Science gains from failure

[Matt77]

In KSP when I crash a ship, I'll often just reload or revert.  But first I check the after-action report to see if it contains any useful info, such as which part failed first etc.

In real life they investigate what went wrong, and learn from it - gaining Science^TM.  Real space programs are littered with accidents and disasters that led to improvements later on.

So why not gain some science to balance the loss of rep, funds - and sadly - brave, experienced Kerbals.

It would give a bit more motivation to allow the failure to happen (rather than reverting) and would also make impactors a valid probe type.  Perhaps it could go so far as gaining science by recovering parts from the ocean floor.

Possibly a bit morbid for stock?

[Rocket In My Pocket]

Could make a good strategy for the Admin center, something like:

"Crash Investigation Committee" Paying Kerbals to find out how other Kerbals messed up isn't cheap, but every once in awhile we learn something important from it.

-Reduces funds gained from recovered vessels by "X%" and converts it into "X%" Science.

[Matt77]

I like the wording!

Strategies would make it nicely optional, though I'd like it to be more developed than that. It's just an idea I had though so not sure how exactly.

 

[nikokespprfan]

I fully agree with this. Feedback is a very valuable -and sometimes invaluable- way of learning, especially in KSP.

[Gaarst]

Getting science from crashes alone doesn't really make sense. I think a Test Flight style thing with parts failures would be better: getting information from crashed crafts (recovered or not ?) can make your parts more reliable.

[Workable Goblin]

19 minutes ago, Gaarst said:

Getting science from crashes alone doesn't really make sense.

Sure it makes sense, if you think about what science in the game actually is: currency to buy new parts with. When you have a dramatic failure, that naturally spurs people to look at making new things that won't fail in the same way, and it lets them learn something about how (not) to design things, or in other words helps them develop new parts.

Edited January 23, 2016 by Workable Goblin

[Frybert]

Well there is the personal science you get from a crash. Now YOU know what not to do.

 

Edit: That was weird my text size changed after I hit save.

Edited January 23, 2016 by Frybert

[pandaman]

I like this concept.  On recovery of debris you get the option to get funds or less funds and some science, simple yet sensible and easy to do.

It could also be used as a variant on the rescue contracts.  Fetch 'this' debris from 'wherever' and get X science bonus on it's recovery back on Kerbin.

[BloodDusk]

This makes total sense. Science not only happens when things are successful, but also when stuff goes wrong.

[Matt77]

2 hours ago, pandaman said:

I like this concept.  On recovery of debris you get the option to get funds or less funds and some science, simple yet sensible and easy to do.

It could also be used as a variant on the rescue contracts.  Fetch 'this' debris from 'wherever' and get X science bonus on it's recovery back on Kerbin.

If it went the simple route it could just be an extra button -

Recover Vessel - Terminate Vessel - Investigate Debris

The latter option would trade recovery funds for science.  I prefer the more detailed approach but this idea seems more SQUAD, very simple.

[mattinoz]

If this existed it would be the first/only strategy I would employ. 

[AbacusWizard]

11 hours ago, Matt77 said:

Perhaps it could go so far as gaining science by recovering parts from the ocean floor.

Now I'm picturing a klaw-equipped submarine picking up a 50-year-old command pod from the ocean floor in the arctic regions, and then discovering that the pilot's body inside is frozen intact... and still showing faint signs of life.

 

-= CAPTAIN KERMERICA =-

[razark]

The only problem I see with this is the question of how you determine what is a failure?

[Guest]

I like this idea. Just one thing however, what would keep devious people like me from exploiting this? What type of mechanic should be included to discourage me from just crashing my ships on purpose to gain science points?

[razark]

2 minutes ago, Otis said:

I like this idea. Just one thing however, what would keep devious people like me from exploiting this? What type of mechanic should be included to discourage me from just crashing my ships on purpose to gain science points?

Decreasing returns on crashes.  You crash a part, you learn.  You crash it again, you learn less.  Wash, rinse, repeat. You crash the part a 10th or 20th, or 100th time, and you learn nothing.

[nikokespprfan]

13 hours ago, Gaarst said:

Getting science from crashes alone doesn't really make sense. I think a Test Flight style thing with parts failures would be better: getting information from crashed crafts (recovered or not ?) can make your parts more reliable.

Then you'd need to combine it with the idea of having failures in experimental parts idea from a few weeks ago. That could be an interesting one. Although I like the funds vs science tradeoff at recovery better.

[pandaman]

5 hours ago, Otis said:

I like this idea. Just one thing however, what would keep devious people like me from exploiting this? What type of mechanic should be included to discourage me from just crashing my ships on purpose to gain science points?

Interesting point.  Naturally the rewards should be low enough to make this an unattractive strategy, but I suppose it's one of those tricky 'balance' problems where there is no overall 'best' solution. 

However...  The game knows which ship the various parts of debris come from. So it could possibly assign a science limit per ship based on  original size/cost/part count or whatever maybe even adjusted on how long it's been in flight and where it's been. So If you crash your science research mission on lift off you get very little reward, but if it crashes on return from Eeloo then you may get to salvage more data. 

Though I guess it would be simpler to have a cross the board per part method where you get (say) 20%  of the funds and 1 science point per 1000 funds original cost, so more expensive, and therefore usually more complex parts, yield more information. 

[Red Cloud]

I like the concept. I don't like my habit of resetting the flight to launch every time if fails, but that seems "the right thing to do". This might encourage me to "live with failure".

I may try never reverting to launch to see how it plays out...

[Andem]

On 1/23/2016 at 1:18 PM, Gaarst said:

Getting science from crashes alone doesn't really make sense. I think a Test Flight style thing with parts failures would be better: getting information from crashed crafts (recovered or not ?) can make your parts more reliable.

If the vessel isn't recovered, it makes little sense. but they figured out that the foam on the SRBs had enough force to punch trough the shuttle wings because they recovered the Challenger debris. They wouldn't have thought about that if Challenger or some other crash hadn't happened.

Edited January 29, 2016 by Andem

[Moesly_Armlis]

This idea is an excellent method to wean new players away from using the revert crutch.

Edited January 30, 2016 by MoeslyArmlis