A small update coming down the pike today, but one big one comes along for the ride. Objective pointers! When a boss appears, the game will help guide you to the opponent, and once the boss is destroyed, you will be pointed to the exit. It’s a frequently-requested feature, and it will only get more robust as time goes on.
Objective pointer appears when boss appears and then the exit gets a pointer when the boss is dead.
Game doesn’t slow down quite as long when an enemy dies, and it doesn’t slow down at all if the victim is offscreen.
Rebuild some shader files to reduce hitches.
Visual quality defaults to Very High rather than Very Low(!). (Auto Fire is 3D but it’s not that taxing on your machine…)
Fixed the spacing for the inventory/buy/sell screens a little, and a few other quirks.
The loadout and settings screens had some fixes to make them more usable with a keyboard.
Fixed some things interfering with the cargo screen when used with a mouse.
Fixed some bugs when using the repair screen.
Removed a rare combat result in the loot crate encounter for the time being. (Caused you to get stuck in the encounter popup)
It’s arrived… a big update, worthy of a full point! Auto Fire has had a lot of work put into it over the last month, not the least of it being the revised onboarding (that’s fancy-pants dev talk for what a new player experiences). A lot of players’ first experience with Auto Fire was tough, because there were some really dangerous enemies out there in the badlands, and you were just as likely to meet them while you were still trying out your training wheels as any other time.
In the spirit of making the game more playable and easier to understand, there’s now an intro that eases you into things a bit… and sets up the conflict against your (generated) rival down the road! Clear out those bandits and get back to civilization!
The game also tailors that early experience when you’re stuck with a couple of junkthrowers and not a lot else, so that you are less likely to encounter enemies that want to come punch you in the fun bits right away. This way you can pick up some loot, head to the garage, and upgrade away!
In addition, the player can now customize their character more fully upon startup. You can choose your portrait, and the dialogue (even the procedurally-generated stuff) will adjust for male, female, etc.
There are a ton of changes that came in, but a lot of it is nerdy stuff that only I care about… Oh, hey, here’s some cool stuff you might actually notice!
There’s a new encounter system that extends the random encounters in the game previously. You now can see your opponent, they can call you by name, and choose a course of action.
The game starts with some intro encounters that help give you a sense of what you’re doing there and what to do next.
Character customization has been extended to support player portraits as well as choosing gender.
Both enemies, players and any NPC can now be male or female. Dialogue is set to use the correct language to address them as it comes up.
You now have a (soon to be) publicity agent and a rival that are generated on campaign start. These are referenced and used throughout the game.
Agents are now handled outside of factions. Thus agents (any NPC) can join you, an opposing faction, or be an independent operator like the player.
A story manager presents basic encounters that drive the initial experience forward. This will hook in as more cause-effect things happen with the player.
Naming has been adjusted to better fit categories of civilians, punks, drivers and corporate suits.
Female names have also been added since we can generate those NPC’s now.
Chests now break open when moved onto in the overworld.
Weapons now start with a higher amount of damage (junkthrower starting around 10) but don’t increase as drastically (the maximum is about 250% of the minimum, rather than the former 1000%)
Weapons have less of a randomized spread in their damages (used to be 70-130%, now is 80-120%)
Armor for the player’s vehicles have correspondingly been increased, from 100 per side to 150.
Armor repair kits now repair 15 per side rather than 10.
When the player is ambushed into a battleground, they start in the middle of the map rather than near the entrance.
There’s a new intro cutscene that sets the situation when the campaign is started.
Because I know dying would be a chore if it couldn’t be, these can be skipped.
Encounter text also has a teletype-style reveal, which again can be skipped if the player is impatient.
Portraits can show or not show in encounters, and can even show as a transmission.
The loading screen shows hints and has a spinner so you know if the game is live.
The cursor keys have been remapped to work properly within menus. Key navigation of menus still needs improvement, but it should be better than it was.
Fame/Prestige is displayed properly on character and city menus.
Adjusted a bunch of buttons to be more visually interesting.
Cars now don’t kick up much dust when travelling on roads.
The gas can in the world is now red, matching the icon(!)
There is a quick stint of slow-motion when cars are destroyed, and the same effect is back when the player died.
Revised some of the stingers when entering maps.
The turn-end sound was revised to be less in-yo-face.
Button presses have more audio feedback in more cases.
This coming month will be all about adding variety and continuing to improve the interface. If I demonstrate the game at a show, what will draw players in? What will it take for them to understand what’s going on? What last flourishes or abilities would really sell the whole package?
This has been an exciting week for me… As Hidden Path puts the final digital shrinkwrap on my most recent VR title Raccoon Lagoon, I’m bidding the classic nine-to-five a temporary adieu. As of yesterday, I’ve started focusing my full attention on Auto Fire! It is my hope to get it into solid, pro-tier shape over the summer, and see what happens!
In celebration of this moment, I’ve put together an early trailer, complete with some pro-tier voice acting:
Auto Fire needs a lot of work yet. I need to make the interface more approachable, flesh out the content, and improve the basic art so that the game starts turning heads. I need to deliver on the fantasy of driving a combat car… that’s skidding, shooting, hauling convoys, maybe even launching from jumps? The possibilities are endless!
Adding gamepad support should help people get comfortable with it quickly, and improving the mouse interface will hopefully do the same. Anything that helps people ramp up and be gripped by the promise of a muscle car bristling with chromed-out weapons.
As an inaugural step for my all-in on Auto Fire, I’ve been deeply examining my turn model… something I haven’t touched in like two years. The way it used to work is that each team would execute their moves when the timeline reached it, so that if you were moving at 60 mph (3 moves per turn), and the enemy was moving at 40 mph (2 moves per turn), the simulation would resolve with:
You move (progressing to 0.333 seconds)
They move (progressing to 0.5 seconds)
You move (progressing to 0.666 seconds)
They move (progressing to 1.0 seconds)
You move (progressing to 1.0 seconds)
Guh. This might sound sensible if you are a realism fan, but when many entities are moving at different speeds, the turns all interleaved and the player never knew who was going to move when. It was confusing and could get frustrating as an enemy vehicle suddenly drove right into your path or out of your line of fire.
The new model is pretty simple: Each turn you execute your moves, then the rest of the world executes their moves. So in the above example, you’d get your 3 moves, then they would get their two. It’s basically X-com style, but you know what?
It ended up playing exactly the same.
Yeah, you can’t really tell there’s any difference at all, it flows great. And while I was worried that you’d be irritated by that long pause for your opponents to do their thing during your fancy driving , it actually feels a lot better than when your opponents interrupted you at odd, unpredictable times during your turn.
This is an important breakthrough because if I can make the player keenly aware of what a “turn” is, I can help them understand what acceleration does (more moves per turn) and how weapon cooldown works (most weapons can only be fired once per turn). I might not need a hojillion progress bars (a weapon is either available or not). Since the core goal is to make the game more accessible and less math-y, I’m optimistic that this is a good step that doesn’t sacrifice the core gameplay.
Finally, along with this exploration I started experimenting with better shaders using Amplify for Unity. It’s another useful step, because there will be a lot I want to convey in-world and good shader control will help me make better 3D and mouse-driven interfaces. Things are looking up!
I’ve been laying the foundation for a trailer over the last couple weeks, so most of my progress in Auto Fire has come in the form of cleanup, although there are a few gameplay tweaks, mainly to make that early experience a bit better. On the upside, there are less-soldier-y gang members, a cleaned up HUD, some new lighting and vehicle trails, check it out.
Starting next week, I’ll be able to work on Auto Fire full-time for a while, so prepare for some meaty updates. The goals are to significantly improve the player movement, reveal more options when driving, and fleshing out systems that have only been teased, like a more alive overworld.
See you on the road!
Upgraded to Unity 2019.1, which seems to have gained a little performance.
Cleaned up UI
The equipment quotes in the info popup are now aligned correctly
Revised weapon icons
Revised window frames, etc.
Removed the skid meter behind the car, since it didn’t look great.
The player can turn the skid meter on or off in the options screen
Remove random encounter dialogues until they are ready.
Revised the road decals to make for softer edges
“Soldiers” are now desert warriors
Foot gangers now use the proper pose to match the weapon they are wielding.
Added headlights to the player car when in a ruined city
Lowered overall brightness of the city map, so that the light sources could stand out.
Increased the light brightness for streetlights and barrels and fire.
Revised the trails from the taillights of vehicles entirely. They use a new system that makes them very smooth.
Improved the detection of offroad when a vehicle is driving, for visuals and gameplay ramifications
Gangers now use battle cries rather than military radio transmissions when they spot the player
Vehicles now rev their engine instead of honk when the spot the player
Improved the visibility of enemy cars (they were very short)
Increased the range of junkthrowers since they were still kind of frustrating to use.
Adjusted population cap to accommodate new fuel dumps and watchtowers.
Oh man oh man it’s been far too long since I’ve published an update to Auto Fire. That’s a terrible thing I don’t want to happen very often, but I started to put in the quest updates and it made sense to get a number of additional features up to snuff in support of it.
Worse yet, I sat on a hojillion changes in my source control before I checked everything in. I think it was like a month. Work was making me a little crazy, but that’s ridiculously bad form. On the upside, this update brings about a bunch of changes in a big sweep.
My primary goal for this update was to add more exploration and stages to the boss fights and quests. For this I needed to support better quest state reporting, and make new emplacements to fight against to draw the boss out. Then I realized that the whole system fell apart when you left the area, so I had to improve how quests were maintained when you leave an area. Then I realized I wasn’t really saving data the way I should and basically had to improve the saves to be near-ready for cross-session saves (hopefully soon). Then I realized that spawning emplacements in random locations was really ugly and made them hard to find, so I added a content socketing system for bosses, emplacements, loot and hazards so that their placement could be more deliberate and hand-crafted.
Along the way I cleaned up the UI, added dynamic music, fixed some lingering physics problems (which caused invisible soldiers when the ragdolled out of the world, as well as making some tiles near rotated large objects to be un-enterable. I even stripped out some of the anti-aliasing that was making the game look muddy.
Entering a map occupied by a boss now requires the player to progress through the map and take out a number of strategic structures in order to coax the boss to face you.
Outpost maps are defended by armored watchtowers.
Ruined cities require you to take out fuel dumps.
The quest title is shown when entering and area, and updates are shown as the player achieves objectives.
The mini quest display is cleaned up and should update properly.
Quests are properly resumed when the player returns to a location.
Reviewing your quests that are in maps other than the current one is handled better.
Added some post-apocalyptic music and a couple stingers. Adjusted existing stingers.
Added dynamic music tracks for city and outpost tactical maps.
Dynamic music now escalates as the player takes out more emplacements and the enemy spawns get more intense, up until the boss is unleashed and the boss music is played.
Added boss-specific music, and adjust the intensity based on how close the boss is.
Adjusted the anti-aliasing so the game isn’t blurry. Temporal anti-aliasing can cause a smearing effect might work for realistic titles but ain’t great for games with precise information to dole out.
Added sustained fire bonuses that improve player accuracy after multiple attacks.
Painting an enemy with the radar will improve player accuracy against them.
Improved some targeting response elements by indicating which entities are people, cars, emplacements, etc.
Emplacements such as watchtowers have new aggro and play distinct spotted sounds.
Extended the aggro duration of enemies and made sure they don’t lose interest in the player while still in sight.
Junkthrowers do 50% more damage. They were supposed to be scrub-tier weapons but they were just sooooo bad.
The Stallion now has a bolt rifle mounted front and two junkthrowers (one per side). Its combat capability was depressingly terrible.
Mines have a lower cooldown again.
Significantly more cash is dropped from loot crates and enemies. Killing a boss and getting $4 was definitely sub-awesome.
Zones have fewer garages.
Quest emplacements like watchtowers and fuel dumps are placed in sockets that are part of map generation. Thus their placement is more crafted.
Loot crates and barrels also have specific hand-crafted sockets for various map generation tiles, for a less haphazard placement.
Crates are off the beaten path, sometimes in nooks or dead-ends, but generally in a place somewhat thought out.
Barrels are placed in clusters around road hazards, fuel stations and large wrecks.
Loot and barrels now have a tunable target number placed per map. Before it was a much wilder range of possibilities.
Population, quest progression and entity placement is now saved when exiting and returning to a map.
Entities, enemies, sites and pickups now save their state (when marked to do so) when leaving and returning.
This is not quite all the way to full savegames, but we’re very close.
Improved the display of enemy misses somewhat. Shots go wide and misses are pretty clear.
Cleaned up the “chrome” UI window borders. They were originally photoshopped from actual chrome dashboards but that didn’t scale as well as I’d like. Buttons have their own appearance now.
Improved some bugs with weapon targeting and the widgets over target vehicles.
Can now display entities as singular or plural for quest readouts.
Boss and targeting popup displays are now cleaner and, well, less terrible.
Fixed the handling of rotating large objects… This means that there should no longer be any invisible barriers.
Improved some poorly-handled persistent effects such as oil jets and skids… These are now handled with greater safety and more robustness.
Enemies no longer can get in a state of attacking inanimate objects or themselves.
There has been a solid amount of progress on Auto Fire in the last month, though not everything has been visible.
There’s a condition system now, where entities can be stunned, set on fire, made to skid, be blind, etc, and that will last a fixed number of turns before automatically removing themselves. Nothing super fancy, but it allowed me to do stuff like cause a vehicle to spin out when it hits an oil slick.
It also allowed me to give the player’s radar more functionality, because it now “paints” targets within a specific radius for a set amount of time. Ideally the player should be able to build up sustained fire on a single opponent, or race through a group (at high speed to avoid being shot) and hit everyone with a radar ping before swinging around and taking advantage of the higher hit rate (and eventually critical hits spurred by this).
There is a new icon system above vehicles to show their current conditions, which hopefully will teach players more about the advantages of speed and choosing targets.
I did a bit of work on weapon resolution to clean up some weirdness, as well as allow for effective area effects over various volumes. I can have weapons with blast radius at impact, cone effects, lines, and more. This gave me some vastly improved versions of scatterguns, flamers, and so on.
I also switched over my missiles to LeanTween (a great Unity package that’s freeeee, although the Editor that goes with is worth throwing a few bucks at) so that I could use more sophisticated arcs (splines, eases, etc) for the projectile travel. This gave me some great drunk missiles and so on.
A somewhat smaller bit of work but vastly important was looking into problems I was starting to see in my city layout.
A couple of years ago I put multiple months into a city generation method that took pre-crafted blocks and spliced them together, street-to-street, with props and so on. It worked pretty well… However, lately the cities seemed to have wayyyy too many skinny alleyways and dead-ends, even though I remember putting a fair amount of effort into reducing these.
Worse yet, I’d started to see some passability issues and unplayable maps, which I know I did checks for. Ugh. I love dusty Mad Max wastes, but the cities are just as important a part of the game and they weren’t fun.
I spent some time trying to re-learn what the hell the 2016 version of me had made. For a little bit I thought 2016 me was a bit of an idiot… but it turns out he was somewhat clever. It was 2018 me who introduced a number of bugs that caused loops to no longer form… that guy was a jerk. Specifically I had some code that overlaid roads over previously populated obstructions to create extra loops, and those no longer overlaid properly. In addition, my passability checks were not properly busting holes through the buildings and obstructions when needed.
I added a bit more two-lane roads and discouraged alleys from forming very often. In addition, I added some new block types to my definition that had fewer buildings, so some extra open spaces could be formed. I can pretty much make an infinite number of city block components, so I’ll keep adding ones that give some more driving freedom.
Anyway, I hope to have a new version out this weekend, it’s been too long. Wish me luck!
Been working on creating some better combat mechanics and options for Auto Fire lately. One of the factors I was working on was tangibly having hit chance be affected by tactical choices such as “painting” a target with radar for more accuracy, or on the other side moving faster to avoid getting hit. I got those systems in but it didn’t really seem to matter.
I had to seriously think of how I’m dealing with accuracy, range and speed. My initial version of the combat system in Auto Fire was inspired by the system in Diablo II. It used this sort of calculation:
Chance =200% * AttackRating /
(AttackRating + DefenseRating)
Consider AttackRating to include things like gunnery skill, weapon accuracy, and other targeting bonuses. DefenseRating in contrast is stuff like range, target speed, and target size.
Multiplying the value by 200% gives the player 100% hit chance whenever the Attack Rating is more than 50% the Defense Rating… So in the average case the player is guaranteed a hit. This was the system I used in X-Men Legends, and it feels pretty good in an action game where you are directly controlling your character and only want to have misses when your opponent clearly outclasses you.
The problem with this method is that the chance to hit tapers off because it never reaches zero… Thus, I was finding that there’s only a small sweet spot where hit bonuses mattered between 100% chance and maybe 60% chance.
I do put a minimum hit chance of 10% just so it’s not worthless to attack, and I’m not sure what increasing the miss chance is going to do to the feel of the game.
I have a mechanic where excess above 100% is used as a critical chance (That is, 130% chance is an automatic hit with a 30% crit chance). I might consider doing the same on the bttom so a “miss” is a glancing blow that still does a little damage, progressing to a real miss based on the percentage. (For example, -30% is a 30% chance of a real miss and a 70% chance of a glancing blow).
I’m wrapping that back into weaponry definitions and looping back around to where I can try the new system in the wild. From there we’ll see what happens.