How far can you get on SWAGs - before you have to do some math?

[JoeSchmuckatelli]

Thus far in my 'career' (2 weeks), I've gotten through everything via guessing, testing and revising.  That's gotten me to orbit (once) and back.

 

At what point - if ever - do I need to start really doing some number crunching before a mission to have a chance of success?  Also - are there resources out there that can help an interested guy with a BA figure out how to run the numbers?

[worir4]

How far can you get? all the way!  You don't need maths on KSP. That's one of the things that's good about it. Of course it is easier if you get Kerbal engineer to crunch the numbers for you but even without that it is possible to do all the contracts without calculating. Either through trial and error or just having more fuel than you need. Play it the way you like. 

 

[problemecium]

I'd advise you to trawl through the Tutorials subforum and see how many of your questions that answers.

[KatzOhki]

I don't bother to calculate anything a lot of the time, but there are some handy resources. Search for transfer calculator, transfer window calculator, delta v map and delta v calculator for the most helpful ones.

[Xavven]

You don't have to do math... you can get pretty far on trial and error alone. But you will do much better if you use a mod that calculates, at minimum, delta-v and TWR.

Or do the math yourself, but I recommend a spreadsheet as opposed to calculating each stage by hand. That could get tedious. Check out the Tsiolkovsky Rocket Equation for starters for calculating delta-v, and TWR is just your weight divided by thrust, all over the acceleration due to gravity at the celestial body you're on (9.8 for Kerbin).

Personally I find the game more fun if I can design something for a particular job and know its exact capabilities. When it works, it's a great feeling knowing my math and mission planning was correct. When it fails, it's a great feeling learning something new. Without the math or theory behind it, it's like throwing darts blindfolded (which, now that I think about it, is probably a lot of fun too).

[Plusck]

I've never really calculated anything (yet), or used any mods (yet). I have created a basic rocket equation spreadsheet (basically for the essential: (i) how much fuel do I need for my interplanetary / landing stages and (ii) if I try to get back with what's on board, will I get there?) 

Having too much fuel going to the Mun / Minmus is not a problem as soon as you build space stations, because you can drop off fuel for your other stuff when you get there. It's more of an issue for interplanetary travel since it could be years before you have a station set up.

I have often used an online "optimal rocket calculator" too. The only problems are that the rockets proposed by the script would look horrible, and the script hogs the CPU, but I find it useful to get a ballpark idea of what I need to get where I'm going. It also helps to identify where I'm making hugely expensive rockets where there would be a cheaper solution.

Edited December 13, 2015 by Plusck

[GoSlash27]

I'm a weenie when it comes to this. I never build a rocket without doing the math first. On the one hand, it removes all the fun and suspense of wondering whether it's going to work (of course it is ), but on the other hand it allows me to complete missions I otherwise couldn't do. And for me (and I know I'm in the minority here)... I actually enjoy the engineering side of it.

 I don't use any mods to help me with this, just some spreadsheets I knocked together.

Best,

-Slashy

[Pecan]

ESSENTIAL formulae

• Somewhere to sit + box of boom = two parts
• OR somewhere to sit + box of boom + boom-controller engine = three parts

Beyond that, it's all up to you.

If you practice long enough then you'll get a feeling for what you can throw at the sky without it just ending in fireworks but you'll never have the faintest idea why unless you quantify it and once you do that you're into the maths whether you like it or not.  In practice - I'd say the moons are ok for 'by eye' but interplanetary is just easier/quicker if you do the planning up-front.  You can certainly do interplanetary without the maths but it'll mostly just be a frustrating waste of time that ends in failure, so why bother?

Now, what the hell are 'SWAGs'?

[Superfluous J]

36 minutes ago, Pecan said:

Now, what the hell are 'SWAGs'?

I don't know what the first S means, but in my experience WAG is "Wild-Donkey Guess" where a different word is used for "Donkey."

Edited December 14, 2015 by 5thHorseman

[Mastikator]

If you're gonna start doing math then I'd get KER, or something else that calculates delta-v and TWR.

That being said you can always just add more SRBs. You can get anywhere with enough SRBs

[OhioBob]

You can probably SWAG your way through almost everything but, if you do that, be prepared for most of your missions to either under or over designed.   If you want to carry out your missions both successfully and efficiently, you are probably going to have to do at least some math.  Mainly you need to know two things: (1) how much Δv do I need, and (2) how much Δv do I have.  The first part includes planning out your missions step-by-step, identifying what maneuvers need to be performed, and determining the Δv of each maneuver, which when added up is your Δv budget.  The second part involves computing the Δv that each stage of your launch vehicle/spacecraft can produce, and making sure it can perform the maneuvers that make up your Δv budget.  You want to match your vehicle to the mission with a reasonable margin.

Much of the first part can be determined using a Δv map, and the second part can usually be determined using a mod like KER.  However, those tools have limitations.  That is were it is useful to understand the math so you can identify instances when the usual data may not be reliable.  You can then manually compute what you need to know.

An example of where the tools might fail you is when you have a two-part spacecraft, such as an orbiter and a lander.  KER will give the Δv based on how the parts are stacked and staged in the VAB.  However, in practice those parts may perform some maneuvers while docked together, and others while separated and flying independently.  The actual Δv that you'll attain depends on how you fly the mission, which is something KER can not possibly know or calculate.  In this case you've got no choice but to compute some of the Δv by hand.

Furthermore, Δv maps just give typical values.  Transfers to some planets can vary significantly in Δv from one launch window to the next.  I highly recommend Alex Moon's Launch Window Planner for planning your interplanetary missions.

One place you can go to start to learn the math is my web page, Rocket & Space Technology.  The site gives much more information than you need to play KSP, but you can skim through it and study the parts that are applicable.  In the Rocket Propulsion page you should at least familiarize yourself with "Tsiolkovsky's rocket equation" (eq. #1.17), "Specific Impulse", and "Staging".  The Orbital Mechanics page is a good overview of the subject, it probably wouldn't hurt to skim the whole thing and then concentrate on the parts of greatest interest. 

16 hours ago, GoSlash27 said:

 And for me (and I know I'm in the minority here)... I actually enjoy the engineering side of it.

You and me both, brother.

 

Edited December 14, 2015 by OhioBob

[mattinoz]

I did the math once... It's just not as fun as people make it out to be.

Now I use this....

https://itunes.apple.com/us/app/ksp-deltav/id908896518?mt=8

Work out my Payload stage in the VAB then mass it up in the app just using ore tanks to avoid finding all the bits and pieces, then use the app to sketch out all my stages. Worked very well so far to get to Eve and Duna. Hoping it'll work well beyond that.

[Xavven]

3 hours ago, mattinoz said:

I did the math once... It's just not as fun as people make it out to be.

 

What? You don't think manually calculating the delta-v of an asparagus-staged rocket is fun? You're not playing KSP correctly

[GoSlash27]

9 hours ago, OhioBob said:

You can probably SWAG your way through almost everything but, if you do that, be prepared for most of your missions to either under or over designed.   If you want to carry your missions both successfully and efficiently, you are probably going to have to do at least some math.

^This.  I personally would hate to try a hard mode career on WAGs. No quick- saves, no reverts, reduced rewards...

If you underestimate the mission requirements or overestimate your ship's capabilities, the best thing that can happen to you is you don't go to space today. If it's the other way around and you overengineer your ships, there's a lot of things you just plain won't be able to do because of the pad mass, part count, or cost.

The math makes the exercise a whole lot easier, especially if you automate it.

The way I do it is by taking the rocket equation and turning it backwards. Click for details

This is even better than using a build aid like KER because it allows me to compare all engine candidates for a stage simultaneously and decide which is ideal for the role at a glance. It also tells me how many fuel tanks I'll need, how much the stage will weigh, and how much it will cost.

Best,

-Slashy

 

[Decoherent]

Bah, math. All you really need is Kerbal Engineer, the dV chart to make sure you can go where you want to go, and maybe some fancy stuff like the transfer window planner plus KAL. If you can't get there with those, you probably can't get there Spreadsheets make KSP slightly less fun than an actual "game", for me.

[Red Iron Crown]

14 hours ago, GoSlash27 said:

I'm a weenie when it comes to this. I never build a rocket without doing the math first. On the one hand, it removes all the fun and suspense of wondering whether it's going to work (of course it is ), but on the other hand it allows me to complete missions I otherwise couldn't do. And for me (and I know I'm in the minority here)... I actually enjoy the engineering side of it.

 I don't use any mods to help me with this, just some spreadsheets I knocked together.

I always crunch the numbers first, too (though I use mods for it most of the time, spreadsheets are how I check if the mod is working correctly or do something unusual). Strangely, I am not afflicted by "of course it will work", I have discovered there are many modes of failure that the math doesn't cover (lack of control authority, lack of rigidity, insufficient heat resistance, bad aero properties, pilot is a ham-fisted idiot, and so on....). The math eliminates a couple of modes of failure, but there's still some tension as to whether the thing will work as calculated.

12 hours ago, 5thHorseman said:

I don't know what the first S means, but in my experience WAG is "Wild-Donkey Guess" where a different word is used for "Donkey."

"S" in this context is "Scientific".

[moogoob]

As said before in their thread by others I advise the use of Mechjeb (though Kerbal Engineer Redux is just as good) to figure out delta-v values and TWR for staging - takes most of the guesswork out of it, leaving the rest of the mess-ups to either insufficient planning or insufficient piloting skill.

[Pecan]

I totally agree with all those saying use a mod.  You need the results from the maths but you don't need to do the calculations yourself, let alone by hand.

It helps to understand the results if you know how they're derived but a dV map, transfer window mod and info-mod (KER, MJ, VOID) are all you need.

[qromodynmc]

I used to be like you, then i took an arrow to the booster.

Joke aside, use Kerbal Engineer redux, it's simple awesome and useful, you dont have to worry about twr and dv anymore.

 

[Curveball Anders]

I prefer going on SWAG or eyeball-engineering.

Mainly because I think it's funner, but also because I know my own OCD tendencies. So if I start to look for the 'optimal' solution I can easily sink evening after evening into research, testing, math and coding.

Without getting anything actually done in the game

[A35K]

Or you could just overbuild everything like I do... For my first Duna mission I sent an asparagus stage lander with about 6000m/s of Dv, which then docked in orbit to return to Kerbin. Once in orbit I realized the lander could almost make it from LKO to Duna and back by itself, let alone being carried by the mothership.

[Rainbowtrout]

I don't/rarely use maths. My ships are fueled by willpower, blood, sweat, and tears.

[OhioBob]

13 minutes ago, Curveball Anders said:

So if I start to look for the 'optimal' solution I can easily sink evening after evening into research, testing, math and coding.

Without getting anything actually done in the game

I've been guilty of that far too often.  I've probably spent at least as much time developing and working with my spreadsheets as I have playing the game.  (Of course, the fact that the game keeps changing with each new release is largely to blame for that.)  On the good side, I very rarely experience a failure because of a poorly design mission.  I will still occasionally get an unstable rocket that requires a trip back to the VAB to adds fins or whatever.  But mistakes with Δv might happen only a few times out of a hundred missions, and, when it does happen, it's more likely due to poor piloting than bad planning.

Most of my failures are due to either poor piloting or silly procedural mistakes.  For instance, I just recently ran out of power on a mission because I forgot to extend the solar panels.  Fixing those kinds of stupid mistakes are what quick-saves are for.  What really aggravates me is when I leave some small but critical part of a ship, like an antenna.  Or when I re-fly an previously designed mission to a new biome on Mun or Minmus, only to I realized when I get there that I forgot to add a newly discovered science experiment.  All the math in the world won't prevent the occasional oversight.

[OhioBob]

17 hours ago, GoSlash27 said:

I don't use any mods to help me with this, just some spreadsheets I knocked together.

I design my vehicles using some basic rules of thumb, and use KER to compute the Δv.  My rules may not always give me the optimum design, but I know from previous studies and experience that I'll get something that works reasonably well.  KER is a great tool that eliminates most of the tedious math, but it does have limitations.  For a simple single payload mission, KER is usually all that's needed.  However, for more complex missions with multiple payloads, or for missions that require multiple launches and orbital assembly, KER just can't provide everything I need to know.  That when I use old fashion paper, pencil, and calculator.

[MalfunctionM1Ke]

At what point - if ever - do I need to start really doing some number crunching before a mission to have a chance of success?

You dont need to if you know how to do it. Kerbal Engineer Redux takes the whole dV calculation out of your hand and compared with a KSP-dV-Chart you only have to add up the requirements along your route (and back... maybe)