[Career] Experimental parts should have a chance to fail

[Rdivine]

This suggestion is mainly for career only.

There are contracts that are currently in the game rewarding you with funds if you test a part at ___ alt and at ___ speed. However, we can accept that part contract, and be able to use that experimental part indefinitely until that contract expires (sort of a way to cheat).

New suggestion: Make that experimental part have a chance to fail. Let's say 20%. When it fails during a mission, the contract will be considered as "complete". Then, when another contract regarding the same experimental part comes up again, the chance of it failing is lowered, let's say 15%.

Additionally, when the chance of failure reaches below 5%, the player has an option to purchase the blueprints for that part using funds.

Edited June 8, 2016 by Rdivine

[Ser]

Yes, that's a bit of exploit when we are able to "borrow" a part for free and use it without any limitations. I've suggested to make experimental parts highly unreliable in DangIt mod some time ago, don't know if they listened. What you are suggesting is some kind of mix of this idea and Test Flight mod's functionality.

So IMHO it would be enough to make experimental parts unreliable and complete its test-part contract regardless has it failed or not - you tested it, you get money.

5 hours ago, Rdivine said:

Additionally, when the chance of failure reaches below 5%, the player has an option to purchase the blueprints for that part using funds.

Up to this I'm agreed with you, but didn't quite understood what you mean here. Should a player not be able to purchase a part that have been tested once until he completes a certain amount of test contracts for this part? Don't think that's a good idea at all. Or maybe you suggest to give player a bonus to purchase the part without researching corresponding technology? Can't agree with that too. I think a better solution would be to give experimental parts a very high chance of failure, say 70% and gradually decrease it with every test contract completed to approach a still high failure chance, say 30-40%, thus preserving the idea that experimental prototypes are suitable for experiments only until you properly invent them by researching a corresponding technology.

[magnemoe]

its always nice to be asked to test the 48-7S early as its an perfect satellite engine but I think it has been moved down the tree.

[cantab]

I mooted a similar idea myself back when part tests were new. But rather than just outright failure, what I advocate is some "fuzziness" in the performance. The experimental part might perform nominally, or it might underperform, or it could even *overperform*. Or some mixture, like too much thrust and not enough Isp.

[cami]

(removed by author)

Edited June 8, 2016 by cami
wrong topic

[Rocket In My Pocket]

Another solution that would be relatively simple would be that you only get 1 copy of said part.

Ie. If you place the "test" part on a craft and launch it, you can't place that "test" part on any other craft until you recover the original craft along with the "test" part. If the craft and/or the part is destroyed then the contract to test said part is failed. This would make using the "test" part on an actual craft possible but risky.

To me it doesn't make sense that you'd get an infinite number of "test" parts to test. (Although now that I think of it, it could already work this way and I just don't know. I don't usually do the test X part contracts nor do I exploit them to gain advanced parts early. Not that I'm judging anyone who does, it's a clever way to do things certainly.)

[sal_vager]

It's a long standing design decision in KSP that the player should never be punished by a random number generator, if your Eeloo mission fails it should be by something the player can avoid through better craft design or piloting.

[Streetwind]

Well, the player COULD in fact avoid it, sal_vager. Namely by not putting experimental parts from a part test contract onto their Eeloo mission

I do like part of the suggestion. Not the incremental contracts to test the same part as a means to unlock it - that would be too repetitive - and I don't understand the part about getting an option to purchase the part (isn't this what unlocking tech nodes already does?). But definitely the idea that "experimental" parts should be something more distinguished than free, zero-downsides access to parts you haven't even researched.

[Stoney3K]

54 minutes ago, Streetwind said:

Well, the player COULD in fact avoid it, sal_vager. Namely by not putting experimental parts from a part test contract onto their Eeloo mission

I do like part of the suggestion. Not the incremental contracts to test the same part as a means to unlock it - that would be too repetitive - and I don't understand the part about getting an option to purchase the part (isn't this what unlocking tech nodes already does?). But definitely the idea that "experimental" parts should be something more distinguished than free, zero-downsides access to parts you haven't even researched.

Exactly that: Don't rely on experimental parts to accomplish a mission.

I kind of like the proposed dynamic: It gives players an incentive to do the (otherwise) boring, grindy test contracts because they get a chance to "earn" new parts for free or for a bargain price, without having to do all the science to get them. I'd like to add some test conditions, like not just "staging" an experimental engine to complete a contract, but have it run for X seconds and require an attached antenna, so the test crew can get reliable telemetry from the test.

Recovering of a test part to complete the contract is not always a good idea, how are you supposed to recover a part that you just flung into an escape trajectory from the Sun?

It would also discourage players from using that experimental part outside of the testing mission because it is prone to failing.

[John FX]

2 hours ago, sal_vager said:

It's a long standing design decision in KSP that the player should never be punished by a random number generator, if your Eeloo mission fails it should be by something the player can avoid through better craft design or piloting.

Just because a decision is long standing that does not mean it is the best decision for the future of the game. IMO one definition of a better design would be to not use experimental parts in it.

Myself I really like this idea. I tend to use this type of contract purely as a way to get parts for free I have not yet unlocked in a career game. I think it`s only sensible to have a penalty for exploiting contracts that way.

It would mean that experimental parts are discouraged from being used in critical missions, as is right.

[sal_vager]

Well if you would like random failures there's some rather good mods you might be interested in

[John FX]

5 hours ago, sal_vager said:

Well if you would like random failures there's some rather good mods you might be interested in

I currently play using testflight and enjoy it tremendously, it brings a real sense of tension with every launch until engines become reliable enough to trust for manned flights. I would not play with dangit as I don`t want totally random failures (I too do not want a random number generator deciding if my mission fails or not) but ones that depend on how much testing I have done, my gameplay and design. The two failure mods seem on either side of the `failure due to random number generator` line. This suggestion seems to be more like testflight in that if you test an unreliable experimental part it then becomes reliable, that`s not failure by random number generator. As the OP says, it would give purpose to testing parts.

Like FAR and DRE used to be "Well there is a mod if you want that in your game" and now they are stock and widely enjoyed so would this suggestion for making experimental parts less reliable IMHO.

[Holo]

18 hours ago, sal_vager said:

Well if you would like random failures there's some rather good mods you might be interested in

But I don't want random failures. I want a reason to not use experimental parts in ordinary rockets that isn't arbitrary or contrived.

[sal_vager]

2 hours ago, Holo said:

But I don't want random failures. I want a reason to not use experimental parts in ordinary rockets that isn't arbitrary or contrived.

Well I'd prefer for experimental parts to have altered statistics instead, you might not want to use a shiny new Vector that had half the thrust, or trust a turbocharged LV-N not to overheat.

This would add design considerations when testing these parts, instead of a random 'Oops you rolled a critical failure, say goodbye to your Kerbals!" situation.

[pincushionman]

5 hours ago, sal_vager said:

Well I'd prefer for experimental parts to have altered statistics instead, you might not want to use a shiny new Vector that had half the thrust, or trust a turbocharged LV-N not to overheat.

This would add design considerations when testing these parts, instead of a random 'Oops you rolled a critical failure, say goodbye to your Kerbals!" situation.

What also might help is more involved testing conditions besides "activate the part through the staging menu." Things like "run at 45 to 60 kN thrust continuously for 300 seconds" or "cycle the [gear] part from stowed to deployed 15 times with suspension unlocked, then lock suspension in the deployed condition and visually inspect via EVA" type stuff.

[nikokespprfan]

On 9-6-2016 at 10:16 AM, sal_vager said:

It's a long standing design decision in KSP that the player should never be punished by a random number generator, if your Eeloo mission fails it should be by something the player can avoid through better craft design or piloting.

That is maybe where the fuzziness that cantab suggested can come in, you never get outright failure, you never get punished, but there might be some things for which it is a good idea to put some redundancies on your craft. 

I know that you want KSP to always work reliably, but players expect experimental parts to at least have something funky. (At least I do.)

Spoiler

Maybe you could add a little description to say what is not guaranteed to work yet like,

Engine: This is an experimental part, the thrust figures might not be accurate.

 

[juanml82]

The problem with part failures is what can you do about it. A full part failure, as in, you put experimental aerospikes in your Tylo lander and when you finally make it there, you land but then you can't get back up... what are you going to do? Field repairs of such a scale would stretch disbelief. And if you can't do anything about it (and worse, if the failure keeps you locked in a solar escape trajectory, for instance), there is no gameplay value.

Getting strange performance (less thrust, more overheating, for instance) could be a way to discourage using experimental parts, without making it so you get a "Oops... you lost the randomness lottery, your mission failed, your kerbals died and there is nothing you can do about it"

[pincushionman]

If this is implemented, then it isn't a case of the RNG screwing you - They are not intended to be free access to higher technology than would otherwise be allowed; the tests are intended to be the mission itself, and the parts are therefore the payload. The fact that you can use them on ordinary missions can be exploited, and the OP is suggusting this as a way to close that loophole in a way that may be interesting.

If you're counting on experimental parts for your Tylo mission to be successful, you'd need to accept that risk. If you need reliable parts, on the other hand, a method already exists to acquire them - tech tree investment.

If these suggestions are implemented, however, then part test contracts need to be rebalanced to be worth the effort of dedicated missions, rather than piggybacks.

[Rdivine]

6 hours ago, pincushionman said:

If this is implemented, then it isn't a case of the RNG screwing you - They are not intended to be free access to higher technology than would otherwise be allowed; the tests are intended to be the mission itself, and the parts are therefore the payload. The fact that you can use them on ordinary missions can be exploited, and the OP is suggusting this as a way to close that loophole in a way that may be interesting.

If you're counting on experimental parts for your Tylo mission to be successful, you'd need to accept that risk. If you need reliable parts, on the other hand, a method already exists to acquire them - tech tree investment.

If these suggestions are implemented, however, then part test contracts need to be rebalanced to be worth the effort of dedicated missions, rather than piggybacks.

Exactly, i agree with your point.

Experimental parts should be part of the payload, not part of the rocket. It will only potentially fail if you use it(e.g. engines), not if you stow it in the payload bay.

Additionally, i would suggest that contract missions that require testing of large parts(srbs,mainsails etc) in space/escape trajectory have their rewards increased, due to the increased difficulty in bringing them to space.

Or, we could remove testing in space of large engines altogether and have contract missions that require dry-firing of engines (e.g. burning of > 30000 units of liquid fuel at launchpad, or launching the engine 1000m into the air and recovering it etc.).