Engineering system to complement Science-based tech tree

[NeilC]

Now that 0.22 is released, let's get the devs to notice this request for an EEP system! Please take a break from doing ALL THE SCIENCE to post your thoughts on this!

TL;DR:

These are nice-to-have perks for good Space Program design, and the general concept is:

- Use a part more often, that part gets cheaper

- Fly a booster more often, the booster gets cheaper

- You learn a lot with the first couple dozen parts, and afterwards not so much

- As you gain experience with a part you get ideas on how to improve it and can start building versions with higher cost but better performance

In more detail:

I'd like to point out that having to go get samples of Mun rocks in order to figure out how to build bigger and better rockets doesn't make much sense. It's not the rocks that teach you how to improve rocket tech, it's the fact that you had to build rockets to get to the Mun. Now, don't start flaming. I'm not arguing against the tech tree. I love the Science-based gameplay element, and I think "sample Mun rocks" is a much more compelling goal to unlock a tech tree node than "Fly 10 missions". Having a purpose for missions is good, a gradual progression from small to large rockets is good, I can see the science-based tech tree implementation being a perfect fit for KSP gameplay.

But I think KSP gameplay could be further improved by adding a much simpler, much less flashy Engineering subsystem that gives the player benefits for not just what their space program has accomplished (Science), but how they did it. This relates to lifter, payload, and overall mission design. Real space programs don't build custom lifters for each mission, or these days even custom boosters for each lifter (eg Delta IVH, Falcon 9, etc). I think players should be encouraged to re-use and develop their designs instead of building from scratch every time. Not that you would be limited from building from scratch if that's what you want - go nuts! But there should be a cost incentive to re-use

Note that we are mostly talking about engines so far, because they are the most central parts in KSP and are easy to define upgrades for. Please feel free to suggest appropriate upgrades to other parts.

How an Engineering system would work:

The system would focus on keeping track of what you have launched - let's call it "Engineering Experience Points" (EEP). Once you have unlocked a part, you gain 1 EEP towards that part for every one you launch. Build a rocket with 1 central LV-45 and 4x LV-45 boosters? That's 5 EEP for LV-45s from a single launch.

Other possibilities for earning EEP include:

- Inspecting parts in orbit after their operation, like taking a Kerbal EVA to take pictures of an upper stage engine after firing

- Recovering used parts, and presumably tearing them down for inspection

- Test firing engines on the pad

There are clearly concerns with balance and the potential to "farm" EEP from the last one - these must be addressed. Economics will play a strong role here, and test firing engines must incur significant expense for this system to work. If cost recovery is 100% for used craft, boomerdog2000 points out that it won't be long before people are launching 25-Nerva-powered rovers and recovering them immediately to gain the EEP. It might be important to define a "launch" more strictly for EEP purposes, and award points only for orbited parts.

In any case, this type of balancing would be done later on in development; the current focus is to get the core concepts of the EEP system fleshed out.

For each part:

Each EEP would marginally reduce the cost of the part. This wouldn't be huge or game-breaking: I'm thinking a maximum of 25% cost reduction after around 500 EEP or some other large number. The cost curve would be smooth, meaning that every EEP gives you a small benefit instead of having EEP thresholds. The curve would not be linear but pretty heavily front-end loaded: after 20ish EEP you would see a 10% cost reduction, it would take 100 EEP to get to 15%, 300 EEP to get to 20%, etc.

Also, at certain EEP thresholds you would unlock new versions of the part that would increase the cost but improve the performance. There are two ideas going for how the upgrades should be handled:

1) NASA-style "block upgrades", where the Block A/B/C/D/E designs are fixed and there are 3-5 of them.

NeilC's original intent was for upgrades that continuously improve the engine within the confines of its type: LV-T30s and Mainsails are first-stage/booster engines so their upgrades improve thrust and sea-level ISP, while LV-909s and Poodles are upper-stage engines so their upgrades would raise vacuum ISP and lower weight. These upgrades would stack, so each block improves on the previous one.

The basic cost structure for these block upgrades looks something like:

A costs $100

I develop B, A now costs $90, B costs $100.

I develop C, A now costs $80, B costs $90, C costs $100.

Now you have an interesting choice to make: do you build the good old Block A for cheaper, or pay for the fancy new Block C?

An alternative idea from Zaeo is to have each block upgrade be more suited to a different role so that you could specialize your engines to your current mission. In this case the block upgrades would not necessarily stack, and each one would be a more substantial bonus.

In either case, this would be implemented as a "tweakable" in the VAB interface, not a separate part - we don't want 5 copies of each engine in the GUI! For example, if you right-clicked an engine you might see a dialog that looks like a normal right-click dialog from the game, but with an added field something like:

LV-T30  B0     BA     BB     BC     BD
WEIGHT  1      1      1.05   1.05   1.15
THRUST  200    220    220    250    250
ISP     270    270    290    290    320
COST    200    220    240    260    280
       | SEL  | SEL  | SEL  | SEL  | SEL |  

|SEL| represents a clickable SELection button to pick which Block you want to build. Only the changing parameters are tabulated, all else is listed as normal.

2) A slider-based upgrade system that would let you freely customize parts within limits set by EEP.

This alternative, suggested by Boomerdog2000 here , is being actively discussed and needs further fleshing out.

For engines, we envision having sliders for thrust, weight, and ISP. They start out at stock values on the left, and each parameter can be increased to a max of stock +15-25% subject to balancing. Every increase adds cost to the part as well. You can move any of them freely, but the max total improvement is limited by your current EEP - spend it all in one place, spread it evenly, or keep building cheaper stock parts.

For each Subassembly:

This concept also applies really well to the new stock Subassembly system. Every time you launch a subassembly - lifter, booster, payload, whatever - you gain EEP with it and the cost starts to come down with the same sort of curve. the shape is up for debate, I would suggest a smaller (15%?) max benefit at less EEP (200 max?) but also less front-end loading - you learn less about welding things together with each part than you do about engine design.

This would only apply to assemblies that are re-used without changes. If you loaded and changed a subassembly, you would be prompted to either overwrite and lose all EEP or save as a new subassembly. The idea is to encourage re-use of common modules and improve your space program's efficiency by developing new things to do with existing hardware - like if NASA had actually done anything the Apollo Applications Program suggested.

Astropapi1 points out that it may be difficult to get the cost reductions to play nicely with the subassembly system when you are repeatedly adding/removing multiple subassemblies. More investigation is needed here once we get to play with the new system.

All these numbers are subject to balancing, of course, and my intent is these should not be overwhelming benefits that give you an incentive to stagnate your program on 1.25m engines. It should not ever cost less to build a 100-ton lifter with 40x 500EEP Block D LV-30s than it does to use 5 0EEP Mainsails you just unlocked.

Edited October 16, 2013 by NeilC

[J.Random]

So, basically, it's about reducing the price through mass production? Me likeses. How about EEP decay if parts aren't used for sufficiently long time?

[i8nemo]

An XP system for rockets?!?

Awesome, pratical, reflects real life engineering (sort of).

[boomerdog2000]

I like the idea but, the problem I could see with this is people building part "farms."

Launch something that is a platform on wheels filled with NERVAs then just roll it off the runway. We'd have to be very specific on what counts as a launch.

[BostLabs]

I like the idea but, the problem I could see with this is people building part "farms."

Launch something that is a platform on wheels filled with NERVAs then just roll it off the runway. We'd have to be very specific on what counts as a launch.

Or a location, like the Jool system.

Park next to the runway, get lower xp points.

Parking orbit around Jool, good xp points.

the Interstellar mod does something like this.

[JumpsterG]

I like the idea but, the problem I could see with this is people building part "farms."

Launch something that is a platform on wheels filled with NERVAs then just roll it off the runway. We'd have to be very specific on what counts as a launch.

I wouldn't really think this would be a problem, because every launch (and every part) will cost money. Sure, launching non-missions to farm experience would be possible, but would also have a price... In fact, firing a bunch of NERVAs on the launchpad sounds like engine testing which SHOULD result in experience gain. The tradeoff of player time and money for EEP might be fun for some people and remain optional for others since the benefits would presumably be minimal (and have to be carefully weighed against the immediate cost).

I like the idea quite a bit.

[Guest]

All rockets are tested on the ground before launches, so it makes sense. Also, truthfully the science system is anything but a science system. It's a xp system with "science points" taped over the xp. In a game based around discovery and invention we don't need a "it does everything science sauce" this encourages game breaking behavior, exploitation and makes less "science points" lucrative activities pointless. Actions you take should have impact on something relevant to that activity. If I go to Eve and collect blutonium it should give me a better understanding of atomic sciences(nervas) not add to a magical pool of science stuff. What we do as players should carry meaning and have impact; Science Points cannot, nor will they ever be capable of that.

[Rockstar04]

I really like the sound of this, and its is one of the best though out suggestions I have read in a while.

+1 for sure!

[NeilC]

I like the idea but, the problem I could see with this is people building part "farms."

Launch something that is a platform on wheels filled with NERVAs then just roll it off the runway. We'd have to be very specific on what counts as a launch.

Actually, I'm not sure this is a problem at all. Remember: one day, launches will cost money.

Test-firings of engines on the launch pad is a valid way to earn EEP even if the rocket doesn't go anywhere, just like in the real world. But doing 25 of them at once on a "part farm" would bankrupt your space program pretty quick. You'd pay full price 25x before getting the benefit of 25EEP for your next construction round.

Once KSP players have to pay for each launch and there's a requirement to deliver science for the money spent, players will choose to do real things with their launches because otherwise Kerbal Congress cuts their funding.

[geb]

Also added a system to award “recovery†data when recovering vessels that have been places. Makes sense that you’d learn something from reverse-engineering a vessel that didn’t explode.

^from the 1st of October KSP Weekly. It seems like there is already going to be a system for engineering experience, and it will award science points for successful missions.

[Monkeh]

I have typed this message, here on this forum thread using my keyboard, with the intent of expressing my feelings towards this particular suggestion...

*Ahem*

THESE ARE GREAT IDEAS!

I do hope Squad see this thread.

Thank you for your time.

[NeilC]

It seems like there is already going to be a system for engineering experience, and it will award science points for successful missions.

Sure, but it will just add Science points towards unlocking more parts. I don't think that's the best way to capture "learning something reverse-engineering a vessel that didn’t explode" - how is that the same type of knowledge as examining samples you've returned? What do Mun rocks teach you about rocket bells? Instead, I'm proposing a parallel system that's about making existing parts better and cheaper, instead of unlocking new ones.

(As I said in the OP, the whole concept of scanning- or sample-return-type science being the key to unlocking new engines is "interesting" in terms of realism. But it's a GREAT gameplay mechanic as the primary driver of the game, so I can suspend disbelief)

Anyway, to totally co-opt your point and put words in your mouth, I think getting an extra EEP or two from recovering parts that have been in orbit is a great idea!

[rkman]

Sure, but it will just add Science points towards unlocking more parts. I don't think that's the best way to capture "learning something reverse-engineering a vessel that didn’t explode" - how is that the same type of knowledge as examining samples you've returned? What do Mun rocks teach you about rocket bells?

Good point. It would be more appropriate if science points are the measure of success of a space program, and engineering would be the way to achieve that success.

On the other hand, there is some correlation between science prowess and engineering prowess. Though i have no idea how to capture that in KSP gameplay.

[TimMartland]

This idea is well thought out and would compliment the science system well - unlock parts through Research and Development, improve them in engineering. +1 definitely (hope someone from Squad sees this post)

[Rockhem]

This sounds like a great idea when they are refining career mode, using a part a lot should help with costs.

[boomerdog2000]

I wouldn't really think this would be a problem, because every launch (and every part) will cost money. Sure, launching non-missions to farm experience would be possible, but would also have a price... In fact, firing a bunch of NERVAs on the launchpad sounds like engine testing which SHOULD result in experience gain. The tradeoff of player time and money for EEP might be fun for some people and remain optional for others since the benefits would presumably be minimal (and have to be carefully weighed against the immediate cost).

I like the idea quite a bit.

I didn't mean put your rocket up to test it and get experience, that makes perfect sense to me. I meant more put 50 NERVAs on wheels to exploit the launch and get experience idea.

The money problem should help balance this but it will depend on how much you get back after recovering the vehicle.

[Lone K.]

What would happen if a spacecraft lands back on Kerbin safely? Would those parts that have been salvaged be, essentially, "free" parts? Like, let's say I recover a Fly-by spacecraft and it lands on Kerbin without harm (excluding boosters and the like, but I'll assume those are part of the sub-assembly topic you covered), would those parts of the spacecraft (that were recovered) act as resources, and as such you can use them again without cost?

[boomerdog2000]

I would think you should get something back but I'm not sure how much it costs to refurbish something to go into space again.

[NeilC]

What would happen if a spacecraft lands back on Kerbin safely? <...> would those parts of the spacecraft (that were recovered) act as resources, and as such you can use them again without cost?

This is definitely not the case in reality. All Shuttle components had to go through serious inspection and refurbishing before being flown again, and the cost wound up being much higher than anticipated in the original Shuttle proposal - higher than comparable expendable systems, in fact. They planned for one launch per week while pitching the program to Congress, and achieved an average of one per 3 months over the program's 30-year life span. This includes the long launch hiatuses after the Challenger and Columbia disasters, but in the later stages of the program each shuttle took 6 months to refurb and launch again. Even the minimum refurb time topped out at 2 months, and that was before Challenger caused a more cautious approach.

In KSP I don't think it's worth managing the equivalent of a shuttle fleet and paying attention to refurb cost vs construction cost. Instead I think there should be a cost refund for recovering craft of around 50% max, and potentially gaining additional EEP as well from inspecting the flown part. The recovered cost is another candidate to benefit from the EEP system: recovery percentage could increase the more experience you have with a part/subassembly.

[KvickFlygarn87]

i have typed this message, here on this forum thread using my keyboard, with the intent of expressing my feelings towards this particular suggestion...

*ahem*

these are great ideas!

i do hope squad see this thread.

Thank you for your time.

this. You are my thoughts.

[vetrox]

I like this idea.

You have my sword!

[Bessy]

love it

Devs .... i now have something extra i want in ksp ^^

[Capt Snuggler]

+1 love this idea!

Edited October 14, 2013 by Capt Snuggler

[ludsoe]

I didn't mean put your rocket up to test it and get experience, that makes perfect sense to me. I meant more put 50 NERVAs on wheels to exploit the launch and get experience idea.

The money problem should help balance this but it will depend on how much you get back after recovering the vehicle.

Detect if the vessel was in a valid orbit at least once. => Give Points if true.

At any rate you would most likely get points for the wheels instead since your testing those and not the engines themselves in that case.

[DChurchill]

Yep. Great idea, adds another consideration to the career game.

There should be some limits, though. Five blocks max or something and they can only get so cheap. Have each successive block make the previous ones cheaper, too.