Acceleration curves

[Marvin]

I don't suppose you could state the change you want in just one or two sentences.

[Maarschalk]

LOL...and I thought I was relatively a muchtalk on relativity.....:P:P:P:P

[Aures]

Sure Marvin, see the underlined bit. My primary suggestion has two parts and is quite simple. I can give each its own sentence. The acceleration curve in Mercenary is a lot like something out of relativity. It is not like relativity because there is no actual speed limit and the curve stops you at speeds well below the maximum you can achieve.

So here is my primary suggestion:

If a speed limit was introduced the acceleration curve could convincingly model relativistic flight on a fun game scale level. To be really consistent the jump mechanic should also be tweaked so the time it takes to hit light speed and jump depends on your existing velocity.

I have several arguments for tweaking the acceleration curve:

The change to the acceleration curve would not be that drastic and would not really affect lower speeds.

The change to the acceleration curve is probably quite easy to implement since it is presumably defined explicitly by an equation in code.

Lots of people already assume the acceleration curve is something to do with relativistic effects. But really it is just an arbitrary limitation. This game is largely about the lack of arbitrary limits. Recast this limit in terms of relativity and it becomes the freedom to kiss the speed of light.

It would act as a selling point for the game to replace the rather disingenuous "no speed limit" that is currently used. It would place Evochron a quantum leap ahead of other space sims in terms of realism. I believe it would greatly add to immersion for many people, not just nut cases like me.


It is not essential that the jump mechanic be changed at all for the acceleration curve tweak to be implemented. But, I think it would be a good idea that complements the other change. I also have several arguments for tweaking the jump mechanic:

The jump mechanic currently suffers some issues with regard to high lateral/vertical velocities. They are of the same type as seen with high speeds in Legends. If it was worth fixing those it is worth fixing this. The work required to fix it affects the same area of the code as would be changed in my suggestion and the amount of work involved is probably not dissimilar.

It is not entirely clear what the fix for the above issue should be in the absence of any constraints. My idea provides clear guidelines for how it should be modified.

It introduces a minor gameplay mechanic where repeated jumps are slightly faster at higher lateral/vertical velocities. It is not enough to upset the current dynamic and would not affect MP balance. But, it would have a noticeable effect of speeding up long journeys slightly. The autopilot already gradually drifts to higher lateral/vertical velocities during long voyages so the game would make use of the new system without any effort required on the players part.

[Nigel_Strange]

Aures,

I just wanted to point out a couple of things.

First, you would probably enjoy Terminus, if you can find it. The space flight is fairly Newtonian up to a point, but there is a speed at which the ship begins to take damage from further acceleration. I forgot how they rationalized that. You could get frame modifications that would let you get up to higher speeds, but you still topped out. I think the fastest ship I had got 1.2 kmpss units of acceleration (most got around 500 mpss or so). However, once you started taking damage, you needed to slow down. Maybe it was running into particles or something. It mentioned something about hull stress, which really only makes sense for acceleration, not velocity, though that might have been a limiting factor on acceleration. I don't really remember. It's a tinkerer's game. You can build a ship from scratch if you know what you're doing (e.g. don't forget to shove life support somewhere).

The other game nobody mentioned is Independence War 2: Edge of Chaos. It modeled the Newtonian flight really well, I thought. You could get up to very high speeds, though space, being vast, would not surrender to any travel plans that did not involve FTL travel, which is realistic. The combat in that game was also very good. The main drawback was that it was an open-ended, but nevertheless linear story. Once you finished it, that was the end. There were mods added later to allow more sandbox play, but this post-storyline sandbox did not really have the same level of interaction as Evochron (no missions, no mining, etc. only fighting and trading). Still, you might enjoy it.

As for "fixing" the acceleration curve, that's a matter for Vice to weigh. These kinds of changes are largely a matter for him to consider against overall popularity and playability. I would be in favor of a "realism setting" or "difficulty setting" or something that would allow me to adjust the speed curve in SP, if not in MP, owing to everyone having to inhabit the same 'verse.

[Marvin]

What if I said the speedometer maxes out at 10,000 but the jump speed is infinitely higher? (Like how speedometers in the 1970s were set to 55 mph but my old Opel GT could hit 120.)

Such that, even at maximum afterburner, you're only moving at less than 1% of jump speed? If that were the case, could you tell the difference between the time it takes to go from max AB to jump speed and the time it takes to go from max throttle to jump speed?

[Aures]

Thanks Nigel, I'm sure I would enjoy both of them.

Marvin, thank you for your comment. I felt it did did not accurately describe how the jump drive works in Mercenary. But, during the course of investigating that theory I have found some things that alter the jump drive part of my suggestion. So again, thank you for taking the time to engage me in my speculations. Your feedback has been very helpful.

The speed you reach before the jump animation ends is not known exactly. But, you can roughly estimate it by looking at the way your velocity increases/decreases during a jump. It does not involve speeds that are even an order of magnitude more than 10k from what I have seen. As far as I can tell your speed increases linearly from your current forward velocity to about 12k more than that. When you pop out of the jump your forward velocity is just above 10k and you linearly decelerate to certain forward velocity determined by what you are flying. So if you are suggesting jump speed is much greater than the limiting speeds in the acceleration curve then I don't think that holds water. There is no fixed jump speed in the current mechanic, it is the vector addition of whatever your current forward velocity is plus about 12k with your lateral and vertical velocity.

I initially thought the effect on the jump mechanic would be tiny and I think it is desirable it is tiny to avoid disturbing gameplay. Based on the above test it would fundamentally change the jump mechanic and have a very noticeable impact. I came up with the shortening of jump times as beneficial side effect of putting everything on a firmer footing. It naturally fell out of my proposed changes rather than being a goal in and of itself. The idea was you wouldn't really notice it per jump, but you would find if you timed a trip to Sol from Orion on autopilot before and after it was implemented it would take a fraction of a minute to several minutes less. I have now abandoned that idea and decided to come at it from a different angle.

A simple alternative to making the jump animation time variable is to leave its length fixed but make the rate of acceleration depend on your initial velocity. As the jump “tube� is an actual model that is spawned when you initiate jump that then tracks you this could be quite a bit easier in terms of coding.

One of the most minimal ways you could create a system like I am advocating for is to leave the part of the acceleration curve I plotted totally unchanged and make the rest asymptotic to some maximum value (maybe a bit over 8000, whatever value is chosen might as well be referred to as c) rather than always very slowly increasing and tweak the acceleration during a jump so you jump at that velocity. You could just reduce the acceleration rate so that whatever your velocity it takes as long to get up to c as it currently does to add the 12k or so to your forward velocity.

But maybe there is room for messing with jump times as separate issue. Has anyone else noticed how much longer jumping takes in Mercenary compared to Legends? It hit me in the face the first time jumped in Mercenary. I haven't got around to quantifying it before now. A Jump in Legends took a little less than 5 seconds. Jumps in Mercenary take almost double that. Multiple autopilot jumps are only a little slower in Mercenary than in Legends because the energy recharge time is very similar/identical and you start recharging as soon as the jump exit animation starts. But still, craft in Mercenary are again disadvantaged compared to Legends equivalents.

I understand and appreciate the beefed up jump animation and greater speed ramp up in Mercenary, but I don't remember thinking jumps happened too quickly in Legends. Was there any actual gameplay related reason for nearly doubling the jump time or was it purely aesthetic? Maybe there is some room there to make some changes, even without considering the other ideas I am raising.

I am back to not really having a firm suggestion. I am just spit-balling and trying to work out what if any improvements could be made to the current system without turning it upside down. Really there are a whole bunch of different suggestions than I am trying to open up discussion on. Maybe the shape of the acceleration curve should be altered, maybe it should stay almost exactly as is. There seems to be some support for a minor rebalance of the curve so it is not quite as restrictive (so you hit the knee of the curve at say 8k-9k rather than 6k-7k for example) quite apart from any high faluting talk about immersiveness and self-consistency. Then there is the amount of time that it takes to jump. Maybe the animation in Legends was way too quick and it is perfect now, maybe it should be trimmed back a bit.

I have also thought more about the high lateral/vertical velocity during jump issue I keep alluding to. I now believe the engine has no trouble with this and that it is probably something deliberately introduced by Vice mainly to stop the autopilot accelerating you to unreasonably large lateral/vertical velocities during long journeys. I still think there is still room for polishing that system too.

The core of my suggestion is still that the jump drive accelerates you to light speed as part of the jump process and that the acceleration curve is caused by the effects of relativity. I haven't seen any explanations of how either work that would contradict that, nor have I seen or can think of any explanations of how they work and why they work that way that make more sense. The current mechanics do not quite back up that explanation but the changes required are not gigantic. I think we can kill a couple of birds with one stone depending what behaviour is deemed desirable and that is why those other issues are relevant to this discussion. If my core idea goes nowhere and we decide we are happy with the acceleration curve without even a fig leaf of handwavium I would still like to see some discussion of those other issues.

[Marvin]

Okay, let's look at this from the other end ... let's assume you've just invented the first jump drive to use a wormhole ... which is how Vice defines his jump drive:

http://www.starwraith.com/spacesims/ind ... Jump_Drive

And, after experimenting with your new invention, it behaves exactly as you see in the game. In other words, its behavior violates your initial assumption as to how wormhole acceleration would work.

So, you have empirical data that doesn't jive with what you expected to see. How do you explain it?

[Nigel_Strange]

I imagine that a wormhole would just be an opening connecting to points in space, without any sense of space or time-travel between. It would be like a doorway, only one side of the doorway opens to a point billions of light years away. Looking at it, you would be able to see the other point in the universe, as looking upon a reflective sphere, only instead of reflecting your end of space, it is refracting images from the other end (a bit like a fiber optic with a light at one end).

So, traveling between two distant points on either end of the wormhole would be nothing more dramatic than stepping through a doorway and being a long way from where you were, like stepping from the living room to the dining room. Another way to imagine it is if you have two monitors hooked up to your computer. When your mouse pointer leaves one monitor and goes to the other, there is no space in between. It does not matter how far away the monitors are from one another because there is no actual "space" between them from the pointer's POV.

So much for imagination...that's just how I see it.

Now, back to the way it is implemented here. The graphics in Mercenary are more complex than in Legends. There are a lot of procedurally generated textures. In other words, the end-point of the jump must be generated when you get there. Hence, the length of time to make the jump is a bit longer: the universe needs longer to be generated.

[Aures]

Thanks Marvin, I can accept that.

But, there are problems with just accelerating to velocities greater than light before you cross the event horizon:

"Eventually, the increasing force of the event horizon and its gravity brings the ship to an extremely high velocity, fast enough for the ship to safely pass through the fulcrum point. When the ship reaches the event horizon at the required speed, it triggers the folding of space between the two points, the ship passes through the event horizon and arrives at the destination location."

If the acceleration curve is nothing to do with relativity or the speed of light then it has no lore justification I am aware of. My description of the jump drive almost exactly mirrors the above, but obviously the velocity is just "extremely high" rather than an actual punching of the light speed barrier. Hitting light speed is problematic anyway so the quoted explanation works better.

My point still stands that everything can be wrapped up nicely with a little bow if the acceleration curve reflects relativity and velocities achieved before you pass through the event horizon are not greater than the light speed limit implied by the curve.

The above explanation actually works very well with my earlier suggestion that it takes less time to get to “an extremely high velocity� if you are already at a high velocity. But it sounds like lateral and vertical velocities wouldn't contribute. So there are a bunch of different ways the acceleration curve can be reconciled with the jump mechanic that fit within current lore.

Nigel_Strange, thanks. It makes sense that a technical reason would demand a longer jump time. If that is the case then obviously it needs to be left as is.

[Marvin]

I have a feeling you're still thinking about this in much the same way as it was once thought that all swans were white. I'm sure swan-ologists had very good theories for why it was true. But you need to look at how a jump works in Mercenary as if you've just seen your first black swan. And come up with a counter theory that explains it ... one that might otherwise go against all you've learned about time, space, and Maars' favorite word.

[Maarschalk]

LOL.....Marv....relativity is relatively not a 4 letter word, so no censorship required......:P:P:P:P

[Aures]

Yes I could just accept the current mechanics and try to come up with appropriate handwavium. I don't think we are there yet and I don't think it would be good quality handwavium.

If everything is happening well below light speed then the current jump mechanic is fine. It would then be a matter of explaining what the acceleration curve is in lore terms. AFAIK there has been no attempt to do so and it would be up to Vice rather than me.

And in that case I would focus on whether the acceleration curve could be rebalanced a little purely from a gameplay perspective.

[Edited on 8-10-2010 by Aures]

[Aesir Rising]

You can try working the model backwards, and arrive at a maximum optimal speed for two humans to fight one another.

For example, a worst-case situation with respect to human reaction times is a head-on merge (180 degree approach for combatants).

Given a range of 700m for a class 1 particle cannon, and a rough estimate of human reaction times (let's say ~0.4 seconds to allow for some error-correction during a dogfight), what is the max closing speed to fire off a shot?

If I'm thinking of this correctly, it would be just 1750m/s closing speed. That is a far cry from the 100k m/s described in the top post. Someone better at math than I can punch in values for different closing angles and arrive at a practical ship speed for those angles.

I personally wouldn't like that type of Boom & Zoom combat that such speeds would require.

[Edited on 10-12-2010 by Aesir Rising]

[Nigel_Strange]

Aures, as a physicist, maybe you can explain something to me.

In space, all motion is relative.

Light travels at constant velocity....relative to WHAT?

If two photons leave the same star in opposite directions, are they not traveling at twice the speed of light relative to one another?

[Edited on 10-12-2010 by Nigel_Strange]

[Marvin]

Originally posted by Nigel_Strange

Aures, as a physicist, maybe you can explain something to me.

If you do, please post the formula too.

[soulsacrifice]

Originally posted by Nigel_StrangeLight travels at constant velocity....relative to WHAT?

You cannot exceed the speed of light. Time slows down to compensate so I guess an observer on photon A (if that were possible) would see photon B move at the speed of light. I'm not a physicist but it's all to do with space-time as distance in space is measured not only by distance but also time.

[Aures]

The answer depends whether you are asking about in the real world or the Evochron universe. In Evochron there is an absolute frame of rest that all observers agree on so the rules of relativity don't fully apply. That is why the appropriate rules to use are some kind of light speed limited Newtonian model rather than relativity proper.

In the real world it is more helpful to think about observers moving in opposite directions at less than light speed. It can get a bit confusing to talk about the speed of a photon relative to another photon moving in a different direction.

Imagine you have two rockets that set off from the same point in different directions. They accelerate to a high fraction of light speed (relative to the starting point), say 90%. What is the velocity of one rocket relative to the other? Naively you would expect the combined velocity to be more than light speed but that is not the case. Their velocity relative to each other is about 99.45% the speed of light (see relativistic velocity addition formula).

Imagine now that we only have one rocket. Once it has reached 90% of light speed (or at any other time) it is still at rest with respect to itself. As far as it is concerned it is the starting point that is moving at 90% light speed. We can take the rockets current reference frame as being another starting point. If the rocket again accelerates to 90% light speed but with respect to the new starting point, what is its velocity with respect to the first starting point? That's right, about 99.45% of light speed.

The same is true of any two time-like (ie less than light speed) vectors. No matter how what you do, adding them always results in an answer than it less than c. But light does not move in a time-like manner, it moves in a light-like manner (time-likec). Here is a link to someone asking pretty much the same question 2 parallel rays of light travelling in opposite directions?.

Basically the answer says that the question is not properly formed because to know the velocity of one photon relative to another you have to find a frame where one is at rest and that is impossible. Let me know if that doesn't satisfy you (certainly didn't when I was still learning the ropes) and I can break out the light cone diagrams and more detailed explanations. If your willing to accept that answer at face value just remember that you could have two things heading away from you at 99.999999999...% of light speed in opposite directions but their velocity relative to each other (found by taking one as reference frame) is still less than light speed. You can have two photons moving away from you in opposite directions but you cannot take one of them as a new reference frame to measure the velocity of the other by.

The answer that they are moving away from each other at 2c is not exactly wrong but is only right in a certain sense that doesn't contradict relativity. The answer c also works, if you go ahead and take one photon as a reference frame anyway and plug the values into the velocity addition formula you get v=1+1/(1+1*1)=1=c. It is also interesting to note that if you did both the above experiments at the same time all three observers would see both photons travelling away from them at exactly light speed.

[Marvin]

My, how physics has changed since we studied the Lorentz Transformation ... nearly half a century ago. I guess the only saving grace is that, fifty years from now, all the current theories will be just as old-fashioned.

[Aures]

It hasn't changed at all, at least the bit I was explaining. The fashionable way of explaining it might be somewhat different. Marvin, if you don't mind me asking, when you learned about the Lorentz transformations and other aspects of relativity all those years ago what context was it in?

The general theory of relativity has had some theoretical developments since the 60's but the special theory isn't in essence any different. The thing I talked about earlier with mass invariance is just a change of convention and terminology, it doesn't affect the theory at all. The only substantial thing I used in my last post was the velocity addition formula and that is definitely the same.

I have a feeling I probably will have to go into more detail about the relative motion of photons. I am kind of looking forward to it. I have explained it in person to a number of people but it would be good to set it out clearly and include some pictures.

[Marvin]

Well, after scrounging up my old textbook, I then went to Wikipedia and looked up "Lorentz transformation" for comparison (old theory to new). The GIF pix at Wiki are nice but did little to clarify the issue.

The old textbook listed the following formulas:

x' = (x-vt)/sqr(1-v^2/c^2)
y' = y
z' = z
t' = (t-xv/c^2)/sqr(1-v^2/c^2)

Where:

x' is the distance measured with respect to the moving system along the direction of motion
x is the distance measured with respect to a relatively stationary system
y' and z' are measured with respect to the moving system, at right angles to x' and each other
y and z are measured with respect to the relatively stationary system
t' is the time interval clocked with respect to the moving system
t is the time interval clocked with respect to the relatively stationary system

And explained them this way:

If the velocity of the moving system (v) is small relative to the velocity of light (c), then the equations reflect the expected addition of velocities. As the magnitude of velocity approaches that of light, then the values of x' and t' change drastically.

Wiki, of course, adds generously to that short list of forumulas. Eventually, I just went back to the textbook ... satisfied with the following explanation:

Example.

A clock attached to any moving system runs at a different rhythm from a stationary clock; and a measuring rod attached to any moving system changes its length according to the velocity of the system. Specifically, the clock slows down as its velocity increases and the measuring rod shrinks in the direction of its motion.

An observer riding along with the clock and the measuring rod would not notice these changes. But a stationary observer would find that the moving clock has slowed down with respect to his stationary clock and that the moving rod has contracted with respect to his stationary units of measurement.

P.S. The quotes are really "quotes" ... in that I did some minor editing of the explanations.

[Aesir Rising]

Given:
A Top Fuel dragster is 8m long at the starting line.
At the finish line, that dragster is travelling at 500km/h

Q: To an observer taking a photograph at the finish line, how long is the dragster?

Just trying to get all this stuff into a context that I can comprehend.

[Rush]

You see it shorter. But the dragster is the same as before.

However there are paradoxes about this, that may suggest that relativity doesn't actually work as Einstein described it.

For example: a cap of length x is moving at quasi-light speed towards a hole of same depth x.
If you are near the hole, you'll see the cap shorter, and you'll say "well that cap won't touch the bottom of the hole".
If you are on the cap, you'll see the hole less deep, and you'll be sure that the cap will touch the bottom.

Can something like volume or length change if you change your position and speed? Probably not, but I don't know for sure...

(I hope the paradox is correct, but I may have said something wrong. So please correct me in this case

[Aesir Rising]

To measure, you observe. You measure anything through observation. The measurement is always with respect to the observer.

The driver (an observer) wouldn't see a difference length of the dragster because the driver is traveling at the same speed and direction as the dragster (or close enough for my question).

The camera (a different observer) taking the photo should see a difference, though perhaps in practice we don't have cameras precise enough to display that difference at such short lengths and slow speeds.

This is why I'm asking about what that formula you all talked about above would come up with about actual length differences. millimeters? micrometers? nanometers?

[Marvin]

Well, according to the same textbook, you must be moving at about 90% light speed before an object shortens to half its original length. For what that's worth.

But, what I'd like to know is why all the formulas use v^2 or c^2 ... 'cause that's acceleration, not speed. What happens if you accelerate at, say, 5 Gs for an extended period of time. Since your acceleration never even comes close to 186,000 miles per second squared, what's to prevent you from exceeding light speed?

[Maarschalk]

It does not matter how you relatively get relatively close to the speed of light the laws of relative physics does not allow matter to travel faster then the speed of light and the same laws apply once you reach close to the speed of light. It does not matter if it is by accelaration or by a steady increase at a snails pace......it just takes relatively longer to reach close to the speed of light.....