How should starmade generate thrust?

[Sven_The_Slayer]

With the recent update we saw changes to the thrust system. At first I thought this was a good changed because it seemed to favor large clumps of thrusters over the checkerboard array I had been using. I personally feel that when it comes to generating thrust the rule of the land should be the bigger more solid the engine the more thrust you will get out of it. The downside to this being it does take some clever design for more efficiency off the table. With the current system this seemed how it worked. I was able to make large clumps of thrusters that were more efficient than my previous arrays. This made for better ships, less space spent on engine arrays meant I could pack more power and shields into the same space getting and increase in all stats.

Then the problem has come up when I dove deeper into the thrust mechanics to see what the best options were and the results were shocking

This is two ships just made out of thrusters, based on how you build it you can have a widely varied result. Even creating arrays of smaller thrust packs is pointless when your gains are this significant by expanding a single thruster.

Additional test "ships" built so far

 

[Sven_The_Slayer]

Dang, I forgot to add an option to the poll for "other, explain below" and you cannot edit polls aparently...

 

[Winterhome]

Personally, I'd prefer if it were a combination.
The complexity of the engine should increase thrust output, but not as much as pure quantity, resulting in the need for both blocking and complex thruster design.

 

[Majorfatboy]

As it used to be was best ( I think).
Under the old system, you had a trade-off. Want big speed/thrust? Then You're going to have to have a large, bulky, heavy, power-hungry thruster to do it. And said thruster's dimensions ate up so much room, that You were forced to cut back on other things, like weaponry, shields, internal space, Etc. Simply making the ship bigger wouldn't help, as that would also add to your mass, requiring an even bigger thruster. If you wanted better weapons, shielding, Etc, while still maintaining your ship's mass, then You'd have to cut Your thruster down.
In other words, the old system was more sensible: Tank/Big gun ships were powerful attackers, but were also slower; High-speed ships were fast, but had to pick their fights, or run from a fight all together.

Now, any ship, and I mean any ship, can have it all: Massive destructive potential, heavy shielding, craploads of internal space (well protected internal hangers), and enough thrust to chase down any other ship in the game.

I spent a few days updating every ship I have to the new system. It's a joke. I was able to maintain or far exceed the same thrust, but ended up with so much empty internal space I was able to practically double firepower, shielding, additional weapons, or all of the above.
I gained everything and lost nothing.

In single player this isn't that big of a problem, as I can build and balance what the AI baddies get, and build every ship off of My own technology base, but I blanch at the thought of what may be going on in multiplayer right now......
The munchkins are probably having a field day.........

 

[Lt_Duckweed]

As it used to be was best ( I think).
Under the old system, you had a trade-off. Want big speed/thrust? Then You're going to have to have a large, bulky, heavy, power-hungry thruster to do it. And said thruster's dimensions ate up so much room, that You were forced to cut back on other things, like weaponry, shields, internal space, Etc. Simply making the ship bigger wouldn't help, as that would also add to your mass, requiring an even bigger thruster. If you wanted better weapons, shielding, Etc, while still maintaining your ship's mass, then You'd have to cut Your thruster down.
In other words, the old system was more sensible: Tank/Big gun ships were powerful attackers, but were also slower; High-speed ships were fast, but had to pick their fights, or run from a fight all together.

Now, any ship, and I mean any ship, can have it all: Massive destructive potential, heavy shielding, craploads of internal space (well protected internal hangers), and enough thrust to chase down any other ship in the game.

I spent a few days updating every ship I have to the new system. It's a joke. I was able to maintain or far exceed the same thrust, but ended up with so much empty internal space I was able to practically double firepower, shielding, additional weapons, or all of the above.
I gained everything and lost nothing.

In single player this isn't that big of a problem, as I can build and balance what the AI baddies get, and build every ship off of My own technology base, but I blanch at the thought of what may be going on in multiplayer right now......
The munchkins are probably having a field day.........

Even with the old system, I could make fighters with op shields, weapons and thrust for their size/mass. It's all about how you set up your systems. My fighters generally have a 4-1 or 5-1 thrust mass ratio, and no power problems at all. The new update just makes it easier.

 

[Lt_Duckweed]

Also Sven, you asked I deliver.



The old equation for thrust was thus;

( x+y+z-3) + (n^1.125)
Where n is the number of blocks



The new equation is this:

[(x-1)(y-1)(z-1)]+(n^1.125)
Where n is the number of blocks


What this means is a 3d cross will give a vast amount of power, for example, a 101*101*101 cross will give

(101-1)(101-1)(101-1)+(301^1.125) = 1,000,000+614.3 = 1,000,614.3 thrust. out of 301 blocks.

That comes out to 3324.3 thrust per thruster block. Under the old system the same configuration would have given 914.3 thrust, for 3.04 thrust per thruster block. This is an amplification of 1093.5 times the old thrust. And it only gets worse the larger you make the cross. At 201*201*201 you get over 8,000,000 thrust.

Personally though, I really like this system, your still limited by your power regeneration but it leaves open some possibilities for unique designs.

 

[Gamer_Z]

I like the new system, no need for mega massive thrusters, but what I would like is that the *direction* of the thruster would have significant impact (hint: faster turning, massive ships take ages to turn)

 

[Lt_Duckweed]

Just thought of something, if you have a 101*101*101 cross of thrusters, and the center gets shot out, you will suddenly go from over a million thrust to less than 1000 thrust. Meaning one lucky shot can cripple a poorly designed dreadnought .

 

[Sven_The_Slayer]

Just thought of something, if you have a 101*101*101 cross of thrusters, and the center gets shot out, you will suddenly go from over a million thrust to less than 1000 thrust. Meaning one lucky shot can cripple a poorly designed dreadnought .

Yeah I had fully considered that when I was toiling this in my head. It is a huge potential weak point to this system but still easily avoided. Even a reinforced thruster with thicker arms, or multiple arms will still be capable of extreme thrusts with not a lot of resources when compared to the alternatives.

Does anybody have an idea or equation that will favor large, dense reactors while still allowing for a unique design element to sway the efficiency of the engine? I really like the idea that smart design gets you better ships while still wanting/supporting big fat engines to push big fat ships around. The current single large thruster being the best option just feels like cheating, like it's an unintended side effect of the system.

 

[der_scheme]

Does anybody have an idea or equation that will favor large, dense reactors while still allowing for a unique design element to sway the efficiency of the engine? I really like the idea that smart design gets you better ships while still wanting/supporting big fat engines to push big fat ships around. The current single large thruster being the best option just feels like cheating, like it's an unintended side effect of the system.

The answer depends solely on the type of layout we want to support.
Let's use Star Wars as a basis, shall we? There we have sublight and hyperdrive engines, that look and work entirely different.

• Sublight engines: They usually have a complex layout of twisted pipes and ignition stuff and what not. We could use this to reward engine layouts that are essentially one long, twisted line of thruster blocks.
  
• Hyperdrive: It seems to me that hyperdrives are composed of layered cuboids on the inside. For instance, we could therefore reward engine layouts that are composed of large planes that are connected in a certain way.
  

That book (Complete Cross Sections) is awesome, btw.

I prefer the first one, for several reasons.

1. It is easier to support small, as well as large ships.
2. I believe it to be easier to understand, but harder to build such engines.
3. Creativity! There could be a sheer infinite amount of shapes that are equally efficient.

I came up with the following:
thrust(group) = (volume(group) / surface_area(group) * (\sum_{block \in group} (1 / |neighborhood(block)|)²))²
Explanations:

• thrust(group): The thrust generated by the given group.
• volume(group): The volume of the group's bounding box.
• surface_area(group): The surface area of the group's bounding box.
• neighborhood(block): This could be either the blocks directly adjacent to the block, or the ones directly and diagonally adjacent to the block. I leave undefined if we count all block types or just thruster blocks.
  
• |neighborhood(block)|: The size of the neighborhood.
• \sum...: We iterate over all blocks in the group, assign a value to each and compute the sum of these values. The less blocks a block has in its neighborhood, the higher the value (max 1).
• All: We penalize lines, as well as solid thruster cuboids. We reward a complex layout with twists and dead ends. (I hope. :D).

Because I don't have detailed knowledge about SM's internals, I can't estimate how that would impact exact computation times. The asymptotical complexity should be the same, though.

 

[Elit3]

The current system is slightly ridiculous as is the power system. I think they should all take a page out of shield blocks book.

 

[Sven_The_Slayer]

The answer depends solely on the type of layout we want to support.
Let's use Star Wars as a basis, shall we? There we have sublight and hyperdrive engines, that look and work entirely different.

• Sublight engines: They usually have a complex layout of twisted pipes and ignition stuff and what not. We could use this to reward engine layouts that are essentially one long, twisted line of thruster blocks.
  
• Hyperdrive: It seems to me that hyperdrives are composed of layered cuboids on the inside. For instance, we could therefore reward engine layouts that are composed of large planes that are connected in a certain way.
  

That book (Complete Cross Sections) is awesome, btw.

I prefer the first one, for several reasons.

1. It is easier to support small, as well as large ships.
2. I believe it to be easier to understand, but harder to build such engines.
3. Creativity! There could be a sheer infinite amount of shapes that are equally efficient.

I came up with the following:
thrust(group) = (volume(group) / surface_area(group) * (\sum_{block \in group} (1 / |neighborhood(block)|)²))²
Explanations:

• thrust(group): The thrust generated by the given group.
• volume(group): The volume of the group's bounding box.
• surface_area(group): The surface area of the group's bounding box.
• neighborhood(block): This could be either the blocks directly adjacent to the block, or the ones directly and diagonally adjacent to the block. I leave undefined if we count all block types or just thruster blocks.
  
• |neighborhood(block)|: The size of the neighborhood.
• \sum...: We iterate over all blocks in the group, assign a value to each and compute the sum of these values. The less blocks a block has in its neighborhood, the higher the value (max 1).
• All: We penalize lines, as well as solid thruster cuboids. We reward a complex layout with twists and dead ends. (I hope. :D).

Because I don't have detailed knowledge about SM's internals, I can't estimate how that would impact exact computation times. The asymptotical complexity should be the same, though.

You know what, I had thought that a snake like path might be a good method for power generators where you gain each time the reactor got longer regardless of the box dimension. Infinitely complex in structure but fairly easy to understand how to build, it could be a good method for thrusters. It shouldn't punish small single line thrusters though as sometimes that's the only option for fighters.

The current system is slightly ridiculous as is the power system. I think they should all take a page out of shield blocks book.

I do like shields the way they are because they make great filler but having a complex system for thrust and power allows for creative players to build more effective ships.

 

[Elit3]

You know what, I had thought that a snake like path might be a good method for power generators where you gain each time the reactor got longer regardless of the box dimension. Infinitely complex in structure but fairly easy to understand how to build, it could be a good method for thrusters. It shouldn't punish small single line thrusters though as sometimes that's the only option for fighters.



I do like shields the way they are because they make great filler but having a complex system for thrust and power allows for creative players to build more effective ships.

I guess, but it just seems a bit unfair on noobs

 

[Sven_The_Slayer]

I guess, but it just seems a bit unfair on noobs

Everybody is a noob at some point, then they learn how the game works and are no longer noobs (well most of them anyway) games shouldn't target the lowest common denominator.

 

[Elit3]

Everybody is a noob at some point, then they learn how the game works and are no longer noobs (well most of them anyway) games shouldn't target the lowest common denominator.

It just means Starmade has an extremely steep learning curve.

 

[PreparationsA-G]

Finding It Hard to Believe This is Not a Bug

Everybody seems to be just assuming that this IS an intended part of the latest update, but everyone seems to have forgotten that when first asked about it in the comments of the first release candidate, Schema said:

I didn't really change anything purposefully with thrust.

And (from what he wrote) he didn't really seem to know what was causing the (dramatic) thrust calculation changes, indicating that he didn't intentionally program them into the game (suggesting that they were probably a BUG).
(Original Post: http://oldsite.star-made.org/comment/98489#comment-98489)

Everyone also seems to have forgotten that these new "changes" in thrust calculation were not only not included in the full changelog for the release of 0.14, but that the word "thrust" never even appears once in the entirety of the changelog (not even anything containing the word "thrust").
(Check for yourself: http://oldsite.star-made.org/content/starmade-014-release)

Instead everyone seems to have jumped the gun and thrown themselves into the solid belief that this is definitely an intended change, simply because of how much less they have to try to make their ships incredible, and how much easier it is to have everything, whilst sacrificing nothing.

However, everyone seems to have forgotten that if EVERYONE is incredible, then NO ONE is incredible. A ship that "has everything" in a universe of ships that consistently "have everything" will have nothing more than the next ship, and therefore through comparison will have nothing. This "change" will eliminate diversity of ships, and will further the problem of all ships being able to do pretty much everything, eliminating the need for specialized ships (which I thought was something we did want).

It seems everyone has just assumed this is "working as intended," but from what I've seen the evidence seems to point the other way. Does anybody actually know (as in they are positive) that this is an intended change? Because I still haven't seen Schema come out and say "Oops, I forgot to mention I'm changing the thruster calculation to make it easier to get more thrust with fewer thrusters," (let alone give an explanation as to why or anything like that). It seems to me this is a pretty significant game mechanic change, and I would think that if this was in fact intentional he might have made a special post about it, (or at least mentioned it... somewhere) especially after the community voiced their concerns with it.

So, can we all maybe take a step back? As far as I can tell we don't even know if this was on purpose yet, and I personally would love not to have to refit all of my ships for something that ends up being a bug.

-Prep :D

 

[der_scheme]

You know what, I had thought that a snake like path might be a good method for power generators where you gain each time the reactor got longer regardless of the box dimension.

Happy to see we're on the same page here.
However, I'm not sure if my proposal wouldn't favor designs like this:



That is essentially a crossbreed between a snake and a gear rack. :D But then I was thinking—would that be so bad? However, we probably would need field tests to know what really is more effective.

 

[Cirdanoth]

Just a note... While you can cut down on the number of thrusters, to have a large thrust output, it isn't actually more practical. Because the thrust still takes the same amount of e/s. And potential e/s generation has been decreased due to docked batteries no longer being a feasible source.
While your engines are smaller.. They still take massive amounts of energy, for high thrust. A 20K mass ship will quite quickly drain its entire energy storage at 45K thrust.

We will find ourselves balancing energy output and storage much more, against thrust. We have a lil' more room for both, now that the thrusters themselves can be build smaller.

 

[CyberTao]

Finding It Hard to Believe This is Not a Bug

Everybody seems to be just assuming that this IS an intended part of the latest update, but everyone seems to have forgotten that when first asked about it in the comments of the first release candidate, Schema said:

And (from what he wrote) he didn't really seem to know what was causing the (dramatic) thrust calculation changes, indicating that he didn't intentionally program them into the game (suggesting that they were probably a BUG).
(Original Post: http://oldsite.star-made.org/comment/98489#comment-98489)

Everyone also seems to have forgotten that these new "changes" in thrust calculation were not only not included in the full changelog for the release of 0.14, but that the word "thrust" never even appears once in the entirety of the changelog (not even anything containing the word "thrust").
(Check for yourself: http://oldsite.star-made.org/content/starmade-014-release)

Instead everyone seems to have jumped the gun and thrown themselves into the solid belief that this is definitely an intended change, simply because of how much less they have to try to make their ships incredible, and how much easier it is to have everything, whilst sacrificing nothing.

However, everyone seems to have forgotten that if EVERYONE is incredible, then NO ONE is incredible. A ship that "has everything" in a universe of ships that consistently "have everything" will have nothing more than the next ship, and therefore through comparison will have nothing. This "change" will eliminate diversity of ships, and will further the problem of all ships being able to do pretty much everything, eliminating the need for specialized ships (which I thought was something we did want).

It seems everyone has just assumed this is "working as intended," but from what I've seen the evidence seems to point the other way. Does anybody actually know (as in they are positive) that this is an intended change? Because I still haven't seen Schema come out and say "Oops, I forgot to mention I'm changing the thruster calculation to make it easier to get more thrust with fewer thrusters," (let alone give an explanation as to why or anything like that). It seems to me this is a pretty significant game mechanic change, and I would think that if this was in fact intentional he might have made a special post about it, (or at least mentioned it... somewhere) especially after the community voiced their concerns with it.

So, can we all maybe take a step back? As far as I can tell we don't even know if this was on purpose yet, and I personally would love not to have to refit all of my ships for something that ends up being a bug.

-Prep :D

Well I believe it was Said that the OLD Thruster calculations were a bug :u and this is the corrected Equation that Schema wanted. That being said, I think the new system is a little Too OP Thrust wise, even if Turret mass is not accounted for, on the grounds that not all ships carry many Turrets :u

Edit; Forgot the link - http://oldsite.star-made.org/comment/101815#comment-101815

 

[Sven_The_Slayer]

Finding It Hard to Believe This is Not a Bug

Everybody seems to be just assuming that this IS an intended part of the latest update, but everyone seems to have forgotten that when first asked about it in the comments of the first release candidate, Schema said:

And (from what he wrote) he didn't really seem to know what was causing the (dramatic) thrust calculation changes, indicating that he didn't intentionally program them into the game (suggesting that they were probably a BUG).
(Original Post: http://oldsite.star-made.org/comment/98489#comment-98489)

I hope I'm not being lumped in with the people assuming it's an intended feature. I was the one who asked and got that response. It is odd that there is no mention of thrust changes in the changelog but I do believe the changes were made in order to correct a previous mistake. I personally think the effects of the new equation is all a huge oversight but we really won't know unless we hear from Schema or another Shine member. The best we can do is keep the discussion going so it gets noticed.

As for needing more power, because you can achieve large thrust without needing massive thrusters you have a lot more room to cram power modules in. The ships I've rebuilt so far I've replaced with big clumps of thrusters and even then they still take less room than my old setups which could be the intended idea behind the new equation.