thrown away, and that’s after a solid month of two full-time team members cranking away.  It’s important to fight the urge to call that time wasted.  Every level scrapped has been a step toward finding what’s fun about Retrovirus, as we rise to the challenge of designing levels for six degrees of freedom.