Engine balancing issues in ARM?

[Amaroq]

I don't claim my metric to be best but I still believe we need to use (at least) three parameters to put new engines into context properly. So if you know better approach, fire away.

1. Lift capacity --> maximum size of payload this engine can lift to LKO with a set initial TWR (e.g., 1.2 for FAR, 1.6 for stock)

2. Lift payload fraction --> payload / launchpad weight

3. Jool transfer capacity --> size of payload this engine can deliver to Jool orbit from LKO with a set initial TWR (e.g., 0.25, 0.1)

4. Jool transfer fraction --> #3 / #3's launch weight

5. Mun transfer capacity --> payload from LKO to LMO

6. Mun transfer fraction --> #5 / #5's launch weight.

Complex, but decently achievable.

[stupid_chris]

Alright, to give everyone a feeling of what I'm saying:

I have mapped the KW Rocketry engines, as well as the NovaPunch engines. As it is possible to see, they are all closely packed together.

There was actually three exceptions. And there are some interesting facts about them:

• They were all orbital engines
  
• They were following their own curve, slightly above the others
  
• They were nearly in line with the values I've proposed for the new NASA engines.

I've removed em from the curve calculation for precision, as obviously abbering data just lowers overall accuracy.

Even by keeping those engines in the curve calculation, the new NASA engines are standing out very clearly.

Point of this is: you don't need to be overpowered to do some heavy duty lifting. KW Rocketry and NovaPunch show it very clearly and are balanced with one another and stock.

Edited April 4, 2014 by stupid_chris

[Themohawkninja]

Alright, to give everyone a feeling of what I'm saying:

The funny thing is though, at least with me, is that the new engines are the only ones that feel like real good engines. Now you can build a rocket, that looks like it could work in reality, not be 6 huge boosters around a single core, and actually get something into orbit.

You couldn't do that before unless said payload was only a couple tonnes.

These new engines allow players to put decent sized things into orbit with rockets that aren't absurdly huge, all while still requiring the need for good orbital mechanics skills.

You can throw numbers around all you want, but in the end, more powerful != overpowered.

If we are going to talk about the engines being overpowered to the point that they need to be nerfed, we need to see some proof-of-concepts that show that there is something obviously overpowered to the point that it detracts from the gameplay.

EDIT: Think of it this way:

• A realistic (few SRB's, one main core, and one insertion stage) 0.5 meter rocket can get nothing into orbit.
  
• A realistic (few SRB's, one main core, and one insertion stage) 1.25 meter rocket can get a 0.5 meter contraption into orbit.
  
• A realistic (few SRB's, one main core, and one insertion stage) 2.5 meter rocket can get a 1.25 meter contraption into into orbit.
  
• A realistic (few SRB's, one main core, and one insertion stage) 3.75 meter rocket can get a 2.5 meter contraption into orbit.
  

P.S. This goes for any game, you can mathematically make something that wins everytime, but that doesn't mean it will work in practice.

Edited April 4, 2014 by Themohawkninja

[Themohawkninja]

Reread the OP. He's complaining that an engine that is the same mass as the Mainsail has 33% more thrust. Effectively complaining about the new engine's high TWR.

Well bolding the thrust, and only the thrust is quite misleading.

[stupid_chris]

The funny thing is though, at least with me, is that the new engines are the only ones that feel like real good engines. Now you can build a rocket, that looks like it could work in reality, not be 6 huge boosters around a single core, and actually get something into orbit.

You couldn't do that before unless said payload was only a couple tonnes.

These new engines allow players to put decent sized things into orbit with rockets that aren't absurdly huge, all while still requiring the need for good orbital mechanics skills.

You can throw numbers around all you want, but in the end, more powerful != overpowered.

If we are going to talk about the engines being overpowered to the point that they need to be nerfed, we need to see some proof-of-concepts that show that there is something obviously overpowered to the point that it detracts from the gameplay.

P.S. This goes for any game, you can mathematically make something that wins everytime, but that doesn't mean it will work in practice.

I've been able to put rockets in orbit that looked like real rockets, without ever abusing asparagus, with the old stock parts. Additionally, the thing that prevents us to do the same is the atmosphere. Play with FAR an you'll realize how the atmosphere makes a ridiculous difference.

But then again, I mean all the engines are aligned together in a very clear manner, I'm not making things up there. The new engines are all way above the curve, in a very clear way. So yes, by all means, with the current balancing the current engines are overpowered. Because as you said, being overpowered is not being more powerful, because then all the 3m and 5m KW and NP parts would be overpowered. Overpowered is standing out of a balancing shceme with better stats. Which is the case here.

If it's possible to have the numbers say anything, go on. I'm very interested to see how you can arrange them to show that the engines are not overpowered, more powerful, or any other winning combination.

Well bolding the thrust, and only the thrust is quite misleading.

He bolded the thrust and the mass. Higher thrust + same mass == higher TWR.

Edited April 4, 2014 by stupid_chris

[Z-Man]

It's not a power law. It's an inverse function.

That is a special case of a power law.

How is it factually wrong? The curve he uses seems to fit the balance of the older stock engines.

Because it implies that a hypothetical engine with I_sp of 1, any thrust and zero mass is better than anything we currently have because it has infinite TWR. Because it implies the old Ion engine is useless. It does not take into account that different situations require you to weigh TWR differently, and in none of them should you be looking at TWR*I_sp exclusively. I gave a more correct formula earlier.

[Themohawkninja]

I've been able to put rockets in orbit that looked like real rockets, without ever abusing asparagus, with the old stock parts. Additionally, the thing that prevents us to do the same is the atmosphere. Play with FAR an you'll realize how the atmosphere makes a ridiculous difference.

You have conceded my point, thank you.

Overpowered is standing out of a balancing shceme with better stats.

I disagree. To me, something is overpowered if and only if the application of it shows a clear advantage over other like parts.

Think of it like a gun in an FPS game. Just because a gun does 5 times as much damage as every other gun doesn't prompt a nerf if everyone using it has approximately the same K/D rato.

If it's possible to have the numbers say anything, go on. I'm very interested to see how you can arrange them to show that the engines are not overpowered, more powerful, or any other winning combination.

You're missing my point. In my opinion, when it comes to video games, the numbers don't really matter. Look at Dungeons and Dragons for example. That game attempts to be as balanced as possible, but people still find builds that break it. In StarCraft II, a game that has professional level gameplay, people keep finding ways to break it even though the numbers might not show it. This goes for just about every game that is stat heavy, including KSP.

No matter how clear the data is, if at the end of the day, nobody can utilize them, then the numbers are meaningless.

[stupid_chris]

That is a special case of a power law.

Because it implies that a hypothetical engine with I_sp of 1, any thrust and zero mass is better than anything we currently have because it has infinite TWR. Because it implies the old Ion engine is useless. It does not take into account that different situations require you to weigh TWR differently, and in none of them should you be looking at TWR*I_sp exclusively. I gave a more correct formula earlier.

No, an engine with an ISP of 1 would not have an ISP of infinity. The lower your ISP, the larger TWR, but you cannot reach an ISP of 0, and cannot reach a TWR of 0 either.

Also, you have just discovered the difference between theory and practive. Theory describes the comportment you would expect, practice is what actually happens. However, the converging intervall here would be much larger than the widths explored in KSP, so it's safe to say that this is very similar. Also, if you look, you'll realize I haven't used the engine factor in the graph, although it is a real thing.

EDIT:

You have conceded my point, thank you.

I don't think I have. I said with /old/ stock parts. Without SLS parts.

I disagree. To me, something is overpowered if and only if the application of it shows a clear advantage over other like parts.

Think of it like a gun in an FPS game. Just because a gun does 5 times as much damage as every other gun doesn't prompt a nerf if everyone using it has approximately the same K/D rato.

And how would such a gun be balanced? By having a slower shooting rate. Of course some will be better because they are higher tier. And the values I proposed are better than th regular tendency, but by keeping some common sense. Basiaclly a higher tier that makes sense.

You're missing my point. In my opinion, when it comes to video games, the numbers don't really matter. Look at Dungeons and Dragons for example. That game attempts to be as balanced as possible, but people still find builds that break it. In StarCraft II, a game that has professional level gameplay, people keep finding ways to break it even though the numbers might not show it. This goes for just about every game that is stat heavy, including KSP.

No matter how clear the data is, if at the end of the day, nobody can utilize them, then the numbers are meaningless.

At the opposite, KSP is a game that boils down to those numbers. Of course there will always be someone to exploit them, but this is the exceptions and you can't keep them in consideration. As it is, the new engines are ridiculously easy to exploit. The KR-2L is the most easy to misuse without even trying.

Edited April 4, 2014 by stupid_chris

[vger]

I've been able to put rockets in orbit that looked like real rockets, without ever abusing asparagus, with the old stock parts. Additionally, the thing that prevents us to do the same is the atmosphere. Play with FAR an you'll realize how the atmosphere makes a ridiculous difference.

Whoa, is it actually easier to get realistic rockets designs into orbit with FAR than without?

[stupid_chris]

Whoa, is it actually easier to get realistic rockets designs into orbit with FAR than without?

With a good design, you can lower the DV to orbit all the way down to 3400m/s

[Rainbow-Dashie]

The way I see it you gain no advantage if you use them or not. We can all choose to either use the new OP engines, or refuse to use them because they ruin the fun. Just like some of us use Mods and some Don't, we've always had the ability to add slightly OP engines through mods, some of us did it because it would be fun, some of us didn't because it would ruin our fun.

I personally play Career Mode a lot and will use them, after I unlock them, so I will still be using the other older engines.

Thats the joy of KSP, you don't compete with anyone so you can make your own rules, "No Killing Kerbals", "No Probes", "No Quicksaves", "No Debris in orbit", "No Kerbal Left Behind", and so on. We all have our own set of rules that we play by, so we can enjoy the game in a way we think is best.

[Themohawkninja]

I don't think I have. I said with /old/ stock parts. Without SLS parts.

Exactly, you can do the same task with both old and new parts, hence showing that they are in-fact similar. Just as a 2.5 meter rocket can hide 1.25 parts in a fairing and put it into orbit, so to can a 3.75 meter rocket hide 2.5 meter parts and put them into orbit with the same general design, and probably get the same (scaled) performance in practice.

And how would such a gun be balanced? By having a slower shooting rate. Of course some will be better because they are higher tier. And the values I proposed are better than th regular tendency, but by keeping some common sense. Basiaclly a higher tier that makes sense.

Who ever said it had to have a lower firing rate? It could very well be equal in all other aspects except for damage, but when put into practice, the gun ends up giving players little to no benefit to their K/D ratio. This may be because the higher damage makes them feel like they can be more aggressive, but in the end are just shot to pieces. In theory, the gun should be OP, but in practice it works just fine.

At the opposite, KSP is a game that boils down to those numbers. Of course there will always be someone to exploit them, but this is the exceptions and you can't keep them in consideration. As it is, the new engines are ridiculously easy to exploit. The KR-2L is the most easy to misuse without even trying.

If, in practice, such ease of exploitation ends up coming to fruition, then I'll concede. Until then, one must keep in mind that there were testers before us that clearly thought nothing was OP about the engines, and therefore felt that nothing should be changed.

Edited April 4, 2014 by Themohawkninja

[Rakaydos]

Exactly, you can do the same task with both old and new parts, hence showing that they are in-fact similar. Just as a 2.5 meter rocket can hide 1.25 parts in a fairing and put it into orbit, so to can a 3.75 meter rocket hide 2.5 meter parts and put them into orbit with the same general design, and probably get the same (scaled) performance in practice.

Old parts cant single stage to laythe.

[Kasuha]

I did some more of my own research. First of all I checked real rocket engines which I found in this table. The result is, apart of the fact that they're completely different category (WAY more effective than what we have in KSP), they're all over the place in parameter space as well. They appear to - very remotely - follow the criteria stupid_chris used in his analysis, but whole KSP engine parameter space fits within that if shifted a bit. Besides, new engines are about halfway between "old" KSP engines and real world engines.

What's undisputable is that Mainsail is almost completely eclipsed by the new LFB KR-1x2. The only application where the Mainsail may be better is when you want to use it with less than one full orange tank of fuel. A Tylo lander might be that category but that's very marginal use.

But that's IMO the only really eclipsed engine on list. The remaining two engines are most importantly different size and designed for large payloads and if you want to transfer small payload with them you end up wasting fuel. LFB or KR-2L + one jumbo are not really better at getting things to orbit than Skipper + jumbo. Give KR-2L more fuel and it is the king, but when you want to get just small payload to orbit, Skipper is your engine of choice.

In conclusion, the new engines do not feel all that overpowered at all to me. I'm not missing Mainsail, I did not like that engine before the patch either.

Edited April 4, 2014 by Kasuha

[stupid_chris]

If, in practice, such ease of exploitation ends up coming to fruition, then I'll concede. Until then, one must keep in mind that there were testers before us that clearly thought nothing was OP about the engines, and therefore felt that nothing should be changed.

As a part of the experimentals team, I can tell you that you are not right on that side of the deal, but discussing experimentals stuff is not allowed.

Anyway, it appears to boil down to a difference in opinions at this stage. I think what has to be said has been said. I'll let you guys carry with the discussion, but be wary to keep it polite and constructive, no need to tear another person down to prove a point.

@kashua

Those engines listed are a) from very different types, you can't compare a Nerva to a jet engine to a rocket engine

from different ages, the technology evolved considerably between the F1 and the Merlind 1D

c) use different fuels, because yes, this has a large impact on performance.

I don't think this dataset can be used to conclude anything at all.

Edited April 4, 2014 by stupid_chris

[Amaroq]

Earlier I suggested "payload lift capacity" and "payload capacity to Jool" as useful metrics.

Using a launchpad weight of 1.6 TWR and a cost to orbit of 4600 dV, and a Jool boost of 0.25g, I get the following:

Engine     -> LKO    LW% -> Jool |  JW%
-----------------------------------------
LVT-30     |  29t | 0.22 |  244t | 0.28
Skipper    |  80t | 0.20 |  686t | 0.26
Mainsail   | 169t | 0.18 | 1464t | 0.24
KR-2L      | 285t | 0.18 | 2943t | 0.29
LFB-KR1x2  | 282t | 0.23 | 2199t | 0.27
S3-KS 25x5 | 451t | 0.23 | 3518t | 0.27

The issue - IMHO - is not that the three new engines can lift more than the mainsail, but that they do so with a better launchpad fraction than the LVT-30.

Those numbers don't leave the Mainsail with a particular role, not even that of "launching a 2.5m stack", as the LFB-KR1x2 is a strictly better engine due to its superior ISP, TWR, and Thrust.

stupid_chris' proposal brings those numbers down to:

Engine     -> LKO    LW% -> Jool |  JW%
-----------------------------------------
Mainsail   | 169t | 0.18 | 1464t | 0.24
KR-2L      | 126t | 0.14 | 1563t | 0.28
LFB-KR1x2  | 192t | 0.19 | 1661t | 0.26
S3-KS 25x5 | 318t | 0.16 | 3194t | 0.25

which at least makes the Mainsail more competitive but still leaves the

LFB-KR1x2 a strictly-better replacement for similar loads.

If you want an interesting choice, the LFB-KR1x2 should slot in somewhere between the Skipper and the Mainsail.

Thrust of 1050, ISP(ASL) of 290 and ISP(VAC) of 340 would accomplish that, giving it a 17.845 TWR, and slotting it in as follows:

Engine     -> LKO    LW% -> Jool |  JW%
-----------------------------------------
Skipper    |  80t | 0.20 |  686t | 0.26
LFB-KR1x2  | 123t | 0.19 | 1067t | 0.25
Mainsail   | 169t | 0.18 | 1464t | 0.24

I'd also gut the S3-KS 25x5's ISP(VAC), to make it very fuel-inefficient if one tries to single-stage-to-Laythe, for example.

[Rakaydos]

If you want an interesting choice, the LFB-KR1x2 should slot in somewhere between the Skipper and the Mainsail.

I'd actually argue for an inversion- give the KR1x2 the mainsail+orangetank's stats completely, and making the mainsail the new super-skipper. The built in orange tank is less of a drawbacck for the more powerful engine stats, and this leaves the superskipper usable in upper stages with a bottom attach node.

[Amaroq]

That's .. actually a very good idea. I like it!

.

...

I mean, Rakaydos' idea, not replying to myself, of course.

[Rakaydos]

There's still the matter of the existing underpowered engines, though. the 909, the Rocomax 55, and the Poodle. If we're going for niche-filling, how should we differentiate them?

The 909 is an early lander engine, so it should both be middle of the road- able to poorly fill other niches by clustering, like the other lvl engines- and high ISP- to give newbies with landers more hover time on less fuel, in space or on kerbin. Call it, just enough thrust to lift an FL400+itself in kerbin gravity, a high ISP, with a so-so TWR and a high gimbal

The Poodle is another high gimbal lander engine that is simply overshadowed by it's contemporaries. It should probably swap TWR with the Rocomax 48 probe engine and be specialized for vacuum- Aerospike being the atmo-specialized engine.

The 55 is awkward- a radial engine too bulky to fit between boosters. I want to call it a radial aerospike- an efficent lander engine on any world.

[Z-Man]

No, an engine with an ISP of 1 would not have an ISP of infinity.

I said TWR infinite. And I did not mean an engine balanced according to your formula. I meant a hypothetical engine clearly overpowered by your formula, but not really.

Mind you, I just disagree with your method. Some of the conclusions are right. The new engines are more powerful as lifters of heavy stuff than anything we had before.

[Rakaydos]

I said TWR infinite. And I did not mean an engine balanced according to your formula. I meant a hypothetical engine clearly overpowered by your formula, but not really.

Mind you, I just disagree with your method. Some of the conclusions are right. The new engines are more powerful as lifters of heavy stuff than anything we had before.

A hypothetical massless engine?

A better example of your point would be a pure electric propulsion, like a Quantum Thruster, with a MWR (including needed power sources) approaching 0... and an actually infinite ISP.

But being a lifter engine isnt the problem. Being a high TWR lifter engine AND a high ISP space engine sounds awesome for moving high mass asteroids around! But then what do your other engines do? it's the same probem with the Rocomax 48, writ large.

[stupid_chris]

I said TWR infinite. And I did not mean an engine balanced according to your formula. I meant a hypothetical engine clearly overpowered by your formula, but not really.

Mind you, I just disagree with your method. Some of the conclusions are right. The new engines are more powerful as lifters of heavy stuff than anything we had before.

That was a typo, you can replace the second ISP for TWR. If you really want to explore those infinity solutions, the engine would have a TWR of very near infinity, and an ISP of very near 0. This means it would burn all it's fuel instantly but give an infinite amount of thrust. You quickly get a 0*∞ classic limit problem and you'll realize that the solution isn't as easy as it seems. But that is just the pure maths side

[Red Iron Crown]

But being a lifter engine isnt the problem. Being a high TWR lifter engine AND a high ISP space engine sounds awesome for moving high mass asteroids around! But then what do your other engines do? it's the same probem with the Rocomax 48, writ large.

That was very well put, sums up the problem nicely.

[Themohawkninja]

Old parts cant single stage to laythe.

Actually, yes they can. There is a video of a single stage to Laythe on YouTube using nothing more than a turbojet and some Ion engines.

EDIT: Okay, that was an SSTO to Eve that didn't use rocket engines. Still, it is possible. Just look up "SSTO to Laythe" on YouTube, and have fun.

Edited April 5, 2014 by Themohawkninja

[Rakaydos]

Actually, yes they can. There is a video of a single stage to Laythe on YouTube using nothing more than a turbojet and some Ion engines.

EDIT: Okay, that was an SSTO to Eve that didn't use rocket engines. Still, it is possible. Just look up "SSTO to Laythe" on YouTube, and have fun.

Exactly. 2 different engine types, each one extremely specialized. Not one engine to rule them all. (and into the darkness fly them. )

Edited April 5, 2014 by Rakaydos