Empiro

As always, it depends on what you're trying to do and where you're trying to go. If you're launching a science probe that's on a 1 way trip, then the craft is just a few tons, and you don't need more than 1500-2000 m/s delta-V. A 909 is more than sufficient. If you're doing the Jool-5 with a giant crewed craft with landers and so forth, then of course you'd go with the nuke engine.

Under what circumstances are you trying to rendezvous with a target in an elliptical orbit? Are you trying to mount a rescue, or trying to dock with a space station around another planet etc.? You can also intercept asteroids inside the Kerbin SOI. The trick is to launch into a circular orbit with the same inclination as the asteroid, with the AP/PE matching that of the asteroid's PE around Kerbin. When the asteroid enters Kerbin's SOI, you can fine-tune your orbital period by burning prograde at the point where your orbit touches the asteroid's PE so that when the asteroid reaches Kerbin PE, your ship is exactly there too, and you can perform a velocity matching burn to rendezvous.

Unless you're playing with a part recovery mod, then you'll be discarding early stages anyways. Liquid engines not too expensive, and liquid fuel is pretty cheap, but liquid fuel tanks are extremely expensive in KSP. SRBs are a cheap way to increase your TWR for your launch stage, and let you get away with a smaller, more efficient engine in your core.

How many m/s is the maneuver? While it's possible that you need more engines, I think it's more likely that your transfer is not efficient as it could be. It can take less than 1100 m/s to transfer to Duna if you do it right.

The Warp-To button should help in the next version. However, I consider things like messing up SOI changes and missing the atmosphere / crashing into the planet to be bugs with the current implementation, and it's those things that really ought to be fixed.

I've never had any issues with Deadly Reentry and FAR when going up, and I often launch with TWRs around 1.6. Unless you pitch over too much, you won't come anywhere close to burning up or destroying your rocket.

You might be able to make repeated free-returns if you're exceptionally precise with your orbital period so that after your return, you'd get another free return encounter. You'd have to keep making corrections since even tiny deviations will quickly put you off course. However, it still wouldn't allow you to ferry things back and forth for free. To dock with a craft doing this, you'd need to match velocities, which means that you're putting yourself in the same orbit. To drop things off at the Mun, you'd need to slow down into a Munar orbit.

I haven't had much use for Ion Engines in regular KSP. I don't have the patience for those extra long burns, and there's just not that much need for so much delta-V.

Science should have a much greater impact on the way you play the game. Right now, you click "run experiment", and you're done. This is true for every single experiment in the game. Plus, there's no reason to do it other than to unlock the tech tree. Science needs to be dynamic and integrated with the rest of the game besides just generating numbers. Science should be a multi-step process, where each step has specific requirements (how high, fast, planetary body, etc.), AND the result should impact the game in some way. For example, you should have to set up something like a scan-sat first, and it maps out potential "points of interest" for you to land. After you do so, you could perhaps use the gravity scanner to detect nearby anomalies, and you get science from investigating those anomalies. In addition, getting that result should open up more avenues of gameplay. For example you might detect a pocket of particularly interesting ore, and that generates a contract to land a giant drill on top of that spot. After you do that, you should get a mission to bring back some big slab of rock from that planet back to Kerbin for analysis (OR land a giant lab to study it, and so forth). Each experiment should have a number of different possible results (requiring different next steps), and it would make things much more interesting.

There's way too many choices in this poll, and it's missing a "I never let Kerbals out of the house!" option. I try to err toward safety. This means not sending Kerbals when not necessary, and having a working abort in case something goes wrong (I play with FAR and don't revert, and I've definitely had launch failures). I do love Kerbal Construction Time's simulate feature though. This does cost me a lot extra, because it can be unwieldy and difficult to put abort modes on a giant space station module, so I usually launch those unmanned, and launch the crew up in a safer, more tested rocket. This is pretty similar to real-life.

"I leave as little debris as possible" should be one of the choices. I will in fact re-design huge portions of my rocket just to achieve this. This usually means a 3-stage-to-orbit rocket. 2nd stage will get me close to orbit -- an AP of 80-90km, and a PE of -10km-10km, which means that it will get removed. Third stage circularizes and also has enough fuel for the rest of the mission, whether that's going into a specific orbit, going to the Mun or going even further. When going to the Mun, I usually will ditch that 3rd stage on and let it crash on the Mun. When going further, I'm not hugely opposed to leaving junk in a solar orbit, but I tried to avoid it if possible to.

Too many options make the game harder to maintain though, especially if they're options that have gameplay (as opposed to just graphical) impact. Basically every time you patch the game, you'd need to test every combination of options. Mod makers probably wouldn't be huge fans either. It also ironically makes the game harder for new and intermediate players, because it's harder for veteran players to answer the question, since they could be playing under very different settings. At most, there should be "easy" and "hard" physics mode, but I'd even just lean toward there being one setting -- hard. When I played using FAR, it took me about an hour and a few videos to figure out what I needed to do to make my rocket not flip over. It's actually not that hard to fly rockets under FAR once you learn. A good tutorial can guide new players in even less time. When it comes to planes/spaceplanes, I'd argue that you don't even need to use them at all to enjoy and succeed at the game. You can visit every body in the game without flying a single plane. Let players who want to take things to the next level learn all about how planes work. It took me dozens of hours to finally make a plane that can lift off and land in FAR, but it was hugely rewarding for me.

I agree with progressive contracts that take you logically from one place to another. For example -- orbiting the Mun to landing, to manned landing and so forth. For base construction, missions should involve rover surveys, manned surveys, and finally picking a great spot to land a base. The one thing is that the game should track achievements so that advanced players can skip ahead, for example, by landing on the Mun directly in the first mission. Finally, I actually think that most science should come from contracts, instead of the other way around. One thing is to make those science contracts non-repeating and not grindy. The reason is that I think science is way too tedious and boring already -- click a bunch of times, land in a new biome, click a bunch more times. Instead, contracts should provide you science, but require you to land in specific biomes. That way, biomes add to the gameplay challenge, but it doesn't encourage you to just biome hop.

I wouldn't mind having more time based mechanics, but there also needs to be greater consequences to spending time on things. Right now, you can just time warp past any delays, and the only consequence for that is missing a launch windows. The only consequence of that is the need to time warp some more... Things like having upkeep costs (it's expensive to monitor satellites, and probes, and the ISS costs quite a bit to keep going) and life support would go a ways to add more consequence to just time-warping. That said, one thing I did like about Kerbal Construction Time is that it made it so that I didn't max out the tech tree within the course of 3 months -- advancing science from basic rocketry to ion drives and nuclear engines should take more than just a few months.

The Poodle has been improved in the past couple of versions. It now has better TWR than the 909. I still rarely use it though.

I've found that under FAR, the following general design elements are helpful: -A full-moving tailplane (like the F-22) is extremely helpful in maintaining pitch authority. A canard can also help, but I rarely have found the need for both. -I like to keep the wings mostly flat at a 0 degree angle, because it makes mounting things easy, but I've found that having small wing parts angled upwards at the tip (i.e most wing parts flat, but the part at the very tip is angled up) greatly improves stability at high altitudes and speeds -I often wind up using a fairly large tailfin to reduce sideslip tendency. It makes the craft look somewhat ugly, but it's what works -For cargo-bearing SSTOs, I like to put fuel tanks in front and behind the cargo bay. While this increases how much the CoM may shift during flight, you also have greater control by transferring fuel around. Ideally, your CoM when fully fueled, totally empty, with payload, and without payload (and all combinations thereof) are all close to the same place. -Use airbrakes. It saves you time when landing, and gives you far more control over your speed. It lets you have much steeper re-entry angles without risk of burning up or breaking apart.

There were a number of threads that discussed this a while back. There is a maneuver planning reason: You get to use maneuver nodes and fine-tune your maneuver, including adding normal / radial components so that you get an encounter straight away. I don't know how much you're burning once in a solar orbit, but if it's more than 200 m/s (and you're not going to Eeloo / Moho), then you're spending more Delta-V than needed. There is also the efficiency / cost reason: By burning straight up, you're either wasting Delta-V due to gravity drag, or you're having to use lots of engines to get a high TWR to minimize gravity drag. You can probably launch a craft that is lighter and cheaper that will do the same thing. - - - Updated - - - The exact numbers are going to depend on your rocket design, but the basic gist of it: 1. You have lots of loss due to gravity drag. How much depends on your rocket. If you find yourself with more fuel than 3), then your rocket has way too many engines, and you'd be able to launch a rocket that's cheaper / lighter and still perform the same mission. 2. You're not maximizing the Oberth effect. You'll find that you're burning 950 m/s to escape, and then about another ~500 m/s to get into a Duna intercept 3. Takes advantage of the Oberth effect. You'll only need to burn about 1080 m/s from orbit to get the Duna intercept. This difference gets bigger the further out you go, and makes a huge difference when travelling to Jool or Eeloo.

Also put me in the list of "48-7S is too good". All you really need to do is to look at the optimal engine graphs linked by smartech below to see it. The 48-7S occupies a huge middle region in every single graph. No other engine is like this. No, the 48-7S isn't always the best, but it is the best in significantly more circumstances than any other engine in the game.

With the default settings, TAC doesn't add too much to your launch window considerations. I haven't played with TAC LS in a few versions, but when I played, the large 2.5m life-support container could sustain a crew of three for several (Kerbin) years, and it didn't weigh that much either. For my Duna mission with a crew of 3, I think I only needed two of those modules (and no water or CO2 filters), and had plenty of life-support remaining when I got back. Both ways, I just did a minimal delta-v Hohmann transfer at the standard launch windows, and I think the total trip was under 3 Kerbin years. One extra consideration for me was that I played with Deadly Re-entry, so I couldn't do too aggressive of an aero-capture, which would be necessary with faster transfers (I still had to expend about 100 m/s back at Kerbin to get into a capture without burning up).

The consideration is basically the same -- make sure your ship has rotational symmetry or you'll have a tough time keeping it straight as you burn. Using the large docking ports is a good idea since those have pretty high strength and resistance to flexing. You're still likely to get lots of wobbling though. The key to turning such a craft is patience. Turn off SAS, use RCS, turn on fine-control mode. Only fire the RCS briefly and let your ship slowly drift to the right orientation. If your ship still wobbles too much, you might need a mod like Kerbal Attachment System that lets you add struts in space, but personally, I've never had issues with just being careful and gentle with the turns. Puller designs work well if your ship is very long, and will reduce flexing. Make sure that the exhaust doesn't hit any part of your ship, however. One piece of advice is to turn off the gimbals on the engines of the tug (and any engine that is in front of your center-of-mass). If you don't, the engines will gimbal in the wrong direction and steer you away from where you want to turn. This bug/quirk might be fixed at some point, but is still present as far as I know. This means that you'll need to have several SAS modules or use RCS to maintain altitude control.

My recommendation is to not worry about atmospheric ISP versus vacuum ISP too much if you're trying to get into orbit. The reasoning is that most of the engines you'd use on the launch pad (SRBs / engines with good TWR) all share similar atmosphere - vacuum curves. If your launch TWR and flight paths are all similar, then the "atmospheric ISP loss" will be a similar percentage across all the rockets. Instead, work exclusively with vacuum ISP. The rule of thumb of 4550 m/s in stock, and 3300 - 3500 m/s in FAR (depending on the mass and shape of your rocket) of vacuum delta-V will serve you very well. Things like better piloting, and picking the optimal TWR at launch will have a far greater impact on the net cost of your ship and total delta-V expended.

Two more points of general advice that might be useful to you in the future: If you wanted to launch directly into the second orbit, then the trick is to timewarp at the launchpad until your desired orbit passes right over the KSC, and then launch into the polar orbit. This way, your resulting inclination change very small. Second, in general, when you perform a maneuver, you affect your orbit at every point except where you are currently at. This is why your first maneuver didn't work out -- no matter what sort of maneuver you perform, one point in your orbit will always be at that maneuver node, which isn't anywhere near your target. To match orbits, you want to burn where the two orbits "overlap".

Also, I believe that for Kolniya and Keo-synchronous orbits, the orbital period matters more than the exact AP/PE. Try to get it as close as possible to 6 hours.

Most likely so -- how much speed are you getting before having to switch over to rocket engines? I'm definitely having trouble getting more than 1250 m/s these days on jets, most likely due to the new drag mode. Even a bit more drag leads to less speed, which leads to less intake air, which leads lower throttle, which leads to even less speed, and so on. This means that you'll need about 1600 m/s on rocket engines to make orbit, instead of the 1200 or so from before.

True, but I think that for this would be a fine trade-off to make for the sake of gameplay and challenge. It would give more variety to the contract than just having different orbits, which you usually can use the exact same rocket to accomplish.