Pecan

If you don't have fuel lines yet then you aren't doing onion staging either. Unstaged (artillery) - whatever can go bang, goes bang and makes the ship go up (or bang). When the bang stops the up stops. Stack staged (classic) - lowest engines go bang, higher ones can't because lower are in the way. When a stage stops banging it's dropped and the beat goes on with the next one. Radial staged - outer engines go bang as long as they can then are dropped and the (next) inner 'ring' starts banging. Much like stack staged but allows you to easily have more engines in the earlier (outer) rings. Parallel staged (which you're probably doing) - since all radial staged engines can go bang at the same time they do, but the inner ones have bigger fuel drums and keep banging longer. Onion staged (symmetry 3+) - parallel, but fuel lines feed all engines 'inside' as well as the engine they're directly attached to. Means the outer fuel drums are used up really quickly but inner rings still have full drums when the outer ones are dropped (earlier). Asparagus staged (symmetry 2) - Onion, but dropping pairs of drums/engines at a time instead of waiting for 3 or more to empty. 2 being the minimum you need to maintain symmetry this is the fastest/most efficient way to use-up and jettison empties. The general point is that with Onion/Asparagus you get the benefit of as many engines firing at the same time as possible (as with parallel) but you don't need to give inner engines bigger drums to keep them banging longer (because they initially take their fuel from outer drums). Slack tanks - Sometimes you find that you only need a few powerful and efficient engines to give you the thrust you need but that each engine needs big fuel tanks to keep banging long enough. It may then be worth moving some of this fuel into an outer 'slack' stage (one without engines). The fuel in these will be used first and the tanks jettisoned, leaving the rest of the vehicle with all the engines but less dry-tank mass to push. The trick is to balance exactly when you drop slack tanks and excess engines so that whatever's left has just the thrust and fuel left that you need. You need a lot of thrust to get moving from launch and - ideally - reach terminal velocity as soon as possible. Then you want lower-but-increasing thrust to follow the terminal velocity curve until it gets so high in the upper atmosphere you throw in everything your rocket can give in order to get to orbit. Finally, you need very little thrust (to mass) in order to circularise your orbit. That profile suggests - small(ish) engine on core stage (last to be jettisoned) that is mainly there to circularise*, big-as-possible engines on inner stage (penultimate jettison) for the upper-atmosphere rush to space, slack tanks for intermediate stages (no more thrust but added endurance for the inner engines) and as-big-as-you-need engines on the outer, launch, stage to help get you up to terminal velocity. Easy. Actually doing it is a different matter, of course ^^. [*I favour a light core stage but temstar recommends a heavy one with more fuel than I do. His designs (eg; forum.kerbalspaceprogram.com/threads/33381-0-20-2-Zenith-rocket-family-(modernised-for-0-20-x-with-perfect-subassembly)) assume that last stage will be doing the 'rush to space' that I assign to the penultimate 'inner stage'. Blizzy's engine cluster calculator defaults to 20-25% thrust from the central stack (core stage) for the same reason]. Apologies in advance for all the mistakes I've probably made in typing this off the top of my head.

You may also like to investigate Rhumb Lines, which are a much easier way to navigate but somewhat longer than Great Circle routes. [says the man who once managed to miss the entire Caribbean and arrive in S. America instead]

All I mean by "strut trick" is using the Cubic Octagonal Strut to make a cluster of engines either under or radially on another part. No idea where that strut comes in the career tech-tree but you can use it to attach just about anything to anything else anywhere and it's massless in flight - although listed as 0.001 in the VAB/SPH. IE; You can usually only put 1 engine under a component but the editor happily allows you to put several struts there (use symmetry!), then you can attach an engine to each strut using the same symmetry. 4 x 48-7S engines will fit nicely under a single 1.5m part, without clipping, for instance. Even if they weren't massless their combined mass is less than a quad-adapter for the same job. Edit for below: Yes, the adapters do look better and the mass difference isn't THAT huge - ~0.1t per adapter. There's also the niggling feeling that using struts is 'cheating' even if the editor allows it. It is a handy way to do things though - your game, your choice :-)

Nah, come on, if you want a beginner's VTOL then: core, 90 fuel, 48 engine. It's stable, controllable, agile and SSTO with enough fuel to de-orbit. Slap a parachute on it and you can still SSTO and recover it. What am I missing in the question?

If you've still got a copy around 0.22 performs much better on low-end machines than 0.23. Until today, when I just decided I had to bite the bullet, I've been doing everything in 0.22 for that reason. It's essential to take vexx's advice about the ocean bug lag to get any acceptable performance out of 0.23 though. [As I've said before, my graphics card doesn't even meet the minimum system requirements, so I can't complain - old nVidia 8300].

While we're waiting ... If you start from 14km up and go down ... it's all fine for the first 13.999...kms

Excellent! I have Inform 7 ready ...

Comrade, when you started this thread I just thought you were being daft. Now my jaw has dropped. Amazing. Thank you.

Here I was thinking that loving '30s was my "guilty secret" ^^ They're included in the demo version so before I started using Travert's mass-optimal charts I assumed they were superceeded by things like the poodle and felt a bit guilty about using clusters of them. Even then I was using the bi-,tri-, quad- adapters to place them since I didn't know the strut trick. They're still the engine I try to find an excuse to use; when Nuclear is just too heavy and 48s too underpowered ...

Wow! Do you know, I think that's the simplest, clearest, explanation I've read anywhere. The whole (vacuum) launch summed-up in only a handful of sentences - well done :-)

I love the idea of a tabletop KSP but this is what I meant - thanks to you all. The telemetry that the space programs of the 50s and 60s were based on was minimal at best. In general it was 'light the blue touch paper and stand well back' then wait to see what the results were if anyone/thing came back. The 3 (?) other 'what if KSP ran on xxx' engines seem to prefer the question of pretty pictures to the one of space flight control. Just making the point here that the numbers are what determine whether it works, even if you don't care about them. ;-) Which is not to say that chucking stuff into the sky and watching awesome explosions can't be fun too, hehe.

Shouldn't this be merged with How do you design?

Is KSP a game or a simulator? If it's a game you need career, et al, so you can keep track of your score. If it's a simulator then you just need all the sandbox options to simulate. If it's both, they both have their place. If it's neither then I hope you're having fun anyway.

Better to get there with an overengineered solution than fail with an optimised one :-) Congratulations, it seems like you've "got it" now, so you'll just get better every time you practice - and so will your rockets.

I thought you were 4.6 - do you have a rounding error there? (joke)

Oh, I just throw them together without thinking about it much ^^ Rockets: 1) Determine mission parameters. 2) Work out the deltaV requirements using a deltaV map. 3) Check Travert's mass-optimal engines. 4) Check Blizzy's engine and GaryCourt's rocket calculators. 5) Create my own engine & design choice. 6) Test and enhance 10 or 15 times. Bases: I go for single-piece because the only way to ensure docking-ports align is by driving a rover under them and dropping legs. I might revisit bases when I install KAS, but until then they're more trouble than they're worth. Spaceplanes: Only for SSTO at the moment and only for space-station crew-shuttle in practice. There's a very limited engine choice compared to rockets, so the rest is trying to get the wings and payload (if any) aligned with the centre-of-mass and trying to make sure the air intakes don't look ridiculous without excessive clipping. Space Stations: Are just big rocket-components without engines. *Grin* - I tend to muck about with them. First was 'Mjolnir' (looks like a hammer), then 'High Moon' (arranged fuel tanks to look like a revolver) and most recently 'Iron Man' (anthropomorphism in space). Whole space program: - I try to consolidate all the things I build into (sub)orbital spaceplanes, lift vehicles, space-stations, transfer tractors (to move space-station parts or to move anything else between stations), mission landers (when the stations just don't go that extra distance). Then there's all the 'other' stuff that I do in different saves that isn't about my main space program - that'll be things like challenges, tutorials, special circumstances, etc.

What, no bases on each body? Have you no ambition! [i am joking, of course] @OP: - and apart from 'just' doing that lot; do it with the smallest craft you can, using the fewest parts. Beat 20% payload ratio with rockets, 40% with jet-assists/spaceplanes. Complete missions with as few launches as possible. ... ... *whisper* Look at what Whackjob does And then there's all those mods to start trying. Personally, I'm trying to optimise a set of rocket-only lifters, then I'll do the same with jets/spaceplanes, then I'll start playing around with RT, KAS, IR and the like. Unless something else distracts me. Oh, and there's still 7 games days until the next Duna transfer window. But since a friend asked I've just made 4 mun-landing-and-return vehicles, i) "Long Tom" ('looks' like a rocket, robust but not particularly efficient), ii) "Fat Sally" (asparagus-staged launcher and integrated transfer/lander replacement for Long Tom. All but halves the start mass), iii) "A" (optimised asparagus-staged across whole mission), iv) [Prototype] "Challenge" (optimised without asparagus staging). As you will see, there is plenty to do that doesn't need 'science' for inspiration.

Be awed by the improvement from CGA to EGA, Thrill to the possibilities. Of course, for the authentic 1950s/1960s feel you'd need to knock everything down much further, disable mouse, etc. etc. (Which would all pretty much eliminate any lag issues on modern computers too). KSP does look good but, thankfully, it offers so much more than flashy graphics. Anyone else still got 'Ascii art' flight-sims?

I have a better than 90% success rate at clicking the right button to tell MechJeb what to do ^^.

1. Nothing over 2.5km is modelled. 2. 'Ringworld's - and similarly any ring-orbiter is inherently unstable. http://physics.stackexchange.com/questions/41254/why-is-larry-nivens-ringworld-unstable

OK, I know I'm posting the link all wrong and other people have done 'space scooter' landings before/better, but I'm new to imgur and I think the expression on Jeb's face in the first picture is worth it: http://ksppecan.imgur.com/scooter

Look carefully at the tank at the bottom of your stack-with-winglet. It does not align properly with the one above it. Symmetry - or whatever - has not attached those tanks to each other, which is why the fuel doesn't flow. As Claw says, you can correct it by adding a fuel line yourself.

My Kerbin station is at 250km, allowing lower and higher phasing-orbits at 100km and 400km. For vehicles that can only just make it to LKO I use a never-landing tractor to pull them to the station (or wherever), or to refuel them. Maximising Oberth for outgoing vehicles just means dropping periapsis to LKO; you're dropping from the station so when you burn at periapsis you're travelling faster than from a circular LKO.

As far as I remember there's no good reason for using the skipper. Travert(?) calculated all the mass-optimal engines and the 2.5m engines really don't win anywhere useful. Mind you, "because I want to" is a good enough reason ;-0

Further to my previous post - the 24-77s improved the design (by 10m/s) so I was wrong, again, and the tool was right :-)