How do you explain a computer failing a skill check?


General Discussion

1 to 50 of 54 << first < prev | 1 | 2 | next > last >>

I'm interested in hearing how people play their androids and constructs in regards to explaining why skill checks, attack rolls, and things of that sort fail. Since the characters are essentially computers, is it chalked up to an algorithm they're still testing and debugging? Is their firmware faulty or glitchy? Were they designed with failure chance to better blend into biological societies?


Androids dont think in programmed minds, they have actual souls like an organic and their minds work like an organics... with a few exceptions like dampened emotions. i assume that true robots failing checks would be due to the circumstances outpacing their ability to cope with ie moving targets swerving or ducking when their programming did not predict it, built in anti-slicing programs adapting to the preset approach a robot uses. Machines like a CNC should have a preset level that, so long as you feed them the right parts, constantly produce anything at or below the level they are capable of. Things that have free thought, Androids or AIs should have to work through it just like a PC would though they might have expert tools and circumstance bonuses inherently.


2 people marked this as a favorite.

First nitpick: if Starfinder androids will be the same as in Pathfinder, they won't be computers/robots but synthetic people with biological and artificial components.

That aside:

Attack rolls: The target successfully defended against the attack. Simple.

Skill checks: The computers are not infallible, they are operating within certain parameters, with certain physical capabilities, and with certain knowledge programmed in. E.g. When opening a lock you need to test a certain number of moves, you might fail if you take a move which blocks the internal tumbles and have to start from scratch. When failing a Knowledge check, computer might simply don't know the answer because no one ever entered it into its databanks. Or if that information was entered it might have not been properly cross-indexed, or contextualized.


2 people marked this as a favorite.

The answer is 42.


The real questions is, if an AI fails a knowledge check by 10 or more do they go all HAL 9000?


Torbyne wrote:
Androids dont think in programmed minds, they have actual souls like an organic and their minds work like an organics... with a few exceptions like dampened emotions. i assume that true robots failing checks would be due to the circumstances outpacing their ability to cope with ie moving targets swerving or ducking when their programming did not predict it, built in anti-slicing programs adapting to the preset approach a robot uses. Machines like a CNC should have a preset level that, so long as you feed them the right parts, constantly produce anything at or below the level they are capable of. Things that have free thought, Androids or AIs should have to work through it just like a PC would though they might have expert tools and circumstance bonuses inherently.

Ah! Bummer. I was lucky enough to binge Westworld the other week and it got me all excited to play as a programmable sentient being. I've seen other threads asking if there was going to be a more basic playable robot race like C3-P0, perhaps I was reading about this very difference without even knowing it.


Drejk wrote:

First nitpick: if Starfinder androids will be the same as in Pathfinder, they won't be computers/robots but synthetic people with biological and artificial components.

If I might nitpick your nitpick, while Androids certainly act organic (reacting to healing magic and the like), they are entirely synthetic. They just emulate organic life so well, even magic has trouble telling the difference. There's some language about Androids that even describes their bodies as "installing" new souls when they rejuvenate at the end of their lifespan.


Brew Bird wrote:
Drejk wrote:

First nitpick: if Starfinder androids will be the same as in Pathfinder, they won't be computers/robots but synthetic people with biological and artificial components.

If I might nitpick your nitpick, while Androids certainly act organic (reacting to healing magic and the like), they are entirely synthetic. They just emulate organic life so well, even magic has trouble telling the difference. There's some language about Androids that even describes their bodies as "installing" new souls when they rejuvenate at the end of their lifespan.

I would like to nitpick your nitpick of a nitpick... but I completely agree with you, so I've got nothing. That said. Ever thought about the fact that androids have the ability to absorb a soul? If that is the case, was the soul intended for the android in the first place? Is there a newborn that ends up born without a soul due to the soul being taken by the android upon rejuvenation? If it's possible for a machine to act as a receptacle for souls, then can another machine be built to repeatedly trap souls before they can reach a body?


McBugman wrote:
Torbyne wrote:
Androids dont think in programmed minds, they have actual souls like an organic and their minds work like an organics... with a few exceptions like dampened emotions. i assume that true robots failing checks would be due to the circumstances outpacing their ability to cope with ie moving targets swerving or ducking when their programming did not predict it, built in anti-slicing programs adapting to the preset approach a robot uses. Machines like a CNC should have a preset level that, so long as you feed them the right parts, constantly produce anything at or below the level they are capable of. Things that have free thought, Androids or AIs should have to work through it just like a PC would though they might have expert tools and circumstance bonuses inherently.
Ah! Bummer. I was lucky enough to binge Westworld the other week and it got me all excited to play as a programmable sentient being. I've seen other threads asking if there was going to be a more basic playable robot race like C3-P0, perhaps I was reading about this very difference without even knowing it.

Even in westworld the Hosts were able to miscalibrate to the point that the maintainence techs had trouble fixing them... and the way that remembered past lives certainly distracted them and caused them to perform below expectations of "machine perfection".


About those android souls... The Androffans were apparently unfamiliar with magic in any context, they had abandoned their gods ages before Pathfinder kicked off and had lost all knowledge of things arcane. The Androffans never expected Androids to develop consciousness as they did and had no knowledge of souls to begin with. But Androids do possess organs that to all but an expert would appear the same as a regular humans, possibly even able to donate or receive a transplant from humans. Which begs the question, if some Androids reject their superficial resemblance to humans and strive to be uniquely Android, are there any at the opposite end who desire to graft on organic components in an effort to truly become human?

Since Android is a gender based term, does Starfinder recognize Androids, Gynoids and those who reject gender as just Droids?


Torbyne wrote:
McBugman wrote:
Torbyne wrote:
Androids dont think in programmed minds, they have actual souls like an organic and their minds work like an organics... with a few exceptions like dampened emotions. i assume that true robots failing checks would be due to the circumstances outpacing their ability to cope with ie moving targets swerving or ducking when their programming did not predict it, built in anti-slicing programs adapting to the preset approach a robot uses. Machines like a CNC should have a preset level that, so long as you feed them the right parts, constantly produce anything at or below the level they are capable of. Things that have free thought, Androids or AIs should have to work through it just like a PC would though they might have expert tools and circumstance bonuses inherently.
Ah! Bummer. I was lucky enough to binge Westworld the other week and it got me all excited to play as a programmable sentient being. I've seen other threads asking if there was going to be a more basic playable robot race like C3-P0, perhaps I was reading about this very difference without even knowing it.
Even in westworld the Hosts were able to miscalibrate to the point that the maintainence techs had trouble fixing them... and the way that remembered past lives certainly distracted them and caused them to perform below expectations of "machine perfection".

Very true, but the Hosts are a race at the brink of achieving sentience. I had imagined Androids as an older race and thus able to provide a massive catalog of experience data with finely tuned algorithms and sub routines. This race would likely have a huge ego hubris and would experience failures in situations that biological races acted at low probability outcomes.

I'm reminded of Bladerunner's replicants a little but don't remember if they were fully organic or programable at the moment.

This also brings to the table the possibility of a synthetic mind having superior processing speed. Very likely they could operate with drastically lower input lag too and make minor corrections mid action. I would see this as high skill or attack buffs or some sort of a reroll system.


McBugman wrote:
I was lucky enough to binge Westworld the other week and it got me all excited to play as a programmable sentient being.

That'd be an AI in a robot. I haven't heard much about AI PCs, but if that isn't an option out of the gate, I'll be really surprised if it isn't eventually added. It'd be fascinating to have a mechanic similar to Eclipse Phase's morphs for AI PCs to move between robot "sleeves".


And now I'm gonna boggle for the next day or two about which word in there triggered the blue brigade...


2 people marked this as a favorite.
Garrett Guillotte wrote:
McBugman wrote:
I was lucky enough to binge Westworld the other week and it got me all excited to play as a programmable sentient being.
That'd be an AI in a robot. I haven't heard much about AI PCs, but if that isn't an option out of the gate, I'll be really surprised if it isn't eventually added. It'd be fascinating to have a mechanic similar to Eclipse Phase's morphs for AI PCs to move between robot "sleeves".

I've had this happen to me. :-)

Edit: Now, I'm pretty. :-)


so offended


Why is my computer coughing?


8 people marked this as a favorite.

OP...have you ever used windows?


1 person marked this as a favorite.

Anyway, back on topic...

There are tons of reasons why software, computers, and robots intermittently fail: they temporarily run out of memory, a processing thread times out, storage media is temporarily corrupted and error checking hasn't caught it yet, hardware interrupts conflict, a temperature sensor malfunctions, something trips an overzealous exception handler, a resource pool temporarily runs out, somebody forgot to gracefully handle a specific type of character encoding, a servo is a fraction of a centimeter out of alignment, some poorly handled edge case in dependent but separately developed libraries results in valid but inaccurate output, a random number generator doesn't reseed properly, an internal clock is off by less than a second resulting in a timing error, an automation script encounters a once-in-a-century scenario it can't handle, an unrelated log file everyone forgot about eventually fills up a necessary storage device, or the classic "a literal insect bug got inside and fried itself on a pathway".

Well-designed stuff fails all the time, especially in uncontrolled circumstances. The result depends on how well (or poorly) all of the interconnected systems handle that failure.


Archmage Variel wrote:
Brew Bird wrote:
Drejk wrote:

First nitpick: if Starfinder androids will be the same as in Pathfinder, they won't be computers/robots but synthetic people with biological and artificial components.

If I might nitpick your nitpick, while Androids certainly act organic (reacting to healing magic and the like), they are entirely synthetic. They just emulate organic life so well, even magic has trouble telling the difference. There's some language about Androids that even describes their bodies as "installing" new souls when they rejuvenate at the end of their lifespan.
I would like to nitpick your nitpick of a nitpick... but I completely agree with you, so I've got nothing. That said. Ever thought about the fact that androids have the ability to absorb a soul? If that is the case, was the soul intended for the android in the first place? Is there a newborn that ends up born without a soul due to the soul being taken by the android upon rejuvenation? If it's possible for a machine to act as a receptacle for souls, then can another machine be built to repeatedly trap souls before they can reach a body?

Redemption Engine spoilers:
In Redemption Engine, one of the bad guys has made a magical machine that can trap a multitude of souls as they leave the body and are making their way to Pharasma. I don't see why a machine that catches souls going the other way couldn't be made.

Lemartes wrote:
OP...have you ever used windows?

Yeah, I've been known to open them on nice days.

Garrett Guillotte wrote:
McBugman wrote:
...
That'd be an AI in a robot. I haven't heard much about AI PCs, but if that isn't an option out of the gate, I'll be really surprised if it isn't eventually added. It'd be fascinating to have a mechanic similar to Eclipse Phase's morphs for AI PCs to move between robot "sleeves".

Exactly! Alright then, well fingers crossed it comes in the Alien Archive if not in the core book! I like a lot of the ideas in your list though.

To possibly clear up the original question, computers can obviously malfunction so how would you role play an AI character handling these malfunctions? What would you make their perspective and reasoning like in those situations that they rolled a 1?


Pretty sure it has something to do with BGP.


McBugman wrote:
Torbyne wrote:
McBugman wrote:
Torbyne wrote:
Androids dont think in programmed minds, they have actual souls like an organic and their minds work like an organics... with a few exceptions like dampened emotions. i assume that true robots failing checks would be due to the circumstances outpacing their ability to cope with ie moving targets swerving or ducking when their programming did not predict it, built in anti-slicing programs adapting to the preset approach a robot uses. Machines like a CNC should have a preset level that, so long as you feed them the right parts, constantly produce anything at or below the level they are capable of. Things that have free thought, Androids or AIs should have to work through it just like a PC would though they might have expert tools and circumstance bonuses inherently.
Ah! Bummer. I was lucky enough to binge Westworld the other week and it got me all excited to play as a programmable sentient being. I've seen other threads asking if there was going to be a more basic playable robot race like C3-P0, perhaps I was reading about this very difference without even knowing it.
Even in westworld the Hosts were able to miscalibrate to the point that the maintainence techs had trouble fixing them... and the way that remembered past lives certainly distracted them and caused them to perform below expectations of "machine perfection".

Very true, but the Hosts are a race at the brink of achieving sentience. I had imagined Androids as an older race and thus able to provide a massive catalog of experience data with finely tuned algorithms and sub routines. This race would likely have a huge ego hubris and would experience failures in situations that biological races acted at low probability outcomes.

I'm reminded of Bladerunner's replicants a little but don't remember if they were fully organic or programable at the moment.

This also brings to the table the possibility of a synthetic mind having superior processing speed. Very...

The idea of a species living for centuries and developing massive pools of knowledge and such for their lifetimes of experiences is actually already a thing in Pathfinder and its... not that great really, any species that has lived over 100 years can take a feat to grant a whopping +2 to all knowledge checks. you can argue about machines remembering better or something but in the end its the same kind of experience, and where as an organic might need to reference back to a book to remember it all the machine will have to defrag its probably ridiculously cluttered drives to get the same information.

As is, Android souls seem to leave on their own when their time is up... which makes for interesting cases, you could have a wealthy and powerful Android who doesnt want to lose everything to the government due to their lack of heirs so they leave all kinds of diaries and how to videos to the new soul in their shell so that they can assume to identity and keep all their stuff.

I think Replicants were basically cloned to order, made to be better in all regards but effetively made out of the same stuff as humans, hence why they had the turing test instead of blood test. If they could broadcast a signal and reset or change them then it would be very easy to broadcast the signal and shut down rogue Replicants. It also makes the stories a heck of a lot darker in that they really were a created slave race with a built in timer to their deaths.

AI's with super minds may be a thing but for PCs it would probably be reflected as something like +2 int and +2 to any INT based check, you could buy both cybertech INT upgrades as well as skill programs to boost INT checks further. I think we have seen all of that in the Pathfinder Techguide already actually.

What do you think happens when a new soul occupies a fully kitted out Android Shell? Something with integrated plasma emitters and a full assassination protocol package installed? Awkward first week?


Some complication came up they weren't expecting. Programs are terrible at dealing with things they weren't expecting.


2 people marked this as a favorite.

Thou shalt not make a machine in the likeness of a human mind.


The Orange Catholic Bible wrote:
Thou shalt not make a machine in the likeness of a human mind.

So says robot God! Beep boop!


1 person marked this as a favorite.
Torbyne wrote:
What do you think happens when a new soul occupies a fully kitted out Android Shell? Something with integrated plasma emitters and a full assassination protocol package installed? Awkward first week?

Ha! They become Guy Pearce in Memento?

It would be interesting to experiment with a homebrew where player characters cycle through souls several times. To fully capture the mood players would probably have to get a randomly rolled new shell with level appropriate wealth. I doubt many people would be willing to wipe their character clean every 5 levels or so though..


Failing an attack roll could mean a number of things. Maybe you missed. Maybe they dodged or parried. Maybe it hit but did an inconsequential amount of damage (if any). Maybe it bounced off their armor or they clocked it with a shield. Maybe there was something in the way you didn't account for. There's a lot of possibilities.

As for skill checks, it depends on the circumstances and the skill in question, but most of the answers here have already covered most of those.


Garrett Guillotte wrote:
McBugman wrote:
I was lucky enough to binge Westworld the other week and it got me all excited to play as a programmable sentient being.
That'd be an AI in a robot. I haven't heard much about AI PCs, but if that isn't an option out of the gate, I'll be really surprised if it isn't eventually added. It'd be fascinating to have a mechanic similar to Eclipse Phase's morphs for AI PCs to move between robot "sleeves".

Go look up the game system Nova Praxis. You just summed up half of humanity in that game.

The Exchange

McBugman wrote:
I'm interested in hearing how people play their androids and constructs in regards to explaining why skill checks, attack rolls, and things of that sort fail. Since the characters are essentially computers, is it chalked up to an algorithm they're still testing and debugging? Is their firmware faulty or glitchy? Were they designed with failure chance to better blend into biological societies?

Floating Point Unit Bug


1 person marked this as a favorite.

I love all of the transhuman RPGs out there. Eclipse Phase is definitely the most fantastical of them all and borders on posthumanism. Nova Praxis feels more like the early stages of transhumanism. Transhuman Space is like the very middle.


1 person marked this as a favorite.
Brew Bird wrote:
Drejk wrote:

First nitpick: if Starfinder androids will be the same as in Pathfinder, they won't be computers/robots but synthetic people with biological and artificial components.

If I might nitpick your nitpick, while Androids certainly act organic (reacting to healing magic and the like), they are entirely synthetic. They just emulate organic life so well, even magic has trouble telling the difference. There's some language about Androids that even describes their bodies as "installing" new souls when they rejuvenate at the end of their lifespan.

To nitpick your nitpick about my nitpick:
Organic and synthetic are not mutually exclusive, despite what marketing tries to sell you. Nothing in the definition of "organic life" says it excludes artificially created life forms, in fact there is a interdisciplinary branch of biology called "synthetic biology" focused on engineering new biological parts, systems, and possibly creatures. Synthetic life is one of its topics and goals.

McBugman wrote:
I'm interested in hearing how people play their androids and constructs in regards to explaining why skill checks, attack rolls, and things of that sort fail. Since the characters are essentially computers, is it chalked up to an algorithm they're still testing and debugging? Is their firmware faulty or glitchy? Were they designed with failure chance to better blend into biological societies?

Fluid dynamics and wind resistance are a little hard to caluclate

The rock i was holding LOOKED solid...

Grand Lodge

3 people marked this as a favorite.
Pathfinder Roleplaying Game Superscriber

"Computer call the grocery store"

"Did you say, 'Start nuclear war?'"


Torbyne wrote:
and those who reject gender as just Droids?

Droid is TM Lucasfilm.

Scarab Sages

1 person marked this as a favorite.

A couple of thoughts:

1). Trash in trash out
Perhaps the robots io (input / output) capabilities are limited. Resulting in faulty information gathering, which results in a incorrect assessment of a situtation and a faulty answer.

2). Incorrect data or damaged data to start with

"It must be true. The Internet says so."
--Abraham Lincoln

Perhaps a younger or newer model has either incorrect data or damaged data that would result in a faulty answer. As the robot grows, his ability to clean and improve his data improves.

3). Does Not Compute
The robot does not have information, enough information or the ability to process the question. So it guesses as best it can. Think iPhone auto-completion errors.


All things considered, human society is fairly patriarchal when it comes to etymology. Under the majority of (especially early) literary circumstances, we utilize terms such as mankind, man made, and manpower etc. to represent humanity as a whole. Though I may not agree with how that terminology should have developed, it's hard to disagree it isn't a fairly common theme in humanities history, and gives the world of Starfinder another (probably unintentional) dimension of realism with links to our own reality and experiences. That said, I'm not sure whether I'm going to want to play an android myself (because I'm not sure I'll like the relationship of magic they have, considering the current pathfinder lore) but I have a friend who's super excited for them.


Owen KC Stephens wrote:
Torbyne wrote:
and those who reject gender as just Droids?
Droid is TM Lucasfilm.

Wow, really? I have always used it as a settting agnostic term... and isnt it a specific brand for Android phones, that whole "Droid does" slogan? weird.

So no official use of Droid than. what about Synth or something? it just strikes me that in their situation, if they wished to establish themselves as apart from gendered organics they would use a wholly new term for themselves as a species.

As much as i want to know the rules of the game, i am very much looking forward to the race write ups and known history sections. between nostalgia for B5 and binging the Expanse i really want some tensions in the Pact Worlds to exploit for campaigns. hoping for some good factions within species instead of generic most humans are this, most elves are like that... but that would probably eat too many pages in the book.

Man, i hope you guys are comfortable knowing how long you will be writing material for this setting :P

The Exchange

Torbyne wrote:
Owen KC Stephens wrote:
Torbyne wrote:
and those who reject gender as just Droids?
Droid is TM Lucasfilm.
Wow, really? I have always used it as a settting agnostic term... and isnt it a specific brand for Android phones, that whole "Droid does" slogan? weird.

Yeah, and after they released the first version of the phone Lucasfilm sued them. They reached a settlement allowing "Droid" to continue to be used for the phone (in return for an undisclosed but likely sizable cash payment). Next time you see an ad for one take a look at the bottom where it will say in tiny type that "Droid is TM Lucasfilm and used under license."

Edit: it's also worth noting that Lucasfilm didn't file the trademark until something like 2007. So anyone who used the term in the 80's, 90's or early 2000's is pretty safe to continue using it. As far as I know no one has challenged the trademark in court.


Belafon wrote:
Edit: it's also worth noting that Lucasfilm didn't file the trademark until something like 2007. So anyone who used the term in the 80's, 90's or early 2000's is pretty safe to continue using it. As far as I know no one has challenged the trademark in court.

I am not sure this is entirely accurate. When Battletech first came out in the very early 80's it was actually called BattleDROIDS and a lawsuit from Lucasfilm forced them to change the name. While I am not 100% positive that Lucasfilm won in court over a copyright infringement judgement or FASA just caved at the potential legal costs of battling Lucasfilm in court but I know for 100% sure the case happened.


1 person marked this as a favorite.

How do I explain a computer failing a skill check?

That's what I have to do every day, 40(+) hours a week to earn my paycheck. We have teams of people making careers out of answering that question on computer systems orders of magnitude less complex. It's actually more surprising when you get a real world computer to MAKE a skill check! ;)

Seriously, however, there's no real reason to think computers would never fail checks for all the reasons Garrett Guillotte awesomely lists out. Computers can be as prone to failure as us biologicals.

However, that answer isn't very interesting for a game.

For those that work in AI its interesting that computers are really good at things that humans find extremely difficult (complex math being an obvious one), but are actually really bad at things we find very easy (vision, subtle pattern recognition, factoring context into problem solving, etc.). If you wanted to make androids and AIs feel very different than humanoid races, one interesting thing to play with is which skills they can Take 10 on (if that's even a thing in Starfinder). Maybe biologicals can't take 10 on knowledge checks - they always have to roll, but AIs and androids can, and other skills are flipped.

Or maybe break up the skills so that androids and AI can "take 12" on some skills (knowledge skills?) but only "take 8" on others (like Sense Motive??) and biologicals would have different sets of skills they could take 8 or take 12 on.

Some flat penalties and bonuses might work better. I don't know, I'm making this up off the top of my head and it would certainly add complexity. But I was trying to think of something that could convey that different-ness in base ability without just being a bonus or penalty to optimize, and instead have it be something just sort of off and different. Like in some situations one can look at the other surprised and say "How can you do that so easily?" but it flips in different situations with the human amazed at what the AI can do with little effort and other times the AI amazed at what the human do with little effort.


One thing I heard in a computer science class was that sometimes, instead of writing a slow program that solves a problem exactly, you can write a fast program that solves a problem well enough. Of course, sometimes that isn't enough, and you fail at a task.

In a robot, you of course don't want it to take a century to solve a problem (ie hitting an enemy), but let it run a 2-second algorithm that hits the enemy a good portion of the time.


Pathfinder Maps, Pathfinder Accessories Subscriber; Pathfinder Roleplaying Game Superscriber; Starfinder Charter Superscriber
Garrett Guillotte wrote:
And now I'm gonna boggle for the next day or two about which word in there triggered the blue brigade...

Regarding that, I have a theory to test: "Phase's morphs"

EDIT: Yep, that did it.


So Droid is out, how about Altmen, Nuone, Artie (derogatory term from Artificial)... i know, i know, off topic. But a fun topic!


's morph


What happens to a poster who types "'s morph"?

The same thing that happens to everything else.


Mighty 's morphing powered androids.


Pathfinder Maps, Starfinder Adventure Path, Starfinder Maps, Starfinder Roleplaying Game, Starfinder Society Subscriber

404, file not found


Drejk wrote:
Brew Bird wrote:
Drejk wrote:

First nitpick: if Starfinder androids will be the same as in Pathfinder, they won't be computers/robots but synthetic people with biological and artificial components.

If I might nitpick your nitpick, while Androids certainly act organic (reacting to healing magic and the like), they are entirely synthetic. They just emulate organic life so well, even magic has trouble telling the difference. There's some language about Androids that even describes their bodies as "installing" new souls when they rejuvenate at the end of their lifespan.
** spoiler omitted **

This. I hold it in fantasy as well (inorganic=/=unnatural) and that's in setting without biology not based in carbon.

1 to 50 of 54 << first < prev | 1 | 2 | next > last >>
Community / Forums / Starfinder / Starfinder General Discussion / How do you explain a computer failing a skill check? All Messageboards

Want to post a reply? Sign in.