Windshields and Pacing


Decisions, Decisions

I first heard the term from Darren Grey of Roguelike Radio fame. I seem unable to find the tweets now but he was trying to convince me that “Windshield Enemies” were good for game pacing – that is, enemies that the player can basically smash through with little expenditure of time, effort, or resources.

Aside: As mentioned below in the comments, Darren feels a bit misrepresented here as he was not so much advocating windshield enemies as opposed to commenting on their design purpose. Apologies! (Still no hex tiles, though.)

I did not care for this idea. I wanted every enemy to be some sort of non-zero challenge. I wanted every step, every action to be thought out in advance or DOOM, DOOM shall befall the player for their carelessness. I feel like I more or less achieved this in Steam Marines, and I thought it made the game interesting and replayable.

Aside: Some quotes about Steam Marines that I enjoy to this day.


“There are a lot of tactics. It’s such a brutal game and it makes it even more brutal that you didn’t realize how brutal it is when you start. Then you realize you died because you made one, literally one, wrong move.”

– Joe Fisher,


“Steam Marines is a dark and compelling game, which will force you to learn to play the hard way and make you deeply regret every mistake you make.”

– Kevin Read,


Players must control their Steam Marines and try to get as far as possible through the ship’s elaborate, procedurally-generated levels. When I say “try”, I really do mean it.”

– The Wargamer



– Jesse Cox

But enough of my gloating. What if I’m wrong?


What is Best in Life?

To win or die trying, I think most permadeath players would say. Or something close to that. It is central to procedurally generated, permadeath games that your deaths should be meaningful and of the player’s own doing. Some people do enjoy being instantly and fatally messed with by the random number generator, but I sort of feel that type of design is degenerate. I’m on team “It Has To Be Your Fault.”

But back to windshield enemies. The reason I resisted Darren’s argument was because I didn’t like the idea of the game letting up. (Permadeath) games are supposed to get harder as the player progresses, correct? The player gets stronger, the game gets meaner. Power curve et cetera, et cetera. But what if windshield enemies are a tool to lull the player into a false sense of security?

You’ve killed a Big Baddie, and you’re on a high. You got some great loot and your character is stronger than ever. Next level. Ah, crap what is this? New enemies? Gotta be careful… and you steamroll them. Huh. Fluke? You steamroll them again.

Okay, okay, okay maybe the developer didn’t balance this part that well. Maybe that loot was randomly above the curve. Maybe you get a little complacent. And then you die miserably.

That does sound like an experience I’d like to inflict on people who paid me money.


Hard, Hard, Hard, Easy, Harder, HARDER

If this is a tool for pacing, then it’s linked to time. Time to overcome the hard stuff, time to get complacent, time to smack the player in the face at that critical point. But shouldn’t a game be filled with interesting decisions? That’s what I’m told and I think I believe it. Offering easy challenges seems to run contrary to that idea. What are windshield enemies if not uninteresting decisions? Whatever you do, you can just squash them; failure rate essentially zero.

I think I’m in the minority here, but I really love the idea of bonuses and penalties coming hand in hand. It didn’t poll very well in Steam Marines – players were adamant in their pursuit of strictly superior upgrade paths – but I think there’s design space there left unexplored by my acquiescence. I think you just have to approach it thematically as opposed to mechanically to get players on board.

Steam Marines 2 deviates from its forebear in that ammunition is a single value shared across the entire squad. Your Support class marine rips off with her machine gun? Squad ammo deducted. Your Leader class marine fires her shotgun? Squad ammo deducted. Marines can no longer melee. If squad ammo gets too low marine damage and accuracy suffers.

This makes easier enemies still a bit of a hassle because you must still expend ammunition on them. Boss/elite units typically crop up later in missions and this will be a consideration. Is maneuvering around enemies a better idea than engaging? Maybe this is not appealing to all players – moving is generally less satisfying than shooting in move and shoot games, after all.

It also begs the question of is it better to constantly force the player to regulate their ammunition or not. Is that simply micromanagement and tedium?


Gears Within Gears

I wanted to remove class talent trees from Steam Marines 2 for months. I finally did this year and it felt great. The system was overly complicated, had too many builds, and made the UI a mess. XCOM EU/EW/2 has soldier talent trees and they work well in that game. The difference, I think, lies in the way they approach microdecisions.

XCOM asks you a question per soldier per turn: Where is my best position with regards to cover?

Steam Marines 2 asks you a question per marine per turn: Where is my best position with regards to retreat or hitting multiple enemies?

Both questions are driven by the same overarching tactical considerations of player unit survival and enemy unit destruction. The main differences are that XCOM has a streamlined action system with regards to movement, attacking, and weapon reloading. Steam Marines 2 has a bit more granularity with movement and attack since they operate on separate resources and there is no weapon reloading mechanic.

There is a much higher chance of death in Steam Marines 2 because there is no cover system – a marine simply relies on 1) enemy accuracy stats, 2) her armor, and 3) range and line of sight to survive. But marines can dance. Marines can move, attack, and retreat behind a corner all in one turn. This is something XCOM does not have – everything is rolled into the cover system and flanking. The class abilities are there to bolster that single impactful thing a soldier can do each turn.

Steam Marines 2 has single tile corridors and doorways and terrain that cannot be destroyed. This means marines can guard chokepoints more effectively and can even completely block pathways entirely by standing on specific tiles. XCOM is much more open, as anyone who has winced as a Thin Man dodged three reaction shots, jumped into high cover, and spit poison can attest to.

XCOM EU starts off relatively easy. It throws a few Sectoids at you and on lower difficulty levels this should pose no real challenge. XCOM 2, I feel, learned from this mistake. It starts off with a brutally hard first mission where a revamped Sectoid can singlehandedly wreck your squad. It establishes what kind of game it is right off the bat and let the rest roll in. Granted they may have had the luxury of that because it is a sequel.

Steam Marines 2 starts off… well, I don’t want to say, really. Let’s just say there are no windshield enemies as of yet…



… which is not to say I ignored Darren. There are pacing elements in Steam Marines 2, just not within the tactical combat. There’s a ship interior/roster area where players can review their squad. There’s a Universe View where you can control a human steampunk spaceship and engage in marine dialogue. I also have further content planned for that little game branch.

Aside: Steam Marines did not take place anywhere but inside ships. Having access to human ships, alien ships, robot ships, and being able to fly around in the universe? That opens up a lot of potential content. It makes me a bit giddy.

*Is promptly eaten by the scope creep monster.*

So the player gets a chance to breathe, relax, and retool before heading off on a new mission. Steam Marines did not have that – it was just 30 levels of more, more, more, New Game Plus! GO GO GO!

So there you go, Darren. I took your advice a little bit. I hope you’re happy. No hex tiles, though.

Square Tiles. Sorry, Darren.
Square Tiles. Sorry, Darren.


Thanks for reading,
Mister Bums

You can contact me at [email protected]Twitter, or leave a comment below!

Visual Clutter


Out with the Old…

Steam Marines 1 had some issues with visual clutter. Lots of tiles, lots of colors, sprites overlapping sprites – it was a bit of a mess, in retrospect.

Steam Marines 1 - lots of sprites on other sprites.
Steam Marines 1 – lots of sprites on other sprites.


More than one player complained about the doors. Visually speaking to a lot of players it looked as if the marines were actually one tile higher than they actually were from the doorway. The red-cloaked marine in the upper left of the screen? He can step onto the tile directly left of him – that’s an open doorway tile.


Beyond that I had flashlight and laser sprites with transparency, fog of war tiles could slowly fade away to reveal their tiles, marines had cones of vision that depended on their facing direction, and items could be on chests and there was animated flashing explosives on doors and and and… it was a lot. Too much, probably.

I’ve tried to rectify that in Steam Marines 2.


… In with the New

The one thing I did like a lot about Steam Marines 1 (SM1) was the black around the playing field. It made for some poor screenshots, unfortunately, but it helped to focus the player on the actual game. This is what Steam Marines 2 (SM2) looks like in the tactical field:

Steam Marines 2 - Tactical Field
Steam Marines 2 – Tactical Field


As you can see I’ve retained the black edges, all around the tactical field and even on top of environment walls. Differences include:

  • Fog of war is binary now. In SM1 fog of war had three levels: 1) unexplored, 2) currently seen, 3) seen before, but not currently. Having only 1) currently seen and 2) not currently seen has a lot of benefits. For one it’s easier to render! Another is it keeps all that lovely negative space around the player. And a third is it keeps a sense of mystery and oppression the entire time because you’ll never have enough marines to see the entire map at any given time.
  • Almost no visual clutter. There are a few item pickups that spawn, and there are details and some grass in the Alien Temple level, but it’s not visually cluttered like in SM1.
  • The doors are a lot more clear now. Unlike normal environment walls doors are not black topped and the level lighting slices into the top of the doors, in this level giving the doors a hot purple look. They are also taller than the surrounding environments.
  • The UI is overall a lot less intrusive. There’s the portrait, name, and stats of the currently selected marine in the upper left, the squad ammo bar on the bottom center, and an action panel in the bottom right corner. There are some contextual UI elements such as the green cog under the currently selected marine, and the blue tile overlays showing where that marine can move to. There are also a few text boxes and such that pop up during attacks and other such, but by and large it is much less cluttered.

Part of that is because I shoved some stuff into a submenu. Another is that I removed squad inventory. Squad inventory, as I’ve noted before, is not a great addition in permadeath games. By tradition most roguelikes would beg to differ on that point, but I’ll maintain my position: Players frequently try to save items because the nature of the game is to place you into holes you have to dig yourself out of. Then, they forget they have them because they spent hours refusing to use them unless they felt the situation was dire enough. In other words, their behavior slowly gets molded from “Okay, don’t use powerful, rare items until I REALLY need to!” to “Don’t use items!” to “What are items?”

Aside: Have you ever ended a (non-permadeath) RPG and noticed you have an inventory full of super-powerful items you saved up and never, ever used? Did you die at all on the playthrough? Yeah, that.

SM2’s fog of war is also soft and blurry at the edges, not hard and discrete like in SM1. Mechanically it’s the same, though marine visual and weapon ranges are by discrete tile and not partial tiles.

I think overall this is the correct direction to go in, although I do have some misgivings about the marine action camera when aiming at enemy targets:

Targeting System.
Targeting System.


That is A LOT of black on the upper half of the screen. I’ve been thinking of putting something there, although it begs the question: why is there no roof in any of these interior environments?!


Thanks for reading,
Mister Bums

You can contact me at [email protected]Twitter, or leave a comment below!