Latest Trailer

Throw a wrench, change the galaxy.


Mostly nebulae.

Release date

No idea - sign up to be told when it's out


Windows for sure, dunno about others


Get on the mailing list, I'll e-mail it when there's a way to do that


Design, Code, Writing

Tom Francis


John Roberts


John Halpart


Tom Francis
John Roberts

Adaptive Images by Matt Wilcox

Mailing List

We only mail you to invite you to closed alphas or tell you we released a game


Heat Signature’s First Four Objective Systems And Why They All Sucked

I’ve been designing and trying various ways for you to make progress towards your objective in Heat Signature, and four bad iterations have led me to a surprising conclusion.


Daniel F: Thanks for sharing, Tom - it's really interesting hearing your thought processes.

For my part, I enjoy it when games give me the opportunity to put those McGuffin pieces together in a way that makes me feel smart, even if it's an illusory feeling. For instance, in the Blackwell games, you run around essentially collecting discussion topics for your notebook in order to unlock solutions, but the game lets/makes you put the topics together (a detective version of the age-old adventure game trope of using everything on everything else) to "deduct" the solution. It made me feel smart and responsible for the solution, even if I knew upon reflection that it was an illusory feeling.

In Heat Signature's case, have you thought about letting the player piece together the individual heat sig traces of the target ship in some sort of puzzle mini-game, or deduct something about the target ship from the hints you've collected? Or if not, what do you think would be wrong about that sort of mechanic?

Damon: Hey tom!

While you were discussing some of those objectives I had a thought. You've previously mentioned that ships in the world will have their own roles and objectives; so What about having the Intel you receive give information about the target's role in the world?

The target ship could be a courier, in which case the snippets of information you receive might be about its usual flight path, or its destination. That info might reference places that you wouldn't know about and require more intel to locate.

Once you've peiced together the location, going there and locating the ship on the job is the goal. That might be tricky if there's stations or pirates or it's in a disputed sector. But you know where it is, and now you just have to pull the heist off in one piece.

I'd play a game like that (:

Fin C: Hi! Great to see the thought process behind combining the objective with general gameplay. I imagined that the objective ship would have advanced defences or something else which means throughout the game you would need to learn more than just its location - like if the main quest in Skyrim required you to be a level 90 in lock picking or something.

Also, is that a serving hatch?

Alex W: It may just be Cosmos talking, but I'd love it if the heat signature you were detecting was the objective ship's blackbody radiation graph reading, finding out the right spikes/missing pieces to indicate what's in the ship's exhaust and combining that with its shape and allegiance to track it down. I'm not sure how easy it would be to generate a convincingly disparate set of possibilities from that in Game Maker, though.

ChuckzM: A trajectory calculator might be fun. That way you could intercept and not chase in some cases.

SmashGuts: Do.. Do you have two watches on? It's that what's on your wrist?

Causeless: I think it'd be interesting if instead of getting hard data on the location, you got "soft" data to analyse (which is non-decaying). If you just check a magic bar that fills up and tells you the ship's position once you randomly loot enough ships, it takes away the most fundamental concept in the game, which is the mystery of finding the ship, at least from my outsiders (admittedly un-experienced) view.

I think it'd be nice if instead you got information like, "the ship is an anti-piracy ship that patrols Sector X5G", and then you could learn more information from there - like, the shifts the ships switch around at, the approximate route it takes, and perhaps even information about the ship's weaknesses ("The ship's right missile launcher is malfunctioning"...) to help board it and take it once actually there.

You could also find completely irrelevant information to serve as humour, as I think we'd all love to see more of the genius writing that was in Gunpoint.

It'd be pretty difficult to put that into a procedural generation routine, and probably a pretty big task to prototype, but I think it could be done as the information snippets are partially procedurally generated:

"The target ship's parent space station is in Sector " + string(targetShip.parentLocation) + ".", "The target ship is " + targetShip.activity() + "."

Where targetShip.parentLocation might be a sector number, targetShip.activity() is a function returning a string like "patrolling", "defending", "idle", and so on.

Hope my ideas had some truth behind them, haha! The biggest immediate issue to me is the complexity, making it difficult to quickly prototype... Thanks for reading.

Driftwood: Tom,

I think Damon & Causeless are on the right track. It seems that the core of what you find dissatisfying about your objective systems is the behaviors of the ships. As far as I can tell (and I may be wrong here) it looks like all the other ships just have a direction and a speed and they just fly in a straight line until interrupted. This means you have, at any given moment, approximately a 50% chance of having an objective ship that is speeding away from you into the infinite black. (Well it starts out as 50%, but if space is infinitely large then everything ends up moving away from the center and that percentage drops. If space is limited then eventually everyone ends up stuck on an edge and the % drops off again. If space is toroidal [i.e. going off the top puts you at the bottom and going off the left edge puts you on the right edge] then everyone is both moving towards and away from you. Math is fun!) Chasing after a ship that is just blindly rocketing away from you seems kind of arduous, especially since getting new info about your target requires you to stop and board another ship which is potentially moving in the opposite direction.

That goes away if other ships have routines or learnable behaviors, such as trade routes or patrolling certain sectors. It also solves the time sensitivity problem if what you're learning from other ships is pieces of your objective's routine. The player now has 1) a puzzle to solve by piecing together the parts of the targets routine and b) more choice in when and how they want to take on their target. If all you know is that the target visits station X then you have a choice of either continuing to gather more info or laying in wait at station X.

So yeah... it may be worth while to shelf objectives for a bit and build up the world simulation some more before you take another stab at it.

Keep up the good work :)

Abe: After hearing that you played Covert Action, I was hoping you might implement a system similar to that: where the player gathers seemingly random and unrelated data, but as he/she continues the data begin to create a cohesive picture of the world and its various players. For example, the player could download another ship's computer banks and find a variety of minor pieces of information on the various ships in the galaxy. As the player continues, he/she creates complete, comprehensive profiles of ships and factions. This would eventually let the player find his/her target by knowing what faction the ship belongs to, where it patrols, and what stations it visits.

Furthermore, the player gathers information, and therefore mastery, over other ships that may not be his/her target. This would help make the goal of finding the target feel more like an optional goal the player can complete at his/her leisure, rather than a daunting, ever-present expectation.

As far as implementation, you don't need to rely on the player ducking into an encyclopedia-like menu. You can let the player click on a ship and display all the information he/she might know about that ship displayed on screen. (crew size, loot, name, captain's name, faction, etc.) There could also be a hot key to display all known information on the target.

Good luck with the game regardless of what you decide, I know I'm looking forward to it excitedly.

RealoFoxtrot: Hey Tom, Everyone here has given their own advice... and i shall too :P

Recently, i've started to rant over story in video games, trying to get under what makes them enjoyable.

I've come up with two basic things to focus on:


Characters can be anything, from a location, to a vehical, to a person. So long as they have character and are memorable and suchlike, they are a character.

Events are memorable things that happen. Anything from a scripted event to you farming a tree that uncovers a massive cave system.

Just some ideas ^^'

CdrJameson: Just scattering the clues around, but make them different not just fragments of the heat signature. Might as well say 'collect four magic frimbees' in that case. Make them a collection of different aspects of the target (ship type, signature, equipment, habits etc.)

Even better allow the player to apply some brains to increase their chances of finding certain types of information. For example, classes of ships collect different types of sensor information. Patrol ships cover certain areas. Big traders will know about main routes, independent traders will cover smaller worlds and so on.

Allowing players to accidentally attack the wrong ship through mis-identification would be a cool moment. Whoops! Sorry guys, my bad! I'll pay for the damage! Throw in some clues that are twists on existing information for that one, eg. In Raiders of the Lost Ark where they have to take some length off the staff height. Perhaps a piece of information that they've recently changed ships (but raiding the old one would get you a clue)

mesotes: hi,

while listening to your vid, some ideas poped into my head: its all about the location of the ship, but what about the ship itself, what it is capable of..

about the time pressure thing:
get to some other ships and they might have seen your target but 5 minutes ago, that might not help you much location-wise, but it might help you to know what was the ship doing there? or if you are lucky you get to know some specific details on that ship to help you foresee its movement/behaviour

one time it might be a refuel station which knows how big the fueltank of the ship is..

another time a military ship which will have info on how fast the ship can go for an amount of time..

or maybe a civilian ship has seen the target with a broken hull, so you can deduct that it might get some repairs soon.

all this information together enables you to foresee your targets movement further and further.. until you have enough to pin point it.

too much brainwork too?

0seven3: What if the target ship were a sort of powerful boss ship, maybe the flagship of a villainous fleet? The player could try to spend the first half of the game exploring, raiding smaller ships, and avoiding the boss fleet while gathering resources for the final battle. Eventually, the player would have enough resources/upgrades/skill to infiltrate the boss fleet and destroy or capture the flagship. It might be fun to slowly build up the resources to take down a force that you had to flee from earlier in the game.

BlubBlub: I'm surprised that you didn't come close to a system I had an idea for while watching the video.

I ask a simple question where are all the ships going?

This lead me to think up a system similar to the lead system you talk about that gives you the destination of a ship. Depending on how good you Intel is it will either be a exact location or a type of location in a sector. You will then at some point go to a location that the objective ship is at.

Types of locations could include Planets, Asteroids, Other ships, Space Stations all of which move around.

You could also use the same system on each ship or just the enemy ship to give it an objective too. This could make for some interesting game-play if you could be a npc's objective. There are all types of variations I can see for this type of system.

Greytower: Hey Tom!

I actually sort of liked the idea you had where you are given clues as to where your target is, and it gave me an idea for maybe a better system for objectives.

What if the objective of the game was to board ships and hack their data in order to discover the location of your target ship that you need to get to for whatever reason?

Shaun Cheah: This also exists as a YouTube comment but it was so long that I decided to put it here, too:

I was thinking about how target pursuit worked in the Sid Meier's Pirates! games. You can gain information on the ship you're pursuing and the last piece of information you get is the ship's destination.

Perhaps your infinite universe could have "destinations"? Each faction could have a flagship whose heat signature (for your long range scanners) is really easy to obtain (maybe you get 25% of it each time you hack a ship of that colour, or you agree to do a quest for that faction and they'll give your their flagship's heat sig so you can turn in the quest there once completed). The flagships patrol certain areas or they move in a circle or they move really slow so that their general location is always pretty static. Meanwhile you also have trader stations and such which will happily broadcast their locations to receive attention.

Then it's just a matter of having it so when you finally hack a ship with information on your target, instead of getting 25% of the target's heat sig (which would kinda make the game always boil down to hack all the ships until you get the four special ships with information, then fly to your target and win) they would give you his last known destination (your target was flying towards the Red flagship, or Tom's Space Bar, or Itrade's Escape Pod Store) and then once you get there the ships around will have a better chance of knowing the next destination and you'll also have a chance to pay for or do a quest for information on why your target went to that destination (Your target is trying to mediate a peace between the Reds and Blues, your target is romantically attached to a waitress, your target was browsing the latest escape pod models but left unsatisfied). That way you have all these sidequests that could merely add a kind of randomly generated backstory for this target or they could even let you know more about his patterns (If he's trying to negotiate peace between the Reds and the Blues, it may be a good idea to go to Blue space and try to ambush him there).

Maybe you could even tie the destinations thing with the 25% thing. Each destination could upload a partial heat sig to your computer for an even higher payment or an even tougher quest, for example.

Anyway whatever happens I am super excited about this game.

ERGlabs: It looks like I'm very late to the party here but I'm compelled to leave my 2 cents. I'm sorry, and I regret this already.

I enjoy the concept of building a sensor profile, mainly because the 1/4 signature mechanic means you have little info in the way of making informed decisions as to who to board next.

Weapon Detonation Signature - probably on a ship who is hostile to the one you are looking for. They might have a good scan of the weapon profile of you target ship. then your sensors can pick up Weapon Signature in a direction. Just an occasional ping should your target ship run into combat.

Communication Frequency - possibly on friendly faction ships or known trader vessels. Again, not precise location, but a directional ping when the target ship comes into range of another ship it wants to communicate with.

Engine Exhaust Profile - Maybe on a engineering or science vessel. This would be the best for close range tracking... once you know it your sensors could pick up a colored trail left by your target's engines. But you would have to be close because the trail decays fairly quickly. or you could get lucky and stumble upon it.

The cool thing with the sensor profile is that it allows some logic to which vessels might have useful information. Get intel for the current political landscape, to make informed choices as to who might have Weapon Signature info, etc...

It would also allow interesting upgrade pathways for finding neat things on other ships. Better Com software might narrow the ping from a 90 degree arc to a 45 degree arc, or ion detection hardware would lengthen the trail for the Drive Signature. Or both those upgrades might be useless if you haven't found the appropriate sensor profile yet.

Vagelis Bebelis: Interesting analysis on your thinking Tom, thank you for making it available.

Watching the video with the automatic caption of youtube on is hilarious. You need to give it a try!

Kirk: Hey Tom!

Wanting to chime in on everyone else's comments regarding ship behaviour and traits - what seems to be getting people excited is the possibility of doing some rather genuine puzzle-solving and guesswork that relies upon building speculative models about how the ship behaves so that you can track it down. You get to feel a bit like a space detective in paying attention to the information the game is giving you and formulating hypotheses which you can test out, without the game ever trying to lead you by the nose. This sort of behavioural anticipation is something that people are really, really good at, and is very rewarding to do. We do it every day.

So, rather than identifying a ship by its technical attributes, which would require a lot of memorization or UI work, things like "this ship has a tendency to attack in X manner" (which could lead to a thrilling sequence of following the blown-out corpses of ships it has predated upon through the galaxy), and "this ship is a gun-runner" (check war-zones with resistance groups).

Just to think through the mechanics of the solution, you could have a randomly generated web of identifiable behaviours for your target produced at the beginning of the game. Throughout the game, any random ship you board could have a percentage chance of just so happening to have heard of your target (percentage perhaps modulated by a number of other factors, like faction affiliation, star system of origin, etc), which would yield broad information about it, allowing your search field to narrow slightly. Any ship your target has actually been in contact with would have much more specific information.

The interesting thing about this solution is you could define "factions" entirely within the fiction and within certain behavioural profiles. Without needing to come up with a complicated faction system, all I'd need to know is that members of X faction behave in a certain manner and leave certain traces, and I could start my detective work from there.

The elegant aspect of this on the player's end is you could seed this information in with random information about the world itself, seamlessly merging your world-building with the overall overall goal. In amongst random intergalactic news articles, confidential emails, star-charts, and staff logs, there might just be the clue you need to narrow your search. And in the process, you gain other valuable information about the world to leverage to your advantage. Maybe the clue you just got doesn't mean anything now, but the next ship might just have information about where exactly the Delta X4 system is, and who the Qualians are (and more importantly, what characteristic things they do).

Mike: Hey Tom!
I think a multiplayer mode (if possible) would fit very well with the game mechanics.

Migael Ferreira: Hi Tom. I know this is very late, but I just had to share a few ideas on your objective systems.

First, I noticed that in all of your systems, you are assigned a distinct objective at the start of each game - eliminate a ship, assassinate a guy on some ship, steal a thing, etc. I realize that in infinite space and with so few systems in place an assigned objective is necessary for some focus, but still I thought, "What if I didn't have to be just the hunter all the time, but the hunted?" In other words, what if you start out the game with no set objectives (the context being that you're some vague type of bounty hunter), start boarding ships, and then MAKE enemies, instead of being ASSIGNED enemies?

So if by chance you board and immobilize, say, a ship from the green faction, who turns out to be a *valuable* ship for Fleet X, sooner or later the green faction will investigate, pick up some clues and put a price on your head. Of course you'll need a system for the player to know who is out to get him. Perhaps you could have an AI on your ship that data mines faction communications (not really, of course) and gives you info on what fleet is hunting you - who you have to surprise before they surprise you. From then on the objective system you discussed here could be implemented to find whoever has to be silenced. The AI could concatenate the clues/ puzzle pieces into usable objectives - suggesting courses of action based on the intel you've found so far. This eliminates the need for the player to remember it all and to build the puzzle. The intel structure would have to be generated as soon as you "officially" receive the mission - you'd have to have a finite number of defined clues spread across random ships, with a limited set of definite deductions from that. For example, if you found two clues, one saying that the objective ship is damaged, and one saying that it was seen headed towards a certain area of space, the AI could give you a list of repair stations to investigate in that area.

Before you receive your first "Destroy this fleet before it destroys you" mission, you're just wandering around, looking for treasure and trouble. I realize that this could quickly lead to you having several fleets hunting for you. To counteract this, make only one "sensitive" ship available at all times (so that only one "main mission" is available at any time). Getting the player to discover this sensitive ship in infinite space would be tricky, though.

Of course objectives don't have to be limited to "Neutralize this ship/person". You could pick up rare clues of treasures/upgrades hidden on a cargo ship, then discover that it's being guarded by a tight fleet that has to be systematically infiltrated. I feel if you could somehow get schematics for a fleet's layout (in other words, in Captain Bailey's fleet, ship X is shaped like this and it has powerful cannons; ship Y has authority to order ships in the fleet to move; ship Z is a weak link in terms of heat sensors; ship A carries high explosives; ship B has access codes for ship C's airlocks, etc) it would add a whole new level of strategy and planning. The need to plan would make infiltrating a fleet to get at the main ship very satisfying. So you have infiltration strategy at ship level, and at fleet level.

Some other loose thoughts: I feel that some fixed locations are necessary. I mean, where are all those ships flying to? On the other hand, the planets/ space stations/ whatever would have to be quite far apart to make it possible to board them before they reach their destination. A basic diagram of the locations of these...locations would also be needed.

A couple of ship upgrade ideas:
- If you dock a ship while hot, you said there's a chance that a different, hypersensitive ship will pass, detect you and open fire. You could have a device that immediately transfers heat to the "host" ship as soon as you dock, preventing this problem.

- Flares to distract a ship's heat sensors.

Sorry for the text wall. :|

I hope some of these thoughts can contribute to usable systems in the game. I can't wait to see where you go with it. Good luck!

postinternetsyndrome: Already loads of ideas here, but I'll add my own.

The thing about "ship from faction x that last saw the target" made me think of a setup where you could unlock a permanent marker that showed the ship of a specific faction that is _currently_ closest to your target of all ships in that faction. This could be obtained by doing lots of missions for that faction, by breaking into their headquarters base/flagship/whatever. Repeat this for several factions and you would get a marker cloud - constantly in flux - that showed you the general area where the target currently is.

That idea may or may not be a good one, but another thought I had on reading all the comments here is that maybe it would be interesting to have several optional paths, and/or randomize the available information along with the rest of the world. So one playthrough might offer you to buy a tracker that some pirate installed on your targets ship, for a ridiculous sum. To get it you need loads of money, so go do missions and explore the world and eventually you will be able to get to it. Another time you could know the position of the ship permanently, from the start, but it's a heavily armed mothership thing so the challenge is not to find it, but to get upgrades and intel on it weak points. And so on.

Regardless of what path you may choose in the end, I'm looking forward to the game! It looks really interesting.

Jabberwok: I thought your observation about 'seeding' the main questline into sidequests in an open world game was interesting. I've been playing a lot of Divinity: Original Sin lately, and I realized that it does exactly that. At the beginning of the game, it starts you on what ends up being a lengthy sidequest, and then lets you stumble across the main story in your travels.