Ninety-Three

Ah, that's what I get for skimming, I thought the bonus items counted towards my budget. I suppose I'll have to do up a fuel-efficient take on it after all.

This was an interesting challenge, stuff is costed really weirdly (mainsails cost as much as other engines, the 1.125 and 9 ton tanks are far and away the most cost-efficient ones, and so on) which makes optimizing different. Here's my attempt to come up with as much of a budget surplus as possible, having tested it I could afford to carry less fuel in general, and more on the second stage, but I'm too lazy to relaunch the same basic build. A more detailed account of the mission (including staging, which deorbits cleanly) can be found here. The ship in the hangar: Notable budget features include drogue chutes mounted on octo-struts, tiny stack separators, and no SAS. I would've used less RCS (the entire docking used a mere 1.89 units), only there's no symmetrical way to attach a less expensive amount of the stuff. The 2 person capsule would've been cheaper, but heavier, and disallows a second science/parachute bonus, so I went with two small capsules. Minus the 240 for KER, it comes out to 19380. I earned every bonus item, and my second capsule had a full science suite and its own redundant parachute, earning those bonuses a second time. The budget adds up to 27200, 7820 unspent. Mission photos: As I understand it, the extra capsule (and associated extra science and extra parachute) bonuses can be claimed for an arbitrarily large number of capsules, given that the whole capsule affair weighs under a ton, it seems like you could run the budget up a lot higher, and claim a net profit, by sticking on half a dozen capsules and just a bit more fuel. If no one else does that, I'll give it a go just to see how silly it gets.

On multipliers, if you get 3 2x multipliers, does that go 2x2x2 = 8x, or is it +100%+100%+100%=+300% ?

I've seen spaceplanes that can reach near-orbit on just airbreathing engines, to the point that they can circularize with ion drives. From there I imagine it's very easy to ion out of Kerbin's SOI entirely. I'd enter it myself, but I'm terrible with planes.

Plur: I'm the one who put up those crazy scores in the altitude thread, the method won't be very helpful. If you use it to achieve any truly absurd speeds, it will tear off every piece of your ship except the core. Also, it doesn't scale well: Ten times the mass means ten times less decoupler speed.

Yeah. Decouplers exert a fixed amount of force, which is usually just enough to make two bits of separated ship move slowly away from each other, but that's because the ships are large. You get an order of magnitude more value out of your force if you use a probe core rather than the manned capsule, but the manned one has lower drag, and since Kerbin's atmosphere is a magic aerobrake, the manned capsule is the better plan. That's true for 200ish decouplers at least, theoretically there should be a point at which the probe core gets up to Mach 300, clears the lower atmosphere in three frames, and is free to blast into space at speeds that could make 100 km in 30s. Whether or not that's doable within the limits of KSP has yet to be tested, at some point when my computer's free to hang for an hour, I'll do a 768 decoupler probe and check.

And here's my high score for now. I may make an even bigger launch tomorrow, but the loading times on this thing get absolutely insane and I'd like to have my computer for the rest of the night. 768 decouplers, nearly an hour of load time. Thanks go to metaphor, for the idea to make a single-stage decoupler craft, and to zarakon, for teaching me that several-minute loading times mean you can still double your part count. Note to Mars: Feel free to not list both this and my 40 km entry in the high scores, as they're the exact same principle, just with more parts alt-clicked onto the side of one.

Based on the screenshot, it looks like an infiniglider, which I believe you're disqualifying (based on the earlier infiniglider that hit something like 80k). Status update on the Decoupler rocket: Just launched 704 decouplers, 40km in 30s. It took half an hour to load. I'm uploading a photo now just to earn those "first to X height" achievements, and continuing to perform larger launches. I've been seeing a pleasingly linear trend of 512 = 30km, +64 parts = +3.3 km, It's looking like I might hit 50!

I was getting 43-44 km/s off 200-ish decouplers, so I bet it's even more impressive. I just did a successful manned launch of 576 decouplers to 33966 m, and I'm not crashing yet, so I will continue to turn up the part count. I also need to run a 512 decoupler drone just to see if the magic atmosphere stops applying at sufficiently high velocities, though I'm doubtful.

Well hot damn. I loaded up 512 decouplers and it did indeed make 30 in 30 (I also got the design down to 5 struts per 4 decouplers). I'm going to see how much farther I can push it in the hopes of taking a high score.

What are your hardware specs? If mine are similar, knowing that ten minute load times can actually lead to a launch rather than a hang, I'd like to crank mine up to say, 550 decouplers and give it a try. If not (or once I do), I'll happily share the technique (which I think might be further improvable from 3 struts - 2 decouplers to 5 struts - 4 decouplers).

To borrow a phrase, how did you manage that? 200 decouplers isn't getting me to 10 km in 10 s, and that's putting up minute load times. Awesome computer, or are you just cranking the decouplers up to 11 (hundred) and KSP sometimes allows it?

Inspired by metaphor's craft, I built a ship that uses decouplers for propulsion. I designed a build method that allows for a 1.5 ratio of decouplers to octo-struts. Unfortunately, it -still- doesn't work well. I can use 200 decouplers (about as much as my computer can handle, the load times get very bad) to launch a probe core straight up, but in ten seconds it's slowed to just a few hundred m/s, and it's only 10 km up. Unfortunately, drag reduction doesn't seem to work under current KSP physics. Using a low-drag manned cockpit leaves you with way too much mass, and you can't attach anything to the probe core to reduce drag because the astounding impulse of launch tears off anything with mass (fun fact, struts and the tiny 0.001 mass blocks don't actually have mass, despite what the game says) (and no, it doesn't tear it off at mach ten, the part reaches some reasonably high speed, tears off, and leaves the rest of the ship to continue accelerating). In conclusion: There's no way to reduce your drag, so the magic atmosphere will prevent this method from being effective here. If there's ever a 10 second altitude challenge though, we know how to do it.

Put a circular decoupler (they vary in how they decouple, so only some will work) on top of your ship, put an RCS thruster on top of that. It's worth noting that switching to a jettisoned RCS port resets the timer to 0 seconds, so remember to count however much time you had on the previous stage too.

I tried something new, but unfortunately it didn't score that well. The idea was to use just two or three stages of extremely dense sepratrons to get up initial velocity, then jettison a single RCS port (that part has a uniquely, astoundingly low drag) and watch as it flies through the atmosphere without resistance. The principle was sound, and maybe in a version of this challenge scored by weight it would have done well, but it was only coming in around 20,000 m. Having gone bust on my last creative idea, I'm pretty sure 50,000 m is just impossible, barring an actual supercomputer that runs some hundred-thousand part partially-burned sepratron beast. That is, impossible barring the few exploitative ways this thread has already produced (infiniglide, very tall launch clamps). Although speaking of exploitative, I've just had another idea allowable within the rules-as-written that should pass 50k, I should have some screenshots by the evening.

Well of course, that's what KSP is all about! Or it is when Jeb is in the pilot's seat at least.

It's not as easy as you seem to think. Sepratrons only last three seconds, and it takes a _lot_ of them to lift a ship. It's not like they automatically defeat the challenge, you still have a short amount of time to get your rocket pointed upwards (especially the way I did it, with vectored mainsails rather than RCS for turning).

I don't think this will be as easy it looks. In my experience, it's hard to flip a ship while parachutes are deployed, they try to right it, and I imagine this effect gets a lot worse as your ship gets bigger. I'd be interested to see if someone can make it work for a large ship though.

Read the rules.

I wish I'd documented it more thoroughly, I did a bunch of testing a few weeks ago which proved that a craft's drag coefficient is equal to the weighted average of all its components. Is that the sort of science you're thinking of?

So I immediately saw a loophole in the rules. I slapped together a lazy mainsail beast, then added sepratrons. The rule that only full fuel tanks and capsules count means that my payload comes in at 26.5 tons. I put the whole thing up on imgur, with a bunch more screenshots of the initial launch to show how it went. One important technique: Since I'm using only vectored mainsails for angular control, I had to build up some quick angular momentum, then turn them off until my ship stopped pointing down.

You don't seem to understand. Some people enjoy piloting in KSP. Some of these people make KSP challenges that involve piloting. You're free to mod piloting out yourself, but telling other people to stop having fun their way seems a bit much.

And pointing your engines at the sky defeats the point of going to space. This challenge is not about practicality. Also, as someone who enjoys piloting, that's one hell of a generalization you're making.

Flying a ship out of an upside-down start is mostly about piloting, MechJeb defeats the point.

I'm loathe to submit second place, but I'm having a lot of trouble launching my larger design, and I don't have the patience to reset until it works today, so here's a solid second. Staging: 120,30,9,3,1 I completely guessed at the numbers to use for fuel left in each stage, but I doubt that optimal performance could take me to first place. For that I'd probably have to get a good launch out of my 288,72,16,4,1 design. 29178. I almost want to install MechJeb just so I don't have to deal with altimeter rollover.