Any single "failed" link between parts anywhere on the rocket, that's not part of a staging event, will make the game say "oh your rocket exploded, let me just take control of it away from you"

[SciMan]

I'm not sure if this was put in intentionally, or it's something not working right, but either way this is a horrible way to tell the player that their rocket broke.

I'm not given the option to continue my flight, meanwhile all the thrusters on my rocket shut off and the camera freezes in place and my rocket sails off into the distance.

So, on my large rockets (over 750 tons launch mass), guess what happens. I reinforce the rocket's core with a mesh of struts placed on short truss segments so that it shouldn't fall apart.

And that's where the trouble starts. It shouldn't fall apart right? Wrong, game says "nope this breaks no matter how many struts you add" I've tried adding literally hundreds of struts on the same dang joint that reliably breaks when launching my rocket (not too far into ascent, but after Max Q), and yet.... it just breaks.

So, on large rockets (the kind you'd need to build to go interstellar or even interplanetary), the struts are effectively useless.

Why are they useless? Well instead of reinforcing the rocket and making it able to take more strain, even with hundreds of struts between two in-line fuel tanks (where the rocket always breaks), the rocket still wobbles just as much as it did with zero struts. And that means the struts break.

And that's where the bug comes in. See, whenever any single structural link on the craft fails for any reason other than "part of a staging event, using a decoupler or stack separator (haven't checked docking ports yet)", the game goes "oh no, your rocket must have entirely destroyed itself, leaving nothing behind, so we'll just take away control from you, shut off all the engines (I think, the effects stop anyways), and give you the "you failed, try again?" screen".

I'd much prefer it if I could see WHAT EXACTLY HAPPENS WHEN THAT PART FAILS. Does the vessel keep going? and make it to orbit despite losing a few parts on the way up (as is long-time KSP 1 tradition)? Does the single failure cause a cascade of further failures?
With the game in its current state, I won't ever be able to tell if that happens, because all the game tells me about is the first failure, and worse than that it actively takes control away from me.

Please, never take control away from the player like that unless ALL the control points on the vessel have been destroyed. And if that's the case, check nearby vessels to see if they have usable control points, and switch to the first one of those that is found.

EDIT: Oh yeah, I should probably include all the data about my system just so we're all on the same page:

• Game version: KSP 2 (Steam) version v0.1.0.0.20892
• OS: Windows 10, up to date as far as windows update itself is concerned.
• CPU: Intel Core i7-9700k, stock clock 3.6 Ghz, overclocked to 4.6 Ghz (water cooled, and this CPU has never given me a single hiccup with respect to blue-screens or anything of the sort)
  GPU: Asus ROG Strix 3070 Ti, 8GB VRAM
  RAM: 32 GB 3200 MHz G.Skill Ripjaws V DDR 4 (2x 16gb sticks), on XMP timings 3200mhz CAS 14-14-14-34
  Drive the game is installed on: Samsung 970 Evo Plus 2TB NVMe M.2 Solid State Drive
• Expected behavior: When a structural failure happens on a vessel that is not the result of a decoupling event (such as decouplers, stack separators, and docking ports), the game should continue to allow you to control your craft so you can see if you can still get where you were going despite the failure.
• Observed behavior: When a structural failure happens on a vessel that is not the result of a decoupling event, control of the vessel is lost, the camera freezes in place relative to the world (meaning your vessel goes sailing off-screen), and you are presented with the "try again, revert to VAB, revert to launch or load a save" screen. This is incredibly frustrating when you've spent over an hour designing a craft to get to Jool and are prevented from doing so by the failure of a single strut on the launch vessel (a strut that might not have even been needed).
• Steps to replicate: Build a rocket as follows:
  Start with a 5m (XL) probe core, attach a 5m (XL) nosecone on the front of the probe core, and on the rear of the probe core attach one of the longest 5m (XL) methalox fuel tanks.
  Now stack a tiny truss segment on the bottom node of the fuel tank (second smallest is the easiest one for building purposes), and then another of the longest 5m (XL) methalox tanks.
  To reinforce this obvious week point in the rocket stack, use 8x symmetry to put struts between the 2 fuel tanks in the gap created by the truss segment.
  Then on the bottom XL fuel tank, radially attach 4 of the shortest 3.75m (Lg) fuel tanks, near the bottom of the XL fuel tank, followed by attaching a 3.75m nosecone on the front of these radially attached fuel tanks, and a Mammoth II engine on the bottom attach node of these radially attached Lg fuel tanks. Add launch clamps if you think you want them, it shouldn't make a difference.
  Now it comes time to truly reproduce the bug. Launch the rocket. and I usually a gravity turn at 1750m altitude, and observe as at least one strut breaks despite your best efforts. No matter what structural link fails first, when it fails, you should immediately notice that visually the engines cut off, the camera stops tracking the vessel, and you are immediately presented with the "interesting quote about failures being part of the process" screen that is supposed to happen when you crash a rocket. However, you'll notice that you didn't crash a rocket. You had a single structural failure on the rocket, that was probably able to be recovered from, were it not for the game saying "nope you don't get to control this anymore".
• The theoretical workaround would be to enable "unbreakable joints" however that would only work if I could open the developer options, which I can not figure out how to do, so there are effectively zero workarounds. And if you consider using cheats to not be a viable workaround, there in fact ARE literally zero workarounds other than "hope you built your craft strong enough without struts", which the game is woefully lacking in information on how to do that (I've been going on info from my KSP 1 play time).
• No mods installed, as I don't know of any for KSP 2.

Edited February 26, 2023 by SciMan
Turned it into a proper bug report.

[ChaoticPheonix]

KSP Version: 0.1.0.0.20892

OS: Windows 10

CPU: Ryzen 5 5600X

GPU: RTX 2070, 8GB VRAM

RAM: 16GB DDR4 3600MHz

Installed on: WD Black SN750 SE, 1TB

 

Description of the bug: A craft will have a single part break, usually a strut. The craft is then recognized as destroyed, even though it is still completely in one piece. Kerbals are lost, vehicle is no longer controllable despite still being complete.

• Expected Behaviour: Craft is still controllable and usable when a non-essential part breaks.
• Observed Behaviour: Bug has presented on planes, rovers and rockets, though is most easily replicable on rovers and happens more frequently. I will be using a rover as an example. Driving along the surface of Kerbin, hit a small bump. A strut breaks, then the mission report window comes up and the craft is no longer controllable. Craft is described by the mission report as destroyed, though only one part is actually broken and the craft is still visibly in one piece on screen.

Steps to replicate: Build a rover with struts, mine tend to have struts along the bottom to support the drivetrain. Drive along Kerbin and hit a small bump to stress the struts. Does not happen at every bump, but frequently enough to where rovers are currently useless on my end.

This has also happened with a rocket coming out of timewarp. Strut breaks, craft remains in one piece but is marked as destroyed and is no longer controllable.

Fixes/Workarounds: Avoid rovers, avoid using timewarp in atmosphere.

List of mods: None used, game is fully vanilla.

Screenshots: To be added when I can replicate it again.

[SciMan]

I can attach a craft file to help reproduce this bug, if it helps. I managed to make a rocket that SHOULD be able to put a large lander nearly anywhere in the Kerbol system, but I can't even get it off the ground because the dang "wobbly rockets syndrome" makes it not work.
Seems like it happens any time I use the 5m Engine Plate (rocket turns into a noodle, no amount of struts will stop it from being a noodle because that's how Unity joints work apparently, basically this bug would have been hidden if the wobbly rockets wasn't a thing).

If I could find out how to extract my craft files from the game, this is where I'd be putting the hyperlink.

 

 

 

 

 

Just so we're clear, I'm saying that BOTH the "wobbly rockets" AND this bug that I reported need to be done away with entirely.

And now for a (for me) short rant on wobbly rockets and what exactly that phenomena is (bug?, feature?, other?, etc.).
 

Spoiler

I'm gonna go further than I normally would, because I've been having nothing but problems (and zero successes) lately when I go to build any kind of "large" rocket in KSP.

My position on the whole "wobbly rockets" thing is this: It is, it was, and it forever will be, a bug. And so it needs to get fixed.
Why? Because it's unintended behavior of the physics engine. It has been tolerated for far too long, so long in fact that people have started calling it "part of what makes KSP the game it is".
Well I'm done with it. Wobbly rockets do not contribute to KSP being the "goofy cartoon rocket building game".
Yes, at one point, KSP1 was indeed the "goofy cartoon rocket building game". But KSP 1 in its current state is... not cartoon, and I'd like to say it's aiming to be not nearly as goofy either, since it's trying to teach you things about interstellar travel using real physics.

The reason wobbly rockets got rolled into being thought of as "part of KSP" was because some people thought it was "cute" in the "lol so random" way.
Those people are IMO now firmly in the wrong, given the context of KSP 2 and its aspirations), and so the wobbly rockets took a long time to get even a half-fix introduced, with that half-fix being Autostrut.

Well guess what. KSP 1 or KSP 2, doesn't matter, wobbly rockets are still a bug.
Even if they're currently an "ascended bug", they're still a bug at the core of it.
Now with the focus being on interstellar voyages the wobbly rockets phenomena needs to go the way of the dodo.
To be more clear, it needs to become history, extinct, spoken of in the past tense only, and if that wasn't clear enough, if I never see another wobbly rocket in KSP 2 again, it'll still be too soon.
I could write pages on how much KSP 2 needs to get rid of wobbly rockets to be a successful game.
There just have to be other ways of introducing and simulating the concept of "you need to build your rocket strong enough to withstand the forces applied to it" into KSP 2, without falling back to any sort of solution that results in any kind of "wobbly rockets".
I'm not saying my rockets shouldn't break ever, far from it. I'm saying they should BREAK, not BEND.
And if they DO break, don't stop the simulation, or the engines, or anything. Let me maintain control, let me try to fly out of it, LET ME KEEP PLAYING THE GAME WHEN SOMETHING BREAKS (that's the bug I originally reported).
If this alternate structural limits mechanic needs custom code for determining the loads on individual parts, so be it. If that takes a long time, so be it.
If I was a rich person, I'd throw more than a million dollars at the problem of removing wobbly rockets from KSP2 and replacing it with something better.

Why am I so adamantly opposed to any kind of wobble or flex in my constructed things? Well, the answer is simple.
It doesn't happen to a visible degree in real life, unless you're so close to the rocket that you can't even see the whole thing within your eyesight where you're standing.
Additionally, wobbly rockets does exactly one thing, and one thing only:
It gets in the way of me getting anything meaningful done.

That's all it ever does. It doesn't provide enjoyment, and if I do laugh at a failure that happens due to wobbly rockets I'm laughing because I'm crying inside, and only for just a short moment, after which I'm lamenting the fact that this game "simply must" simulate the physics of the entire vessel using the very fallible Unity joints system.
Wobbly rockets does not provide a challenge that I can surmount with the tools given, and because of that I'm left with a "full" game that I can only access a small portion of due to failings of the tools the game gives me.
I want to send something to the other planets. I really do. I can pilot them there, I can build them that SHOULD be able to get there, I can do all of that (I need to look up transfer windows, but that's not that much of an ask).
What I can't do is get them off the launchpad without the game taking control of that vessel away from me permanently, when the probe core was most certainly not destroyed.

I'd love to test those huge LH2 spherical fuel tanks and the SWERV, but I can't build a launch vehicle big enough to get them in orbit without it turning into a noodle and "oops something broke, you don't get to play anymore"

If I had native XL (5m) engines, I might be able to do it, but I'm very leery of that idea, because I'm nearly certain it won't work out like that.

 

[James Kerman]

Overlapping topics have been merged.