Realism in KSP

[cantab]

Right, but not because your thrust increases but because your weight decreases

Which is in general the biggest reason behind TWR rising as a rocket ascends, at least. Compared to many of KSP's issues, I don't think the constant thrust is the end of the world. But then as discussed it'd hardly be difficult to change either.

[Motokid600]

Which is in general the biggest reason behind TWR rising as a rocket ascends, at least. Compared to many of KSP's issues, I don't think the constant thrust is the end of the world. But then as discussed it'd hardly be difficult to change either.

But it's not constant. Your TWR does not remain at a fixed value as you ascend in stock KSP. I realize the difference between stock and FAR/KIDS, but to say stock ksp trust remains constant ( fixed? ) Is what confused me to begin with.

[Red Iron Crown]

But it's not constant. Your TWR does not remain at a fixed value as you ascend in stock KSP. I realize the difference between stock and FAR/KIDS, but to say stock ksp trust remains constant ( fixed? ) Is what confused me to begin with.

The thrust of an engine (not the rocket's TWR) is constant currently. A Mainsail produces 1500kN at any atmospheric condition; its rate of fuel consumption changes with air pressure.

The more realistic mechanic is to have fuel consumption rate remain constant, and thrust would vary with atmospheric pressure. E.g. The Mainsail would produce less thrust on the pad than in orbit.

[NathanKell]

But it's not constant. Your TWR does not remain at a fixed value as you ascend in stock KSP. I realize the difference between stock and FAR/KIDS, but to say stock ksp trust remains constant ( fixed? ) Is what confused me to begin with.

Thrust-to-WEIGHT ratio is not the same as thrust period. No one is saying TWR remains constant.

If you don't believe me, just right-click on an engine and watch during ascent. As your TWR increases, what happens to the "thrust" line in the right-click menu?

[Motokid600]

Ah.. there we go. Little light bulb just went off. Thanks fellas.

[MisterDoubleSevens]

To a degree - it already happened. In a split between Sandbox and Campaign players (it's most obvious is a suggestions forum)

But even if we ignore that - split will happen in 0.25 and it's new "difficulty options". It's pretty much unavoidable.

There's another really obvious split you somehow missed.

Hardcore Realism (you guys) vs Loose Realism (us guys) vs STOCK-IS-PERFECT (not us guys).

Oh yes, and then there's the few (regex, maybe ferram) who do want Total Realism, but are also smart enough to understand that they'll never get it from KSP and just try to get as close as they can.

[Ippo]

There's also the split between mods and no mods.

Let's be honest: there is just no excuse against some realism proposals, the Isp one being the most obvious. Some things aren't even about "it's a game" or "it could be a simulator", they are just plain wrong no matter what your stance is.

[NathanKell]

Interestingly, the Isp thing allows a further way to "balance" high efficiency engines: you *will need* lifter-class engines to lift off if your upper stages don't produce much thrust. (This especially holds true for the LV-N; it would be a lot less "OP!!11one" if it produced 1/4 the thrust, rather than consuming 4x the fuel, at sea level).

[Daishi]

Interestingly, the Isp thing allows a further way to "balance" high efficiency engines: you *will need* lifter-class engines to lift off if your upper stages don't produce much thrust. (This especially holds true for the LV-N; it would be a lot less "OP!!11one" if it produced 1/4 the thrust, rather than consuming 4x the fuel, at sea level).

Agreed.

(Is there a deadly radiation mod? We need a deadly radiation mod. MOAR TUMORS)

Edited September 5, 2014 by Daishi

[O-Doc]

I've had a good think about this for the last few days and played a bit of KSP in the mean time. I think I'm ready to articulate my opinion about realism in the game.

There should be a realism game mode in terms of size and scaling. If N-Body physics is not too difficult to implement the, sure, why not. The default should be what it is now, tiny celestial bodies with the current 'lite' physics implementation(if N-body is going to hog CPU cycles even with Unity 5). Life support, deadly re-entry, etc should be game options.

So, that's my own opinion. I personally prefer the smaller size Kerbol system because if it was too big then I wouldn't get any enjoyment out of the exploration component of the game. The other day I spotted my first easter egg and it felt great. There's no way you would find easter eggs on a larger scale. Also, I'm a big believer there should be more easter eggs and a reason to be on the surface exploring each 'planetoid'.

So, the exploration part of the game is fun in the current scale. The other way I enjoy the game is building stuff and testing, testing, testing. Most of the game for me is spent in the VAB/SPH building machines because I'm a tinkerer personality. Then, I test ascents and mission performance over and over. If it took 6x as long to get into orbit that's 6x less testing I can do which is a big reduction in gameplay. Not to mention, when I want to build a space station or ground base it will take that much longer to achieve my goals. This game already takes a huge amount of time for me complete my personal projects in it. The last big station I built was 30hr+ of gaming and was built with and SSTO lifter to LKO to undock payload to tug which lifts it to KSO then, passes the football to my station builder which finally places it onto the station. The station was only about 25 parts, but still, that's a fair amount of gaming. I don't know if I could be bothered with it all if getting to orbit took 6x as long.

[Hannu]

So, that's my own opinion. I personally prefer the smaller size Kerbol system because if it was too big then I wouldn't get any enjoyment out of the exploration component of the game.

I agree. Larger solar system does not give anything but increased waiting time, if performance of parts is well balanced and physical laws are correct. You have to take same physics into account on every maneuver. Even larger planets would not give anything before procedural detail generators are developed by astronomical factor (which is unrealistic in game). My opinion is, that better physics should be the most important thing now. Aerodynamic model (including re-entry effects), resource management should be technical (living support, maintenance) instead of financial etc. This should be a space flight game and not a bureaucracy game with beautiful buildings in KSP which breaks nicely when you hit them.

Sorry, I know that I mentioned some illegal words. But too hard limitations will destroy reasonable discussion. If I was SQUAD's boss, I took most stuff from illegal suggestions -list to the first places on the list of the highest priority things to do (yes, you are right, I am a nerd, not a businessman).

[tater]

I don't think any of us in the 'realism" camp disagree at all, actually. The only rationale posted for any size increase had to do with balance once better atmospheric effects are in. A few were testing a slightly larger Kerbin.

[Sky_walker]

There's another really obvious split you somehow missed.

Hardcore Realism (you guys)

Me? Hardcore realism? Where?!

I'm all for balanced realism. I'm far from being a fan of radiation modelling, realistic effects of the G-forces, first person camera, realistic damage modelling, limited ignitions, real fuels, etc. etc. I don't even play Realism Overhaul mod - which all by itself still doesn't offer a simulation-level of realism, but that's pretty much what I would call "hardcore realism in KSP". Changes I'd love to see are linked in my signature, and that's pretty much all (with a proper balance - eg. life support shouldn't be a concern unless I go away from Kerbin and it's moon or build a space station). It's really far away from "hardcore realism".

And regex very specifically said many times that he doesn't want total realism.

Don't try to put words in other people mouth, cause it's rather... unkind.

Edited September 5, 2014 by Sky_walker

[NathanKell]

Well, there's also the rationale "you really shouldn't be using planets with insane densities when you're teaching kids about space"--Kerbal.EDU should have the real system at least as an option.

Also, I think a reasonable case can be made for something like 6.4x Kerbin (or maybe only 3x? who knows). Basically, juuust large enough that parts can have realistic sizes and masses and retain the same payload fraction they do now. If that happens, then we can have reasonable Isp, reasonable dry mass fractions, etc., but keep the same "easy sandbox" feel.

That said, given the sense of the community (although one might question just how many have *tried* a smaller variant of RSS, but eh) I don't think that the second will be happening, and I won't spend time and effort arguing for it.

EDIT:

Yeah, if all these "TOTAL REALISM!!!" folks would actually *name* one person who has said they *do* want total realism in stock KSP, that would be great. Heck, even the thread title was edited to get away from that strawman.

[Laie]

Let's be honest: there is just no excuse against some realism proposals, the Isp one being the most obvious. Some things aren't even about "it's a game" or "it could be a simulator", they are just plain wrong no matter what your stance is.

Well, I'm not sure whether I really *want* it, realism be damned. The first stage gaining like 10% raw thrust over the first 2000m (on Kerbin) would add another layer of complication to rocket design. Probably no biggie -- just plan your rocket with an initial TWR of 1.6; due to weight loss and atmosphere it will get to TWR=2 within the first 30 seconds or so. But that's something I know now. I didn't know that way back when.

I still remember how I looked at engines' stats and tried to make sense of them. I was willing to ignore ISP for the time being (don't know what it is anyway, let's ignore it for now, just patch together a rocket and see how far it goes). I don't think that I'd have ignored variable thrust in a similar manner: Everybody knows that thrust is important, nah, essential.

In the end, variable thrust doesn't make that much of a difference (see above), but how should I have known? The first message of the game would have been that rocket science *is* tricky, after all.

[cantab]

Agreed.

(Is there a deadly radiation mod? We need a deadly radiation mod. MOAR TUMORS)

KSPI prevents you running certain engines if there's a nearby Kerbal on EVA. But NERVA-like designs shouldn't emit significant amounts of radiation.

[m4v]

Well, there's also the rationale "you really shouldn't be using planets with insane densities when you're teaching kids about space"--Kerbal.EDU should have the real system at least as an option.

Why? when is planet density even a variable when launching something or in space? In education the important thing is to transmit concepts and ideas, not the magnitude of numbers. There's a lot of stuff in school that is taught in a simplified way or outright wrong (how wings generate lift is one that comes to mind) for the sake of it.

[Ippo]

And you are ok with that? Seriously: why should we just surrender to mediocrity? "Other people are wrong: let's be wrong too!".

There is a difference between "simplified game mechanic for the sake of fun" and "contradicts everything we know about the universe", especially in a game about space exploration.

[Jhonny007]

So after I followed the discussion for a little while I had an idea, that lets the beginners explore the Kerbal 1.0x System, while knowing that this is not realism. 6.4x/10x should be a difficulty setting.

I suggested that here:

http://forum.kerbalspaceprogram.com/threads/92886-Make-scale-of-solar-system-a-configurable-difficulty-setting

[Red Iron Crown]

So after I followed the discussion for a little while I had an idea, that lets the beginners explore the Kerbal 1.0x System, while knowing that this is not realism. 6.4x/10x should be a difficulty setting.

I suggested that here:

http://forum.kerbalspaceprogram.com/threads/92886-Make-scale-of-solar-system-a-configurable-difficulty-setting

The devs are understandably reluctant to go with multiple systems in the stock game, a position I agree with. A rescale that allows good aero while maintaining the current dV to LKO requirement would be the best compromise IMO. And by that I mean rescaling Kerbin itself, not the rest of the system. Making the whole system larger just increases floating point error magnitude for little gain.

[ferram4]

Making the whole system larger just increases floating point error magnitude for little gain.

Except it doesn't. Floating point errors are well taken care of by the floating origin system and Krakensbane, and players of RSS do not run into more floating point errors. Hell, the higher velocities overall would result in Krakensbane being kicked in at much "lower" (relative to orbital velocity) speeds, which would reduce floating point errors overall. I should point out that Mun orbit is at velocities below the Krakensbane limit, and so the Kraken still causes issues in Munar orbit; increasing the size of the Mun as well as Kerbin would increase orbital velocity, possible bringing it above the 750 m/s needed for Krakensbane and the removal of some of those floating point errors.

[Red Iron Crown]

Except it doesn't. Floating point errors are well taken care of by the floating origin system and Krakensbane, and players of RSS do not run into more floating point errors. Hell, the higher velocities overall would result in Krakensbane being kicked in at much "lower" (relative to orbital velocity) speeds, which would reduce floating point errors overall. I should point out that Mun orbit is at velocities below the Krakensbane limit, and so the Kraken still causes issues in Munar orbit; increasing the size of the Mun as well as Kerbin would increase orbital velocity, possible bringing it above the 750 m/s needed for Krakensbane and the removal of some of those floating point errors.

It was my (possibly very flawed) understanding that there are still floating point rounding errors when plotting a large eccentric orbit due to the distances involved, which I thought would be exacerbated by a larger scale. I wasn't so much referring to the Kraken break-your-ship-apart bug, which I agree has been dealt with satisfactorily by Krakensbane. Interesting to know that a larger scale would actually reduce the Kraken issue in some cases.

[ferram4]

Most of those floating point errors are simply due to the fact that the shape of an elliptical orbit can be greatly affected by tiny changes in velocity at periapsis. Well, once Krakensbane kicks in, there are only two sources of changes to velocity if the engine is shut off: any errors due to failure of conservation of momentum due to rotation, and errors in the simulation itself. With Krakensbane active, the latter are a constant size, which means that higher velocities will reduce floating point errors due to that. Errors due to rotation are also relatively constant, since higher orbital velocities don't magically make rockets spin faster in space. So then that would result in lower floating point errors as well. Note that all those statements are relative to the absolute magnitude of the exact velocity magnitudes.

The ship-breaks-apart bug was not the original bug to be fixed by Krakensbane. It was floating point errors due to loss of precision in velocity, which would cause phantom forces on the vehicle, whihc always cause you to lose control long before the rocket could be destroyed.

Anyway, all of that said, in time warp the only significant numerical errors are in crossing SOIs, simply because the numerics required to solve true anomaly of the orbit (angular position relative to periapsis) as a function of time are relatively simple-ish and can be clamped every half-period. Interestingly, since a larger universe would increases the size of SOIs (especially if the distances were increased to much more realistic sizes, since they're a little too close for proper scale), the SOIs would have a larger area of lower gravity gradient, reducing errors due to jumping a little too deep into the SOI, since gravity doesn't change as significantly.

Hmm... I wasn't for scaling up the planets more than absolutely necessary to deal with aerodynamic effects, but considering those effects (unless I'm horribly wrong), it would be better to make it even larger... This requires some more thinking.

[cantab]

Scaling up - or down - the celestials in order to reduce the impact of bugs is pretty kludgy though. Better to actually fix them.

[Sky_walker]

Scaling up - or down - the celestials in order to reduce the impact of bugs is pretty kludgy though. Better to actually fix them.

You mean... like... designing 128 bit CPUs and distributing them with the game?

Yea..... that's not going to happen.

Seriously though - SQUAD would do it already if they'd know how to hammer out issues just like that. Fixing some of the bugs with scaling up the system is just a side-effect of other fixes that it introduces, not a main reason to do it. And we all know how joyful devs are each time someone mentions rescaling Karbin .

Edited September 5, 2014 by Sky_walker