Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Detoxicated]

EAC-TML2 Demi-Sarissa (2): Andrea, Detoxicated

[andrea]

EAC-TML2 Demi-Sarissa (2): Andrea, Detoxicated

[joha4270]

EAC-TML2 Demi-Sarissa (3): Andrea, Detoxicated, joha4270

[Chiefwaffles]

Revision: Athena - Flexibility

We have had two failed attempts to incorporate Athena in non-space combat roles. Industry overseeing, and now the Sarissa. While unfortunate, this represents an opportunity. The failure of Industry Overseeing gave us a wealth of knowledge on how to improve Athena and her learning components. And while we didn't have any such realization with Sarissa, it's another step closer to realizing how we would theoretically properly implement Athena in these non-traditional roles.

So let's improve Athena instead. Make her much more flexible and robust. Use everything we've learned from these failed revisions to actually make Athena able to properly learn things that don't revolve to "dodging spaceships" and "shooting laser turrets." So she can actually start being useful in ground combat and industry. Though (while it'd be nice) we don't need a complete understanding of ground combat. Enough to allow good missile targeting.

The results should be obvious and immediate. Athena should be able to actually effectively operate guidance for the missiles as well as she manages orbital evasion and turret targeting. Athena should just be better in the ways already described. And we'd love to be able to plug Athena into the Hephaestus plants and see if she can handle it this time, since we should have already prepared the plants for Athena control in Industry Overseeing even if we completely failed at Athena successfully controlling them. But no special effort other than "I guess try it" will be given to the industry overseeing. The main point of this revision is making Athena more robust and flexible, and thus actually able to guide missiles as well as or better than intended.
Hopefully the missiles should top misfiring too - we believe it's a software issue where the guidance is faulty, resulting in premature detonations. So better guidance should hopefully fix mixfiring.

And the range problem?

With not very impressive ranges, at that. Our efforts to make a guidance system were stymied to the point that just dumb-firing them was more accurate.

If what our report suggests is accurate, then the range is caused less by "not enough fuel" and more "can only go in a kind-of straight line", which, as it happens, is really bad for range. So fixing guidance should help notably in this department without any extra effort from us.
(EDIT: According to Nuke on discord, the range is poor [but not "terrible"] for its own non-guidance-related reasons. Range can still be improved a bit by guidance, so I'm leaving this here.)



Also, we replace the engine in the Sarissa with an electrical one while we're at it. I mean, really? We're pretty sure electric engines aren't that hard to make.
And fire whoever ordered "500x Internal Combustion Engines" instead of "500x Electric Engines." Again: this is a really perplexing mistake. Maybe it was Quillus sabotage? Surely no engineer could be that incompetent.


TL;DR: We have a specific bonus in getting Athena better at learning and whatnot from the failed Industry Overseeing. We have more experience on how to (not) get Athena to do stuff outside of orbital combat. So we use all this experience to get Athena to a state where she can actually properly fly the missiles and not explode them before they launch. _{Bonus points if the new Athena can operate our industrial plants with no extra effort from us regarding that specific thing.}
And replace the engine. Sheesh.
TL;DR TL;DR: Use our experience from 2 failed actions learning experiences to make Athena have more flexible learning. This is all we do. Everything else - namely a functional Sarissa - should just be a (desired) consequence of this revision.

EAC-TML2 Demi-Sarissa (3): Andrea, Detoxicated, joha4270
Athena - Flexibility (1): Chiefwaffles

[3_14159]

Aside from the "I told you so", I believe fixing the Sarissa is the most important thing. And I mean fixing it, not upgrading our AI while at the same time fixing everything about the missile. To quote Nuke: "It was Hard, but unless you rolled very well you wouldn't have gotten everything." Let's not make the same mistake twice.

EAC-TML2 Demi-Sarissa (4): Andrea, Detoxicated, joha4270, 3_14159
Athena - Flexibility (1): Chiefwaffles

[Detoxicated]

Indeed though our athena shall be revised soon... maybe i will work on a new manta with missiles...

[Chiefwaffles]

The idea is that the revision solves one problem which effectively solves all the symptoms as well.
Instead of just solving the symptoms one-by-one.

Design: Athena Tactical Network: http://www.bay12forums.com/smf/index.php?topic=168247.msg7630181#msg7630181
Make Athena based on the broken targeting system that couldn't compare to modern targeting. Easy.

Revision: Athena - Robust Learning: http://www.bay12forums.com/smf/index.php?topic=168247.msg7631525#msg7631525
Improve Athena's learning capabilities. Easy.

Both are easy.
We're doing the same thing as Robust Learning ("make Athena learn betterer!"), Robust Learning was Easy, and we have a large bonus to improving Athena.

Hephaestus - Athena Overseeing
...
The potential for some really interesting things was there, and we have at least learnt a lot about how learning algorithms work (or don't work, as the case may be).

So, to sum it up:
Athena Flexibility will almost certainly be Easy or better, and it address the root problem with our missiles: Athena. The fixes are a (desired) side effect of fixing that problem. We get an Easy revision, we improve Athena, and most importantly, we fix the Sarissa.


EDIT:
The Demi-Sarissa, in a revision, implements actual guidance, fixes the launch problems, and adds range.
These are all minor problems (guidance being the biggest), but the Demi-Sarissa tries to do them all individually at once. This is a problem - the way I see it, revisions are best when they address one thing. Not to say that revisions touching on >1 thing are automatically bad or Hard or anything, but that when possible, I think it's a better idea to isolate the problem into one fixable thing.

"But Chief!" you may say, "Your revision does the same thing!"
Ah, but it doesn't. My revision is "make Athena better at flexible learning." That's it! Everything else in my revision is an outline of what I expect to immediately come from doing that. It's not nice to make the GM have to guess about what people actually want their revision/design to do.

Athena Flexibility could very well be easier, and would definitely improve guidance - explicitly the worst problem in the Sarissa - a lot more than the Demi-Sarissa. The Demi-Sarissa primarily addresses guidance, but nowhere near the degree of Athena Flexibility. It focuses on doing things like adding target recognition to the missile so it can track moving targets after being given coordinates. Except we already made a targeting system. A really good targeting system. One that doesn't need extra effort to implement.
Athena.
And better yet, implementing this already-existent targeting system can fix most of our other problems as well.

Guidance is both the worst flaw of the missile and the one thing we need to really focus on. We have a small blast radius and they have really fast targets. Guidance is what determines exactly how useful the Sarissa will be. If it can't hit its targets, it doesn't matter if it has really good range or good speed or anything.

[Blood_Librarian]

EAC-TML2 Demi-Sarissa (4): Andrea, Detoxicated, joha4270, 3_14159
Athena - Flexibility (2): Chiefwaffles, Blood_librarian

[Detoxicated]

While your points are valid and true we still need to fix the technical issues of this thing first. If we manage to create another weapon or ship next turn we can upgrade our system in the way it fits us then.

[Chiefwaffles]

My points are about fixing the technical issues of this thing first.

1.) Guidance - We have to have really good guidance to actually have an impact. Our hits have to be precise, and they have really fast tanks.
2.) Difficulty - Athena Flexibility solves the problems of Sarissa by addressing one thing that we have a bonus on (and a previous Athena revision that was basically the same thing - "better learning" - got Easy). The Demi-Sarissa tries to fix each individual symptom, and does not have a bonus.

Otherwise, it's pretty much the same thing except Athena Flexibility also gives us benefits in other areas.

[sprinkled chariot]

EAC-TML2 Demi-Sarissa (4): Andrea, Detoxicated, joha4270, 3_14159
Athena - Flexibility (3: Chiefwaffles, Blood_librarian, Sprinkled chariot

[piratejoe]

EAC-TML2 Demi-Sarissa (4): Andrea, Detoxicated, joha4270, 3_14159
Athena - Flexibility (4): Chiefwaffles, Blood_librarian, Sprinkled chariot, Piratejoe

[Khang36]

EAC-TML2 Demi-Sarissa (4): Andrea, Detoxicated, joha4270, 3_14159
Athena - Flexibility (5): Chiefwaffles, Blood_librarian, Sprinkled chariot, Piratejoe,khang

[joha4270]

EAC-TML2 Demi-Sarissa (3): Andrea, Detoxicated, 3_14159
Athena - Flexibility (6): Chiefwaffles, Blood_librarian, Sprinkled chariot, Piratejoe,khang, joha4270

[NUKE9.13]

Revision: Athena - Flexibility

We have had two failed attempts to incorporate Athena in non-space combat roles. Industry overseeing, and now the Sarissa. While unfortunate, this represents an opportunity. The failure of Industry Overseeing gave us a wealth of knowledge on how to improve Athena and her learning components. And while we didn't have any such realization with Sarissa, it's another step closer to realizing how we would theoretically properly implement Athena in these non-traditional roles.

So let's improve Athena instead. Make her much more flexible and robust. Use everything we've learned from these failed revisions to actually make Athena able to properly learn things that don't revolve to "dodging spaceships" and "shooting laser turrets." So she can actually start being useful in ground combat and industry. Though (while it'd be nice) we don't need a complete understanding of ground combat. Enough to allow good missile targeting.

The results should be obvious and immediate. Athena should be able to actually effectively operate guidance for the missiles as well as she manages orbital evasion and turret targeting. Athena should just be better in the ways already described. And we'd love to be able to plug Athena into the Hephaestus plants and see if she can handle it this time, since we should have already prepared the plants for Athena control in Industry Overseeing even if we completely failed at Athena successfully controlling them. But no special effort other than "I guess try it" will be given to the industry overseeing. The main point of this revision is making Athena more robust and flexible, and thus actually able to guide missiles as well as or better than intended.
Hopefully the missiles should top misfiring too - we believe it's a software issue where the guidance is faulty, resulting in premature detonations. So better guidance should hopefully fix mixfiring.

And the range problem?

With not very impressive ranges, at that. Our efforts to make a guidance system were stymied to the point that just dumb-firing them was more accurate.

If what our report suggests is accurate, then the range is caused less by "not enough fuel" and more "can only go in a kind-of straight line", which, as it happens, is really bad for range. So fixing guidance should help notably in this department without any extra effort from us.
(EDIT: According to Nuke on discord, the range is poor [but not "terrible"] for its own non-guidance-related reasons. Range can still be improved a bit by guidance, so I'm leaving this here.)



Also, we replace the engine in the Sarissa with an electrical one while we're at it. I mean, really? We're pretty sure electric engines aren't that hard to make.
And fire whoever ordered "500x Internal Combustion Engines" instead of "500x Electric Engines." Again: this is a really perplexing mistake. Maybe it was Quillus sabotage? Surely no engineer could be that incompetent.


TL;DR: We have a specific bonus in getting Athena better at learning and whatnot from the failed Industry Overseeing. We have more experience on how to (not) get Athena to do stuff outside of orbital combat. So we use all this experience to get Athena to a state where she can actually properly fly the missiles and not explode them before they launch. _{Bonus points if the new Athena can operate our industrial plants with no extra effort from us regarding that specific thing.}
And replace the engine. Sheesh.
TL;DR TL;DR: Use our experience from 2 failed actions learning experiences to make Athena have more flexible learning. This is all we do. Everything else - namely a functional Sarissa - should just be a (desired) consequence of this revision.

Athena - Flexibility (Easy): (4+1)+1=6: Above Average
With a little spit and polish, we can make Athena do what some engineers always believed she could do- integrate new experience types into her learning algorithms. In this case, we've added "Flying Missiles" and "Ground Targets" to her repertoire. And after running her through a few simulations, she seems to be performing quite well.
She can take control of our missile's terrible guidance systems, and has somehow figured out how to make them work anyway. Once nearing the target location, she can make a relatively intelligent analysis of incoming sensor data, which allows her to steer the missile to where it will do the most damage.
The misfiring bug... turned out not to be a code bug. However, Athena is able to disable the warhead should the missile not fire, so at least we won't be blowing ourselves up- we'll just have to reset the firing system every couple of shots.
We also managed to replace the combustion engine in the truck with an electric one, which combined with NQ Cells might actually result in improved performance even in-atmosphere.


It is now the Tactics Phase of SY103.

Spoiler: Tech List (click to show/hide)

Ships
ITC (Interplanetary Transport Craft): The ships that carried colonists to your homeworld, now repurposed as military transports. They use rocket shuttles to ferry things to and from the surface. These shuttles have a limited capacity, meaning that with the exception of troops, the ITC cannot carry any equipment of a size greater than 1TC. It has 4TC. It uses chemical engines for propulsion. It has no armour or armaments. It is Very Slow, and quite large.
Cost: 2SPP

ESC-FS 1 'Cuttlefish' laser corvette: A medium-sized spaceship. The Cuttlefish's hull vaguely resembles its namesake. The 'beak' is a "Hermes" variable-water-injected fusion thruster, which draws plasma from the Helios-S reactor. It has very light armour, and has three AX4-Spear turrets, which are automatically aimed by the Athena RL System, which also handles evasive manoeuvres. The more ships in the fleet, the better Athena performs, due to her distributed computing. Its speed is Moderate, though it can inject extra water into the thruster for a temporary speed boost.
Cost: 2SPP
  Island-Alloy Armour: A few plates of Island-Alloy, formed into a basic armour covering for the Cuttlefish. Despite being thin, they increase survivability by as much as 100%.
  Cost: +1SPP


Equipment
Printed Pistol: Issued to security guards before the war, the only real weapon both colonies started with. They are 3D-printed out of plastic, a process which is not entirely reliable, meaning the gun has a tendency to break after prolonged use... or any use, really.
Cost: Cheap

X100 Space Suit: A civilian space suit, totally unarmoured and cumbersome to walk in. At least it works well enough, and has a self-sealing lining meaning a single shot isn't an immediate death sentence... most of the time. Necessary for deployment to worlds without breathable atmospheres.
Cost: Cheap
  X100-I: Literally just a few plates of Island-Alloy bolted onto the outside of an X100.
  Cost: 2GPP

EAF-X1 "Ranger" Laser Rifle: A big, bulky firearm, weighing in at 15kg- just barely wieldable whilst standing. Works best against stationary targets. Can fire 30 shots in pulse mode before reloading, or 15 seconds in beam mode. Beam mode has pretty good armour penetration against conventional armour. Comes with basic scope and bipod. With the scope it is accurate at Long range.
Cost: Cheap
  EAF-PX3 Raider: A variant of the Ranger, using the new Picasso Pattern. Weighs a 'mere' 7.5kg. Fires shorter pulses, and deals more damage against moving targets. Does not have a beam mode.

Interplanetary Radio: Essentially just a very big radio set, capable of reliable communication between planets. Too big to fit on an ITC. Currently only useful for taunting Quillus.
Cost: 2GPP
Weight: 2TC


Vehicles
EAC-TML1 Sarissa Missile Launcher: A nuclear-missile launcher consisting of three shuttle-loads worth of stuff: a small electric truck, a launch trailer, and the missiles themselves. The launch trailer loads 8 missiles, tipped with extremely small nuclear warheads for maximum 'tactical' performance. The missiles are controlled by Athena, who is now capable of flying missiles and identifying ground targets. Their range isn't great, and the launcher misfires (harmlessly) about 25% of the time, the guidance system is a bit wonky even under Athena's capable control, and the truck+trailer is slow and unwieldy, but still, nuclear-missile launcher!
Cost: 2GPP
Weight 2TC/1TC in shuttle.


Infrastructure
Colony Defence Grid: The reason you can't instantly invade the enemy homeworld. A series of surface-to-orbit missile launchers, the only real concession to self defence the planners of the colony made. How they work is a mystery, and no one wants to risk taking them apart to find out.

"Helios" Fusion Reactor: A large, Tokamak style reactor. Performs He3-He3 fusion, though it does not take full advantage of this. Provides 2SPP and 2GPP.
  Helios-S: A variety specifically designed for use in spaceships. A 'twisted' plasma flow allows for plasma to be siphoned off without disrupting the reaction.


Misc Tech
Nanoquantum Capacitor Cells, Escher Pattern Mk2: AKA NQ Cells. About 15 times better than chemical batteries. Palm-sized, weigh about a kilo each. Despite their apparent complexity, manufacture and usage is simple.

Libum-265K: A near-room temperature superconductor. Manufacture is complicated.

Hephaestus Synthesis: A complicated process that produces small amounts of Insulanium- an isotope of a certain element that resides in the "Island of Stability"; it is very heavy, yet relatively stable. Produced in too small amounts to make anything practical out of it, it is instead alloyed with more common metals, which substantially increases their strength even with only 0.25% Insulanium added.

Spoiler: Observed enemy tech (click to show/hide)

Ships
ITC (Interplanetary Transport Craft): New shuttle spotted:
  QS-RAS-00 'Dove': A heavier shuttle for the ITC, powered by a Loki-class Stardust reactor and fitted with Apollo thrusters. It has light armour and two 30mm Hayks. Can carry 3TC. Adds 1SPP to the cost of an ITC.

QS-CA-99 'Odin': A heavily armoured medium-sized spaceship. Her hull is all slope, the front being a point, from which slopes lead up to the widest point, then slopes lead down at the same angle, before being cut off, leaving a large flat area in the back, where the main thrusters are. Aforementioned main thrusters are three Apollos. She has five smaller thrusters mounted pointing in other directions, providing excellent manoeuvrability. Her powerplant is a Stardust reactor. She is armed with four Mjolnir turrets providing excellent coverage, and a single spinal-mount extra-long Mjolnir. Her armour consists of a layer of Heavy titanium with faraday insulation, with a layer of Light ceramics mounted a full three meters away from the main hull. Despite her bulk, she moves at a Moderate pace.
Cost: 3SPP

QS-SF-00 'Loki': A small-ish fighter, capable of independent operation for a few days before needing to resupply. A substantially downsized Stardust reactor provides power to the Apollo thrusters, as well as the singular Mjolnir coilgun. Has light armour, and a radio capable of same-orbit communication. Requires a transport to act as 'mothership'- two Lokis take up 1TC
Cost: 1SPP for two.

Equipment
Printed Pistol: No differences noticed.

X100 Space Suit: No differences noticed.

Interplanetary Radio: No differences noticed.

QS-IRW-99 'Hayk': The Hayk is a micromissile-firing rifle. It is magazine fed, has a reasonable rate of fire, and is accurate at a surprising distance- theoretically out to Long range, although few are capable of such superhuman feats. There is an ammo variant which explodes on contact, though it is less accurate and in shorter supply.
Cost: Cheap

QS-GL-02 'Fog Machine': The Fog Machine is a simple grenade launcher, basically nothing more than an oversized Hayk with a shortened barrel. It fires either smoke/chaff or explosives. Range/accuracy is a bit of an issue.
Cost: Cheap


Vehicles
QS-MBT-02 'Hrungnir': The Hrungnir is a big tank, with fairly heavy armour in the form of a layer of ceramics atop a layer of titanium (all sloped). It is powered by a Stardust reactor, basically the same one used by the Loki. The energy is used to drive electric motors which in turn drive the treads upon which the Hrungnir sits. It is equipped with a 50mm stabilized auto-loading Mjolnir coilgun, the barrel of which can also be used to fire upsized explosive Hayk rockets. It has an electronics package that handles fire control and sensors. It also has smoke/chaff dispensers, which automatically cloud the vehicle in smoke/chaff (for atmosphereless planets) when sensors detect unusual heating.
Cost: 4GPP
Weight: 3TC


Infrastructure

Colony Defence Grid: No differences noticed.

QS-CPP-96 'Stardust': A fusion reactor, copies of which can be found across Quillus, powering industry. Utilises powerful niobium-tin superconductors, cooled by liquid helium, to contain plasma in a Tokamak-style. The "Brokkr's Forge" improvements have made it slightly more effective, to the point where it reaches ignition. Grants 2GPP and 2SPP. A smaller version designed for use in spaceships also exists.


Loose Parts
QS-MAC-97 'Mjolnir': A coilgun (and definitely not a railgun) capable of firing 150mm slugs at considerable speeds. Requires a substantial power supply. Ground installations cost 1GPP for four. Ship mountings cost 1SPP for eight.

QS-MHD-98 'Apollo': A magneto-plasma-dynamic thruster of exceptional efficiency. Uses hydrogen plasma by default; heavier elements can be injected into the mix for a boost in thrust.

Spoiler: Production (click to show/hide)

Ship Production Points: 16+2+2(E)
  4 Points: 2xITC
  2 Points: 1xCuttlefish (Destroyed)
  12 Points: 1xArmoured Cuttlefish (Destroyed), 3xArmoured Cuttlefish (Damaged)
  2 Points: -

Ground Production Points: 16+2+4(E)
  4 Points: 4xX100-I
  18 Points: -
 
Units: 5
  2 Units on planet A (X100-I)
  2 Units on planet D (X100-I)
  1 Unit on planet E (no extra equipment)

Spoiler: The Map (click to show/hide)

Don't even worry about this looking nothing like a solar system. Its, uh, optimal transfers or something. Look, it's dumb, but what do you want from me?

Key:
A: A volcanic world. Constant eruptions and earthquakes result in a thoroughly unpleasant environment. 4SPP, 2GPP
  E:1/6, Q:1/6
B: Rocky world. (Almost) no atmosphere, no major tectonic activity, this planets is... boring. The most notable terrain features are the craters that litter the surface. 2SPP, 2GPP
C: An ocean world. Covered in 94% water, with a breathable atmosphere, and a few inhabitable islands. Prepare for lots of island-hopping. 2SPP, 4GPP
  Q:6/6
D: An earthlike world. Covered in 69% water, with a breathable atmosphere, most of the surface being inhabitable. Terrain varies wildly, from open plains to thick jungles- and of course vast seas. 5SPP, 5GPP
  E:4/6, Q:0/6
E: A frozen world. Almost entirely covered in ice, with an atmosphere too thin to breathe. Ice sheets make for treacherous terrain, with snow-covered mountains being not much better. 2SPP, 4GPP
  E:6/6
F: Rocky world. (Almost) no atmosphere, no major tectonic activity, this planets is... boring. The most notable terrain features are the craters that litter the surface. 2SPP, 2GPP
  Q:4/6
G: A desert world. With a few salty 'seas' covering a mere 12% of the surface, the majority of this planet is dry as a bone. Vast sandscapes dominate, punctured by low, rocky hills, and the ocassional oasis. The atmosphere is breathable. 2SPP, 4GPP.
Ertex&Quillus: Earthlike worlds. Capturing the enemy homeworld results in victory. Yield no Production Points. I mean, they do, but those are already accounted for in your basic industry.