[1.4.1] Kerbal Construction Time 1.4.0.69 (2018-03-24) - Unrapid Planned Assembly

[severedsolo]

Maybe I'll double the mass by default and reduce simulation costs next update. I've been feeling that simulations may be too costly to use frequently, especially in orbit around distant planets.

You think? I find simulations to be very reasonable (admittedly I only use them to test launchers 99% of the time) but I don't think 600(ish) for a launcher simulation is bad.

[Nori]

You think? I find simulations to be very reasonable (admittedly I only use them to test launchers 99% of the time) but I don't think 600(ish) for a launcher simulation is bad.

Yeah for launch testing they are fine, but try a couple day long sim on Duna. Shouldn't be free but it is pretty expensive.

[magico13]

You think? I find simulations to be very reasonable (admittedly I only use them to test launchers 99% of the time) but I don't think 600(ish) for a launcher simulation is bad.

Launches from Kerbin I think are fine, but ones in orbit (especially around Laythe. Laythe costs more than Eeloo) might be a bit too high to use frequently. I'd be curious as to what others think. The trade off with them is that you typically aren't simulating an entire rocket (launcher+payload) but instead just the payload when you're starting in orbit (or soon-ish, landed) so the total cost goes down. I'm mostly thinking of reducing the distance modifier and the atmosphere modifier, especially if I'll be adding another modifier for landed simulations. I could also make the time modifier be an actual function instead of arbitrary so you can enter whatever time limits you want.

[Nori]

Configurable multipliers would be nice. But don't want to make too much work for you.

My main issue with it currently is more that each simulation costs money. If I'm working on a complex lander I might need to do 5 simulations. Over Eve that could add up to a ton of money.

Given that maybe leave the cost as is but set it up that if a simulation has been done to that planet in the last, say day, then make it cost significantly less.

Though that might be harder to implement so I suppose just reducing the multiplier for other planets would be nice.

[NotExactlyRocketDesign]

Another interesting way to get around the annoyance of repeat simulation costs would be to have simulations on or around particular bodies be "unlocked" with a fixed lump sum of funds. For example, if you buy the "Tylo orbit" or "Eve landed" simulator for X00,000 funds, then you can henceforth simulate vessels orbiting Tylo or landed on Eve for free (or cheap, if free seems too unbalancing).

[DopeGhoti]

I like the idea of being able to simulate places you have been in orbit of. Once you orbit Mun, you get the "Simulate at Mun" options, for example.

[shand]

perhaps a set cost without distance multipler - remember that this is simulating, not actually sending it there (although in game it is but shhhh), perhaps to add to the "cost" make it so you can only run tests of what you know. so if you havent been in any atmosphere it won't model any, or the sim finishes if you enter it or something. and if you havent been on the land you cant simulate on the land. i imagine the simulation ending is the easiest for you to code.

this is just my personal views. i find the "launchpad" test slot too cheap, less than the cost of a parachute, while the orbit of min ones probably the higher end of what you'd want.

EDIT: sidenote. which way does the "reconditioning effect" go, larger number larger effect therefore larger time? thats what i'm assuming?

EDIT2: yeah i found the ingame settings for it *whistles innocently*

Edited October 14, 2014 by shand

[magico13]

Maybe I'll save this for a later date and do a larger overhaul of simulations where I make the modifiers customizable and do some other stuff. Maybe let you buy "packages" or whatever to reduce the cost of simulating at a specific place (costs break even after 10 moderately sized sims or something along those lines).

I like the distance multiplier because it makes it so not everything is the same. And how its implemented it will work with any mods that add extra planets. I think orbital sims just need to be brought down in cost pretty substantially. I do kind of like the idea that subsequent sims shouldn't be as costly, so maybe if you revert to the editor any sims at the same place don't cost as much until you leave the editor. That way you can make small changes as necessary without blowing the bank.

Just thought of a way to make it so you couldn't just simulate a command pod and then get the discount on your mega-awesome expensive ship. Keep track of the total funds spent on simulations that session and once it gets above, say, 2 times the cost of the most expensive one you don't have to pay any more. Otherwise you pay the total cost as normal, or the difference of the max amount and the current sim cost, whichever is less.

Ex: Sim costs 10k funds. Max amount then is 20k. You simulate that, then go make some changes. New sim costs 11k funds so the max goes to 22k. Total is now 21k spent. You make some small changes and the newest sim will be 10k funds again. Max is still 22k, but you only have to pay 1k. Future sims, as long as they're below 11k initially are free. If one is, say, 12k the max goes to 24k and you pay 2k.

I had originally considered ending simulations when you entered places you hadn't been before (space if launching from Kerbin early on, atmosphere/landed of other planets/moons) but decided against it for simplicity. I do want to require that you actually have landed on another planet/moon before you can start simulations there though. I haven't decided between letting you input any landed coordinates you want, or requiring you to start near a flag/probe/base that currently exists. At the very least there will be a list of those objects you can choose from and it will pull their coordinates and input them automatically.

I will add a maximum amount of time for reconditioning. Maybe like 7 Kerbin days before the Overall Multiplier gets multiplied in. Alternatively I could set it so it can't have a BP higher than the vessel's.

@DopeGhoti Simulations in orbit around bodies you have visited is already implemented. Just choose a different Body in the simulation configuration window

@shand Glad you figured out the setting for reconditioning. The one in the file is BP per ton at an Overall Multiplier of 1. For whatever reason it made more sense to me to save that number (since that's what's actually used in the formulas) rather than the mass (which is what can be changed in game).

[Noventta]

I think the simulation costs should be separated effectively in two pieces forgeting about the distance modifier in favor of a preset complexity parameter as it doesn't make sense that a simulation costs more just because it is harder to get there. The two charges should be:

1. Enviroment Modeling (once): You need to create a model of the enviroment to run a simulation, whichever the vessel is (fixed price depending on complexity).

2. Vessel Simulation (for each simulation): Depending on weight (or other criteria) the simulation has a base cost multiplied by complexity of the simulation.

The cost of Enviroment Modeling should vary with complexity in mind:

Orbit simulations can be modeled without ever been there from the begining of the game just based on the size and properties observed from Kerbin. Nevertheless, Jool should be more expensive to run a simulation than Eve or Duna and those more expensive than Dress or Moho because the moons add complexity (pretended of course) to the system. All these should be relatively cheap as you can calculate an entire tour with manouver nodes.

Atmospheres on the other hand are very complex to model, so that, are very expensive and cannot be modeled without aquiring data, A.K.A. having been there. A good reason to send a probe ahead. I don't know how bodies without athmosphere are treated by the game, maybe the time warp limit can be used for those at a lower price than atmospheres.

The atmospheric/surface model could be "purchased" from the upgrades menu (or similar) or charged on the first simulation. Simulation ending (or aditional charge message) upon entering athmosphere would be needed of course.

Summary:

1. You can simulate KSC launches and orbits as it is now from the begining although orbit simulations should be cheaper than launch simulations.

2. You can simulate orbits arround other bodies after paying a small amount to first create the model. These should be relatively cheap.

3. To simulate an athmospheric flight or landing first you need to visit the body to aquire the data and pay a relatively big amount of cash for the model. These simulations are more expensive but not too much.

I know it may sound complex but I think it ends being less complicated than what you said.

[hjrrockies]

I don't mind the high simulation cost, simply because I think it reflects a realistic cost for designing a mission far outside of Kerbin SOI.

[Noventta]

I don't mind the high simulation cost, simply because I think it reflects a realistic cost for designing a mission far outside of Kerbin SOI.

The thing is you are not simulating a whole mission to simulate an orbit arround Jool. If you want to simulate a whole mission you have to pay for the infinite time simulation and the various SOI and then you have a high cost. At least I see it that way.

[jwest]

The thing is you are not simulating a whole mission to simulate an orbit arround Jool. If you want to simulate a whole mission you have to pay for the infinite time simulation and the various SOI and then you have a high cost. At least I see it that way.

I see it the same way you do, Noventta. Taking it one step too far, I'd like to see simulation costs for similar vessels approach 0. I tend to be very iterative in my design process, in a very rapid tweak-simulate-tweak cycle. In my head, the simulations would also be fairly iterative. They wouldn't start from scratch just because some absent-minded engineer forgot to set the action groups...again. I wouldn't expect them to ever actually reach 0, because there's still a cost to setting the simulation up and running it again. Not having spending the full cost might be nice, though.

But that's pie-in-the-sky wishful thinking. Even with the costs as they are, I love this mod. It adds so much to the game. Thank you, magico13!

[Noventta]

Taking it one step too far, I'd like to see simulation costs for similar vessels approach 0. I tend to be very iterative in my design process, in a very rapid tweak-simulate-tweak cycle.

We mostly disagree at that point as that way you are getting simulation time for free. The less iterations you need to finish your rocket, the more efficient the process is.

[Matrix Aran]

One very specific case of the simulation system that I think could be addressed is regarding simulations starting in low Kerbin orbit.

Say I want to test my relatively small spaceplane on re-entry. I have two choices. I can either buy 1 hour of simulation time that starts landed on Kerbin, fly up to orbit, test, fail, lather, rinse, repeat; or I can start the simulation already in orbit and simply make the decent burn, saving about thirty minutes of flight time, give or take.

However, the cost to start in orbit is very much inflated considering the short amount of time it takes to manually fly a mission up there.

My spaceplane has a base cost of 182.288 to launch on the ground for 15 minutes. 1 hours costs 2x that, or 364.576.

To start in orbit, for 15 minutes of flight costs 2278.6 funds.

For the same expense, I could buy 14600 days of flight time if launching from Kerbin's surface and still have some pocket change left over. Or make 12 simulations covering the same test for the same price.

While I appreciate the intent is to make it an added expense for the sake of convenience, I think the modifier for low Kerbin orbit could do with some tweaking. A 12x increase in the cost just to save 30 minutes of sim time or 15 minutes of real time is a touch unreasonable in my mind.

[jwest]

We mostly disagree at that point as that way you are getting simulation time for free. The less iterations you need to finish your rocket, the more efficient the process is.

You're right. Let me try this: what I really want is a larger cost for the initial set-up of the simulation, even if the cost of the subsequent simulations remain the same. My gut reaction to the first simulation of a craft and the simulation of the same craft with a negligible tweak costing the same amount is that it isn't terribly realistic. The bulk of the cost in effort should have been paid already in setting up the initial simulation. Revision 2.01 of the craft isn't a complete do-over. So even if the cost for revision 2.01 remains what it is today, I'd like revision 2.00 to have cost maybe ten times that much for the initial set-up of the simulation.

But this model feels like you're outsourcing the simulations to a third party instead of doing them in-house. It's an excellent starting position. I could see a path for progression so that one of the upgrades is buying a simulation facility and paying full-time staff to work on simulations. You might pay a big cost up-front to get the simulation facility up and running, then recurring costs to keep it staffed. Upgrades could apply, too, to get simulations done faster or have more being worked on at a time, following the same model as the VAB and SPH. In the long term, it might save you money compared to outsourcing the simulations.

[DivisionByZero]

How about make this more game-play related. Scale the simulation cost for starting in orbit inversely with the amount of science completed in a given SOI.

All you've done is visited the SOI? Huge cost multiplier.

have all the crew reports and temperature scans? much smaller cost multiplier.

Have everything? make it very cheap.

This, I think, would make for some interesting gameplay decisions between the cost of sending up simple probes to get a couple more science scans vs. full-on simulation for more complex missions.

Just a thought.

[tygoo7]

Will this mod work with FMRS and not use automatic recovery like StageRecovery? I'm thinking of getting it but I like FMRS.

[magico13]

Will this mod work with FMRS and not use automatic recovery like StageRecovery? I'm thinking of getting it but I like FMRS.

You know, I'm not positive what will happen, but it likely will do its recovery and let you do stuff with FMRS, which isn't really balanced. I will add an option in next version to let you disable the automatic recovery aspect.

As for everyone else, I'm listening to what you've got to say about the simulations. Keep the ideas flowing. I haven't picked anything in particular yet, though I'm thinking that update is going to have to be a separate one at a later date. Hoping to get some dev time on Thursday to finish up some stuff for the 0.25 update so that can get officially released.

For those of you new to the party, the current official version doesn't work on 0.25, but there is a development beta that is mostly working a few pages back. I'll put a link in the OP.

[OverloadUT]

For those of you new to the party, the current official version doesn't work on 0.25, but there is a development beta that is mostly working a few pages back. I'll put a link in the OP.

Yessss thank you! Spent hours today getting my new .25 career all set up, and KCT is the only one missing now

[Hyomoto]

I'm surprised I missed this mod, and I look forwards to getting to try it out when the 0.25 version is stable. The contract times will finally make sense, and this seems to be very much like the system I've been preaching since we first got contracts. Time is a under-utilized resource and it's nice to see someone try to actually work it into the game strategy. Granted, I've only read over the first post so I'll have to reserve judgement until I can try it proper but you've piqued my interest to say the least.

[soulsource]

As for everyone else, I'm listening to what you've got to say about the simulations. Keep the ideas flowing.

This is just an idea, and probably not a very good one, but you could consider linking the availability or price of simulations somehow to the difficulty settings, more specifically the options to enable/disable reverting and quicksaves. Although it's not realistic, from a gameplay point of view it might make sense that simulations on a save where reverting isn't enabled are (much?) more expensive than for a save where the player could (on stock KSP) simply press the revert to VAB button if something goes wrong.

I guess I'm not the only one who started a "hard" difficulty game (no reverting), but later on at some point could not resist the temptation of the simulate button, just to see if FAR aerodynamics would kill an expensive craft...

[hatterson]

This is just an idea, and probably not a very good one, but you could consider linking the availability or price of simulations somehow to the difficulty settings, more specifically the options to enable/disable reverting and quicksaves. Although it's not realistic, from a gameplay point of view it might make sense that simulations on a save where reverting isn't enabled are (much?) more expensive than for a save where the player could (on stock KSP) simply press the revert to VAB button if something goes wrong.

I guess I'm not the only one who started a "hard" difficulty game (no reverting), but later on at some point could not resist the temptation of the simulate button, just to see if FAR aerodynamics would kill an expensive craft...

That's kinda the point of simulations though, at least IMO. You need to be able to test if a craft will work prior to spending days/weeks/months building it.

I'd rather see simulation price/scale factor included as a setting. That way if you're really intent on not really using simulations you could jack up the cost and be forced to fork out thousands of credits.

Edit: In addition to adding a scaling setting, I like the ideas Noventta posted a couple pages back about having all orbits unlocked from start, but needing to visit atmospheres (ideally do some kind of scan/science there to "learn" about it) before you can unlock that. Although it would mean you shouldn't be allowed to start a sim in orbit of Eve and then just deorbit yourself into the atmosphere, the sim should stop at that point.

If you wanted to take it even a step further (I love thinking of ideas for other people to work on ) you could say that atmospheric flight requires having been there and landing requires having been to a location in the vicinity or having scanned it (possible tie in with scansat or similar). Want to sim a Moho landing? Better have a scan of the surface, otherwise your sim engine won't know what it's doing.

So you'd have orbits unlocked by the fact that you have giant dishes at KSC, atmospheres unlocked by being there (either manned or unmanned), and landings unlocked by either already landing within x kilometers or having done a surface scan of the area. If you enter an area you haven't unlocked (either enter the atmosphere or get with y kilometers of the surface) you simulation is ended.

Edited October 15, 2014 by hatterson

[marchingknight11]

Is it possible to begin a simulation on the surface of another body? Whenever I try, it jumps me to the planet and Immediately explode

[magico13]

Is it possible to begin a simulation on the surface of another body? Whenever I try, it jumps me to the planet and Immediately explode

Not yet. I've got a very work in progress version of it in my dev version, but haven't made that public yet (because it frankly doesn't work yet). If you put in 0 for your altitude, it will actually try to put you into a stable orbit at an altitude of 0, which is definitely not a safe altitude You either need to land a ship from orbit or wait until I get that out, which might be a while since I have to rewrite half the simulation code due to the above discussion

[forsaken1111]

I'd actually like to see simulations have a large 'up front' cost and then have each individual simulation have a lower runtime cost. This is analogous to buying the hardware for the computer simulations and then paying maintenance and power to run them.

You could even have it scale with runtime, range, etc. Cheap hardware may only support simulations of ~30s and only within the kerbin SOI, with various upgrades to unlock more runtime and range or to reduce running costs.