[Space] Fighter-to-ship ratio: what is it and why?

[Ulzgoroth]

Originally Posted by jason taylor

I am proposing a fleet coming out of jump in a high traffic zone. Or guarding a high traffic planet. Where terrorists are harassing merchant ships by coming in in ships that look like frighters and carry missiles instead of cargo.

Or a blockade of an interdicted world where the blockade runners use the same hulls friendlies are using for normal commercial traffic.

A fleet that is totally composed of AI is vulnerable to asymmetric desception.

Back in space (opera) now? Some warning about the theater change would help...

A) In space, 'high traffic' means maybe there are ships close enough to see with the naked eye. Which is to say that at the very worst everything is still spaced out enough that the 'crowding' doesn't do anything.

B) I'm pretty sure you won't find a single endorsement of an independent fleet totally composed of AI in the thread, barring some of those AIs being human-equivalent or better.

C) Why do you think AIs perform threat identification purely by low-quality surface appearance? That might be close to fair for for day-after-tomorrow autonomous UCAVs, but now you're talking about spacefuture AIs used for autonomous police work.

D) Would you really be using fighters for policing freighters anyway? Unless they've got magic scanners they can't really do anything except threaten to shoot people who get out of their lane. How would they identify a 'blockade runner' whose got a legit-looking ship and passing credentials?

[jason taylor]

Originally Posted by Ulzgoroth

Back in space (opera) now? Some warning about the theater change would help...

A) In space, 'high traffic' means maybe there are ships close enough to see with the naked eye. Which is to say that at the very worst everything is still spaced out enough that the 'crowding' doesn't do anything.

B) I'm pretty sure you won't find a single endorsement of an independent fleet totally composed of AI in the thread, barring some of those AIs being human-equivalent or better.

C) Why do you think AIs perform threat identification purely by low-quality surface appearance? That might be close to fair for for day-after-tomorrow autonomous UCAVs, but now you're talking about spacefuture AIs used for autonomous police work.

D) Would you really be using fighters for policing freighters anyway? Unless they've got magic scanners they can't really do anything except threaten to shoot people who get out of their lane. How would they identify a 'blockade runner' whose got a legit-looking ship and passing credentials?

A): A terrorist ship disguised as a merchant ship is not necessarily hindered by the fact that there is a lot of space in space and having a lot of hulls around will automatically increase the number of inspections that have to be made.

B): I am pretty sure I saw plenty of endorsements of a fleet composed of so few humans that there are none at the sharp edge except in the event of a fleet action and all recon being conducted by AI. Or in other words all the humans being so far away from the decision that they might as well be on the other side of the galaxy.

C): There is no other way to conduct inspection but by surface appearance unless your sensors can penetrate the hull.


D): An AI does not even know what "out of line" means.

[sir_pudding]

In the air, humans aren't any good at identifying things by visual recognition at the kinds of engagement ranges and speeds of air combat, they have to rely on radar and and IFF (and note that all of their weapons, even iron bombs, are now aimed by computers). Space combat is only going to be orders of magnitude farther and faster.

[Ulzgoroth]

Originally Posted by jason taylor

A): A terrorist ship disguised as a merchant ship is not necessarily hindered by the fact that there is a lot of space in space and having a lot of hulls around will automatically increase the number of inspections that have to be made.

How is number of inspections more of a problem for an AI than a fighter pilot?

Neither of which can actually do any inspections, since the things that need inspecting are inside the ships...

EDIT: My point was that there's no 'blending with the crowd', you'll never be close enough to another vessel to produce any confusion.

Originally Posted by jason taylor

B): I am pretty sure I saw plenty of endorsements of a fleet composed of so few humans that there are none at the sharp edge except in the event of a fleet action and all recon being conducted by AI. Or in other words all the humans being so far away from the decision that they might as well be on the other side of the galaxy.

Your final sentence is a complete non sequetor. Being maybe a few seconds light-lag away is no great problem on any timescale where anybody's doing serious thinking. For situations where people don't have time to think, the machine can react both faster and better.

EDIT: Also, what recon?

Originally Posted by jason taylor

C): There is no other way to conduct inspection but by surface appearance unless your sensors can penetrate the hull.

Obviously you think there's something a human can do in this situation that the AI cannot. The human will have the same sensor limitation - I assume you're expecting them to draw inference from action over time, but that's hardly something a computer is incapable of.

Originally Posted by jason taylor

D): An AI does not even know what "out of line" means.

Whyever not? Computers have no problem understanding traffic control.

[jason taylor]

Originally Posted by sir_pudding

In the air, humans aren't any good at identifying things by visual recognition at the kinds of engagement ranges and speeds of air combat, they have to rely on radar and and IFF (and note that all of their weapons, even iron bombs, are now aimed by computers). Space combat is only going to be orders of magnitude farther and faster.

Space combat depends in the first place on the assumption that the relation between the amount of space and the speed of the vehicle are sufficient to allow space combat to be a practical proposition which relation is defined by authorial decree. Otherwise you might as well give up because you have objects faster then bullets trying to find each other and fight each other in infinity.

[jason taylor]

Originally Posted by Ulzgoroth

How is number of inspections more of a problem for an AI than a fighter pilot?

Neither of which can actually do any inspections, since the things that need inspecting are inside the ships...

EDIT: My point was that there's no 'blending with the crowd', you'll never be close enough to another vessel to produce any confusion.

Your final sentence is a complete non sequetor. Being maybe a few seconds light-lag away is no great problem on any timescale where anybody's doing serious thinking. For situations where people don't have time to think, the machine can react both faster and better.

EDIT: Also, what recon?

Obviously you think there's something a human can do in this situation that the AI cannot. The human will have the same sensor limitation - I assume you're expecting them to draw inference from action over time, but that's hardly something a computer is incapable of.

Whyever not? Computers have no problem understanding traffic control.

Being a light-hour away is certainly a great problem and at that distance there is certainly need for "recon" or if that is impossible, some means of projecting power. And you are being contradictory to assert that all the objects being inspected are within light seconds of each other and that there is no such thing as intense traffic.

[Rupert]

Originally Posted by sir_pudding

In the air, humans aren't any good at identifying things by visual recognition at the kinds of engagement ranges and speeds of air combat, they have to rely on radar and and IFF (and note that all of their weapons, even iron bombs, are now aimed by computers). Space combat is only going to be orders of magnitude farther and faster.

IF the ranges are sufficiently long, and delta-vee sufficiently constrained, combat may not be fast at all (apart from the moment when you apply point-defence against incoming missiles). It could well involve long periods of manoeuvring and waiting, with plenty of time for humans to identify possible opponents using telescopes and eyeballs as well as radar, etc.

[Ulzgoroth]

Originally Posted by jason taylor

Space combat depends in the first place on the assumption that the relation between the amount of space and the speed of the vehicle are sufficient to allow space combat to be a practical proposition which relation is defined by authorial decree. Otherwise you might as well give up because you have objects faster then bullets trying to find each other and fight each other in infinity.

It's fairly well impossible to invent a situation where spaceships of similar capabilities are unable to engage one another.

And you have to commit blatant violations of science for finding each other to be problem.

Originally Posted by jason taylor

Being a light-hour away is certainly a great problem and at that distance there is certainly need for "recon" or if that is impossible, some means of projecting power.

Being a light-hour away puts you farther off than the orbit of Jupiter. At that point you're not actually in the same fleet.

But you wouldn't need recon across that distance.

[Fred Brackin]

Originally Posted by sir_pudding

Do you suppose that ALPHA isn't a TL8 program on a Complexity 3 computer in GURPS terms?

I suspect it wouldn't be anywhere near as good a replacement in reality as it was in the simulation. That may not have been any fairer than Watson's appearance on Jeopardy.

[jason taylor]

Originally Posted by Ulzgoroth

It's fairly well impossible to invent a situation where spaceships of similar capabilities are unable to engage one another.

And you have to commit blatant violations of science for finding each other to be problem.

Being a light-hour away puts you farther off than the orbit of Jupiter. At that point you're not actually in the same fleet.

But you wouldn't need recon across that distance.

Then by definition any system with a sufficient amount of objects in it is crowded.