Print

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

[Flying Dice]

Fair enough. But, again, it's already gamey enough just using square WH strengths (though that's slightly less meaningful at high tech thanks to shock damage being introduced) and metaknowledge-based sensor resolution. As I said before, Aurora has a number of metrics which can be nailed down with sufficient study and experience into most-efficient designs for just about any set of needs.

And, also again, that's fucking boring after the second or third time you do it. For much the same reason that being hyper-optimal in DF is boring as shit, actually. It's neat, it's safe, it's easy, and it does nothing of interest. Aurora is first and foremost a storytelling vehicle, not a traditional 4X--thus, making the same decisions and building the same ships just because they're the most efficient is going to result in every game playing out the exact same, unless you get unlucky with a wormhole or Queen spawn and your homeworld gets glassed/irradiated before you can build up.

I've made a conscious effort to avoid that by deliberately eyeballing designs and making use of considerations beyond meta-optimization. So. Uh. Thanks, but I'll stick to stirring the pot and seeing what comes out. I've played the "curbstomp everything with optimal comps" game well to death. I gave up on that when I caught myself farming Invaders for their tech and intentionally triggering multiple 1000% difficulty NPRs just to give my ships a target-rich environment or three.

I would, because 0.5 is also available and a non-commercial freighter sounds terrible.

Inertial Confinement Fusion is a 90k tech.
Minimum Power 0.25 is a 4k tech of the same field, so "don't have the right specialist" holds little water... surely you could at least spare the 3-6k total for the first 2 or 3 techs.
Of course there may be overarching roleplaying considerations, but that seems horribly mismatched research... especially since you agree that you'd "obviously" use lower power multiplier when available.

I burned through drive techs because I was dealing with a multi-front war against two NPRs and assorted Invader incursions, plus the usual Swarm and Precursors, scattered across five different jump lines off of Sol. I've still only got colonies in Sol and one neighboring system, because that's the sum total of my secure territory, so I don't have much need to move stuff around, especially with civilians available to ship colonists and infrastructure; it was basically just a few hundred automines and a handful of mass drivers. I've had so little need for freighters this game that I've still only got a handful of old M-P drive ones sitting around rusting, and haven't had the need to build newer designs. I don't have proper carriers and fighters yet, so I only got up to a power rating sufficient to get my system-defense FACs going at a decent clip on the other side of things.

So yeah, you know what they say about assumptions. I'd "obviously" use lower power multiplier when available, no shit Sherlock. It's not available, and I don't have any particular need to research it since I'm making essentially zero use of commercial ships on looping orders.

As for conventional-start drive design... have you ever even played a conventional start? It wasn't too bad before the drive rework, but shit's terrible now. A 0.5x power 25 HS Nuclear Thermal drive only provides 62.5 EP. A Nuclear Pulse drive of the same size and power multiplier only gives you 100 EP. That's fucking terrible, especially in the very early stages of a conventional game where you're likely operating with single-digit (quite possibly only one or two) freighters and colony ships as your sole means of transporting literally everything you need to mine out different system bodies and establish colonies (since the civilians aren't up and running yet either).

Never mind actual conventional drives, which would give you an astonishingly powerful 2.5 EP for the same setup.

But yeah. Let's assume that you're a little bit into the T/N tech tree after a conventional start. You've got Nuclear Thermal drives and no power multiplier techs. You can't make it too big, since shipyard expansion is both expensive and liable to eat up working population that could be used to support more labs or industry. So ballpark it at 30k tons--that's already two 10k expansions from the starting point of your first-built commercial yard. You can just barely get three 25HS drives onto that--at 0.5x multiplier, that gets the ship going at a lovely 319km/s when you're desperately trying to get stuff to every usable body in Sol. If you upgrade those to Nuclear Pulse drives at 0.5x, you'll zip right along at 510km/s.

Let's say that, instead, you designed those Nuclear Pulse drives at 1.0x. Now that primitive freighter is breaking, barely, the 1k km/s mark.

This is the one point where I'd argue that speed matters more than fuel efficiency for commercial shipping: the initial colonization boom in Sol before civilian traffic starts showing up. This is because the sooner you seed populations and start mining, the more of everything you'll have once you slip out of the 30-day research tick cycle because you need to ramp up military production.

So yeah. Same thing I've been saying all along: theorycraft ideals from people that apparently design ships with every tech unlocked to the exact same point (usually shortened to "every tech unlocked") and no thought for long-term viability or the design quirks that can arise when dealing with specific situations. Take, for example, the habit of designing active sensors at 16R. That's because Swarm Soldiers are 800t and you need to see them coming as soon as possible. Why overengineer fire controls? Because Invader ECM is an absolute bitch for a missile-based fleet to deal with. Why don't you have really low power multiplier techs? Because the situation didn't warrant researching them.

And also because you run into diminishing returns pretty damn quick. The power multiplier increases/decreases by 5% steps at every level. It's linear. The fuel consumption per EP modifier isn't. It's 41% at 70% EP, 18% at 50% EP, and 5% at 30% EP. Going from 1.00 to 0.70 removes 59% of the drive's fuel consumption. Going from 0.70 to 0.50 removes a further 23% from the baseline (a total reduction of 82% off of the norm). Going from 0.50 to 0.30 only removes a further 13%. Meanwhile, going from 1.00 to 0.70 removes 30% of the drive's power. Going from 0.70 to 0.50 removes an additional 20%. Going from 0.50 to 0.30 removes another 20%.

It shouldn't need to be said, but the power techs have diminishing returns. This is the same reason as why you don't pump the power multiplier way up for anything except fighters and missiles. Sure, you could use 30% power drives for that sweet 5% fuel use... but they're going to be slow as shit. And for me, at least, there's a line where further reduction of fuel costs on the power multiplier end stop being worth it. Especially since you can get up to a 50% reduction on total fuel consumption from the size of the drive.

FWIW a 0.30x 25HS Inertial Confinement drive only gets 40 EP MORE than a 1.00x Nuclear Pulse drive of the same size.

This is why I don't pay too much attention to people who spout off about theoretical design ideals without actually building and playing the ships. There's a lot of shit that seems logical and straightforward in a tidy, uncomplicated world, but as always it's what happens in the field that actually matters.

[coleslaw35]

Unrelated to what is currently being discussed, but whenever I use SM mode to create an NPR for a custom system start, the NPR will spawn in a bunch of ships, military and civilian, and lag up the game despite me telling it not to.  Am I doing something wrong?

[da_nang]

So yeah. Same thing I've been saying all along: theorycraft ideals from people that apparently design ships with every tech unlocked to the exact same point (usually shortened to "every tech unlocked") and no thought for long-term viability or the design quirks that can arise when dealing with specific situations.

On that, I can agree. Designing for every tech unlocked is ridiculous. Designing for a specific "tech level" is good and all until you need a good design now, which is why I prefer to optimization techniques that model the mechanics and uses technology and specifications as parameters. Basically the "Give me what you have and what you need at the moment and I'll give you my best effort" kind of optimization. Sure, you're spending time for short-term designs (because tech and specs always advance) but you can only predict the playing field so far before it becomes too much dice play.

Generally speaking, you have a specific minimum speed, range and possibly mileage you are looking for in your design and maximizing remaining space for the given ship size and tech at the time is usually a good idea and flexible. Often this also results in standardized speeds and range as the optimum is typically close to the minimum speed and range, though the MINLP nature of the problem obviously allows some excess. Costs could also be included, but I haven't seen any formula for predicting them (power modifier seems to be the main cost driver).

Hence why I'm nowadays doing all my optimization in GAMS or LPSolve if its small and linear. (Because MATLAB global optimization is limited and a resource hog.)

[Alastar]

@ Flying Dice:

Can we do this without blowing smoke?

If it truly was the case that you never had a cause or breathing room to research at least the first few low-power techs before Inertial Fusion, it's one game in 1000. Much more likely: "I don't have the time to fix my leaking roof, so I'll spend hours every day hauling buckets". The small investment usually would pay off rapidly and free up more resources for military matters.

Many players sink a lot of resources into their fuel logistics (both tech and ships) while building wasteful, overstressed deisgns.
Like your Solemn Oath class cruiser, which uses 3 times as much as it should for its performance.
Why the heck are you using 1.5 power engines when the fuel load for your required range only slows you down? Try 20 engines of the same size at default power multiplier, with correspondingly reduced fuel load to keep the size constant. Better speed, better range, much lower logistics burden.

FWIW a 0.30x 25HS Inertial Confinement drive only gets 40 EP MORE than a 1.00x Nuclear Pulse drive of the same size.

This is why I don't pay too much attention to people who spout off about theoretical design ideals without actually building and playing the ships. There's a lot of shit that seems logical and straightforward in a tidy, uncomplicated world, but as always it's what happens in the field that actually matters.

And you're dismissing things you should't.
the Inertial Confinement design...
- doesn't make it a military ship with the logistics trouble that implies
- costs a quarter of the BP and minerals
- requires a quarter of the crew. For cheap bulky designs crew quarters contribute significantly to total cost, same for long-deployment ones.
- has fuel requirements lowered by 97%

I already mentioned that for freighters, optimising for theoretical cost/fuel efficiency may not be as important as for more expensive ship classes:
The theoretically more efficient ship will require lower-power engines than we'd fit on anything else. It will also be bulkier. This is bad if the true cost is tying up shipyards rather than minerals. Also, a bulk means more expense on armour, which makes up a bigger fraction of costs with cheap internals: we may not cut total costs by all that much.

Good design can be reached by theory or by an empirical approach, but both have their strong and weak points.
Theory can easily check for things that are easy to formalise... like most efficiency concerns.
Experience is very useful for things that would require excessive work to model theoretically.
It seems you're fond of erecting strawmen to dismiss all theoretical considerations that don't agree withyour preconceived notions though.

[ThtblovesDF]

Good discussions, keep going. I personally do not use any Kind of Armor for anything except maybe fluff/RP ships/Carrier s. More missles are a good enough defence (PS: yes, certain melee ships ruin me on the regular).

[MarcAFK]

My designs are getting more optimised, however I still knowingly do certain things inefficiently for the sake of not being bogged down by mechanics, or simply because it's what I think is suitable. In theory I could make optimal freighters and warships, but if there's an option for stupidly overpowered ships which require an epic logistics train to support, well I'm making that logistics train and burning through every gas giant in the galaxy damnit. But nobody can say my ships are slow.

[Dozebôm Lolumzalìs]

Contemporary spacecraft have that much engine+fuel to that much mission payload because we're still using chemical rocketry.

Indeed, current spacecraft have direly low payload fractions. (come on, scientists, figure out cold fusion already! ) But I was noting how a 20% payload - hell, even a 5% payload! - is not unexpected if you want fast, long-range ships at early tech.

Code: [Select]

Refuge class Salvager 100000 tons     203 Crew     3126.125 BP      TCS 2000  TH 1562  EM 0
781 km/s     Armour 9-191     Shields 0-0     Sensors 24/24/0/0     Damage Control 1     PPV 0
Maintenance Capacity 20 MSP
Cargo 75000   Spare Berths -2    Cargo Handling Multiplier 5   
Salvager: 1 module(s) capable of salvaging 750 tons per day

Justinian-Stubbs Aerospace 312.5 EP Commercial Magnetic Fusion Drive (5)    Power 312.5    Fuel Use 3.98%    Armour 0    Exp 5%
Fuel Capacity 1,500,000 Litres    Range 67.8 billion km   (1005 days at full power)

Zoren & Ventris Techsystems CIWS-320 (3x10)    Range 1000 km     TS: 32000 km/s     ROF 5       Base 50% To Hit
Sotha-Hezekiah Thermal Sensor TH1-24 (1)     Sensitivity 24     Detect Signature 1000: 24m km
Sotha-Hezekiah EM Detection Sensor EM1-24 (1)     Sensitivity 24     Detect Strength 1000: 24m km
This is a combat-zone salvage ship. It's designed to go into contested systems and hoover up mountains of wreckage while potentially coming under attack. Thus, it's heavily armored by commercial standards and has substantial cargo space (so as to avoid risking more fragile freighters). By tonnage it's only 7.75% drive and fuel. That's because it is going to spend the vast majority of its time either in port or sitting on top of wrecks. It doesn't need to be fast, it just needs to survive and carry a shitload of recovered materials and modules. If this were 84% drive and fuel, it would need to be accompanied by easily-destroyed freighters, be exceptionally vulnerable itself, and be markedly less effective at its purpose to boot.

I only need one or two of these to clean up after a major fleet action. If I designed them according to your hypothetically optimal distribution, I'd need four or six times as many ships, drastically increasing costs both material and managerial, as well as making each more likely to be lost if something hostile slips past the screening force.

So based on your description and the ship readout, I'd say that you want about a 1000 km/s ship with a pretty long range, about 70b km, right?

You can do that! My algorithm is perfectly fine with that. (You were addressing me, right?) I never said that speed was important. You plug in your speed, mass, and range, and the optimal distribution of engine space and fuel space is calculated for you. You could achieve that speed and distance with a giant, inefficient engine and a cup of fuel, or with a super-tiny, super-thirsty engine and an entire year's worth of fuel production. My algorithm finds the happy medium.

You could tell my algorithm you want a ship that can go 10'000 km/s and travel a few trillion km. But it's your own damn fault if it gives you a ship design that has a payload fraction of 0.1%!

My algorithm isn't a be all, end all. It isn't a panacea. It's not intelligent. It just shows you the best way to achieve the mission requirements. There is almost never any benefit to ignoring it more than a few notches - maybe x0.75 is better than x0.70 for some reason, but x1.40 would be insane.

Code: [Select]

Solemn Oath class Missile Cruiser    57,600 tons     1241 Crew     27142.55 BP      TCS 1152  TH 1382.4  EM 6000
10000 km/s     Armour 20-132     Shields 200-300     Sensors 24/24/0/0     Damage Control Rating 137     PPV 60
Maint Life 8.46 Years     MSP 31510    AFR 248%    IFR 3.4%    1YR 785    5YR 11773    Max Repair 1200 MSP
Intended Deployment Time: 120 months    Spare Berths 0   
Magazine 1408    Cryogenic Berths 200   

Adonai Drive Systems 960 EP Inertial Fusion Drive (12)    Power 960    Fuel Use 55.11%    Signature 115.2    Exp 15%
Fuel Capacity 15,000,000 Litres    Range 85.1 billion km   (98 days at full power)
Taurus Defence Industries Xi R300/360 Shields (40)   Total Fuel Cost  600 Litres per hour  (14,400 per day)

Barak Technology 52cm C6 Far X-Ray Laser (1)    Range 480,000km     TS: 10000 km/s     Power 71-6     RM 8    ROF 60        71 71 71 71 71 71 71 71 63 56
Endanor-Salvin Systems CIWS-320 (6x10)    Range 1000 km     TS: 32000 km/s     ROF 5       Base 50% To Hit
Valefar Incorporated Fire Control S06 240-24000 (1)    Max Range: 480,000 km   TS: 24000 km/s     98 96 94 92 90 88 85 83 81 79
Kalick-Endymion Magnetic Confinement Fusion Reactor Technology PB-1 (1)     Total Power Output 10    Armour 0    Exp 5%

Bale Armaments Company Size 4 Missile Launcher (50% Reduction) (22)    Missile Size 4    Rate of Fire 70
Eisenstein Technology Missile Fire Control FC460-R16 (2)     Range 460.8m km    Resolution 16
Javelin-4 III (352)  Speed: 40,000 km/s   End: 123.7m    Range: 296.8m km   WH: 16    Size: 4    TH: 386/232/116

Eisenstein Technology Active Search Sensor MR768-R16 (1)     GPS 12800     Range 768.0m km    Resolution 16
Sotha-Hezekiah Thermal Sensor TH1-24 (1)     Sensitivity 24     Detect Sig Strength 1000:  24m km
Sotha-Hezekiah EM Detection Sensor EM1-24 (1)     Sensitivity 24     Detect Sig Strength 1000:  24m km

ECCM-3 (2)         ECM 30
This is a military ship designed for limited independent action--long-term harassment of enemy operations, deep strikes into hostile territory, extended picket duty, hunter-killer operations on stray Precursor groups, periodically culling Swarm ships when I don't have the spare assets to kill the Queens, &c. It's much more drive-heavy; the drive+fuel component is slightly over 51% of its tonnage. That's primarily because it needs to be fast enough, despite its tonnage, to outrun just about anything that it can't outfight, with a few unfortunate exceptions.

This, too, could benefit from Dozebom's Algorithm. You want a ship that goes 10'000 km/s for 85b km? Then my algorithm shows you the way to allot the fuel/engines to get the most space in your ship.

What about a FAC, then?

Code: [Select]

Foehammer class Fast Attack Craft    1,000 tons     29 Crew     552.2 BP      TCS 20  TH 512  EM 0
25600 km/s     Armour 1-8     Shields 0-0     Sensors 1/1/0/0     Damage Control Rating 0     PPV 2.4
Maint Life 4.18 Years     MSP 173    AFR 16%    IFR 0.2%    1YR 16    5YR 238    Max Repair 256 MSP
Intended Deployment Time: 12 months    Spare Berths 1   
Magazine 16   

Tridon Aero Engines 512 EP Inertial Fusion Drive (1)    Power 512    Fuel Use 130.11%    Signature 512    Exp 20%
Fuel Capacity 165,000 Litres    Range 22.8 billion km   (10 days at full power)

Adramelech Design Bureau X/O Rail-4 (4)    Missile Size 4    Hangar Reload 30 minutes    MF Reload 5 hours
Eisenstein Technology Missile Fire Control FC460-R16 (1)     Range 460.8m km    Resolution 16
Javelin-4 III (4)  Speed: 40,000 km/s   End: 123.7m    Range: 296.8m km   WH: 16    Size: 4    TH: 386/232/116

Compact ECCM-1 (1)         ECM 20
Still barely under 57% drive+fuel. And that leaves only enough space for four box launchers, a fire control system sufficient to future-proof it for coming iterations of its missile series and help offset enemy ECM, and a basic EW package.

You want a 1'000 ton ship that goes 25'600 km/s for ~23 b km. My algorithm will find the optimized engine-fuel system for those requirements.

So yeah. That's my point. A ship that's 84% fuel and drive is going to be woefully ineffective at whatever it's supposed to do. The only circumstances where anything even approaching that would be practical are very limited: tugs, which are just drives strapped to tractor beams anyways; boarding shuttles, which are drives strapped to drop pods; and gauss/reduced-size laser fighters, which both need extreme speed and have a very low-volume combat load.

It doesn't help that there's generally a cutoff where increased speed stops being useful in combat. In the vast majority of cases no enemy you encounter will exceed 15,000km/s. Most won't break 10,000km/s. Beyond either of those points the only real effects are: decreasing enemy CTH (in which case your volume would be better devoted to a higher grade of ECM or more PD equipment), and allowing E/KW ships to close more rapidly (in which case they also need to have good armor, a decent helping of shields, and enough weapons to decisively end the fight once they reach their range--without the first two they won't live long enough, without the third there's no point using them). All you ever really need is "enough" speed. That is, enough speed that you can catch enemy ships and they can't catch you. At low TLs that's just flat-out not possible, and at high TLs it's trivial.

I'm wondering if you were responding to me now, because I don't remember saying that a ship should have less than a 17% payload. I think at least 40% should be the goal, even for low tech.

Anyway, I agree with this. You, the player, have to decide how fast you want each ship to go, and how far you want it to go. If you choose a fast, far-ranging ship, you might end up with a 10% payload, but that's bad strategy normally. My algorithm just helps you avoid the inefficient "1'000 ton engine and a thimbleful of fuel" designs, while not actually changing the speed or range.

===========================

I'd say that there are only two reasons to ignore Dozebom's Algorithm - RP and laziness. Otherwise, there's no reason to find the optimal engine-fuel system for your specifications.

Edit: And standardization and fuel-efficiency.

[Sheb]

Well, also you might not want to design and reserch a new engine for each ship class. Also, your algorithm give us optimal size use, but no optimal fuel use, right?

[da_nang]

Well, also you might not want to design and reserch a new engine for each ship class.

Standardization is a good idea, but the question is how. Obviously if all ships have the same tonnage, it's no problem assuming functional space requirement doesn't differ too much. Otherwise, you need to make a decision on how you plan to scale the engines to other ships.

The simplest idea is to design an engine for ships within a given tonnage range. For instance, say you a have a military fleet of 2500, 5000, 7500 and 10000 ton ships. You then design an engine for the greatest common denominator, the 2500 ton ship. You then scale the design up for the other ships by adding more engines. This maintains a constant speed across all ships, but you'll lose out on either range or space depending on how you plan to deal with the reduced fuel efficiency.

This technique would probably work best as long as GCD isn't too small and (largest tonnage)/(smallest tonnage) is a small enough integer (< 5-10) to limit the loss of fuel efficiency.

Also, your algorithm give us optimal size use, but no optimal fuel use, right?

My experience optimizing ships show that almost every time, globally space optimal is not globally mileage optimal unless your design parameters make a very narrow search space.

[Sheb]

Well, even if they have the same tonnage, the Dozebon algorythm would only give you the same engine tonnage for all ships if you always had the same range/speed requirement.

[da_nang]

Well, yes, if design parameters are equal then I'd imagine the optimal configuration would be identical. But all you need to do to throw off the design is introduce a large enough reserved space and the old design will turn out to be infeasible.

[Dozebôm Lolumzalìs]

Hmm, good point. I didn't think of the standardization problems. But then again, that's not a problem with my playstyle. I tend to set a standard for all my ships: 5'000 km/s and 20'000b km, for instance. This may change as tech increases, but it's the same for most military ships above FAC at any given time. I also tend to build all ships of the same "superclass" the same size: PD frigates, sensor frigates, etc. will all be 2'000 tons.

And while Dozebom's Algorithm may not be the most fuel-efficient, it tends to be okay. I'd say you shouldn't worry too much about fuel-efficiency over space-efficiency unless you have a giant fleet or are in a sorium crunch, as there are other minerals to worry about - including duranium. Space efficiency also matters for minerals.

What I meant was "if you're concerned with space rather than fuel efficiency, there's only two reasons to avoid D's A: laziness being unwilling to dedicate hours of time and years of researchers every time you research the next engine tech, and roleplay.

[da_nang]

Well, technically if you're roleplaying, what engineers wouldn't want to make efficient designs? Unless their civilization is comprised of idiots, or they simply don't give a fuck about good practices.

...

Probably involves drunk science.

[Flying Dice]

That's what I was getting at earlier, though. The physical laws (as such) in Aurora regarding T/N ships design have a lot of pretty clear points where they're objectively more effective or efficient or whatever. Because it's a game. In Aurora if you don't design every missile with a square number WH, you're deliberately reducing the efficiency of your missile fire. But in real life it's not that simple, you don't just say "Oh, X energy discharge is the ideal for a missile in this role, why would you ever use something different?" Modern AAMs don't even all use the same warhead type, never mind functionally identical warheads.

This is what I was talking about when I said that I deliberately avoid overoptimizing (which, again, people seem to ignore). From a storytelling perspective the way that a lot of Aurora's systems can be gamed for optimal design points is dull. Making the same ships according to the same design principles in every campaign is not interesting.

[Rince Wind]

Well, technically if you're roleplaying, what engineers wouldn't want to make efficient designs? Unless their civilization is comprised of idiots, or they simply don't give a fuck about good practices.

...

Probably involves drunk science.

(WH40K) Orks