Spring RTS Engine

Hello, I want to post some ideas I read in the SupCom forum.
I think it is some ideas that you think are obviouly good from the moment you read them.








I thought of a couple things that would be useful to the Supreme Commander regarding the construction of units and buildings, and the automation thereof. The first is the ability to have a factory continuously produce units according to an abstract rule. Instead of queuing up a specified quantity of units, which will obviously not work for a game like SupCom, you instead queue up a percent. That percentage is the odds that that factory will randomly choose to build that unit. So if you have two units (A and B) each set to 50%, then the factory will produce A, B, A, B, etc. on into infinity, or until you change it. More complex ones might be A, B, C, D, and E, where A is 40%, B is 35%, C is 15%, D is 7%, and E is 3%. You get the idea.

Building on that, you might have the ability to set up different plans and projects for factories (or construction units). For example, you set up the project of a preset squad of units. Perhaps 10 Flashes and 5 Stumpies. You then queue up three of these projects. Or, you can assign this project a percentile just as if it were a unit. A different possibility is the ability to create plans, which are percentile sets which you preset, and can switch between. A factory with three plans might have one for 100% A, one plan for 40% A, 40% B, 20% C, and a third for 50% C, 50% D.

A modification to the system previously outlined which I feel is superior is to instead assign point values to different selections, as opposed to percentages. The factory can figure out what percent each item is of the total points issued to it. A factory with 99 points on A and 1 point on B is the same as 99% A and 1% B. A factory with 1 point A and 1 point B is the same as 50% A and 50% B. Adding a point to C reconstrains its output to 1/3 on each. If you were to then add a point to A, A would have 2, with B and C having 1. So A is 50%, and B and C are each 25%.

I realize I'm stretching this post out. Bear with me. So what about creating plans and projects for construction units building facilities? It can work. You outline a sequence of facilities to construct and create a plan of it. This plan can then be assigned to any construction unit. It can be assigned a priority, percent or points, which causes construction units to go do the tasks with varying probabilities. Once again, the points proves itself superior. When a building project is finished, its points are simply removed. If there were three items at 1 point each, the remaining items each have 50% priority.

As an additional building tool, the power to control the flow of resources would be extremely useful. In TA, you had no control over the amount of metal and energy used by nanolathers, other than adding or subtracting nanolathes. In SupCom, you should probably be able to declare that "I want this project to halve resource use" so that you free up resources for another project. Direct control over quantity would be excellent, of course with a maximum. A rush production with added cost might be cool, but is wholly unnecessary.

Posted by Thomcaster

Please take my disagreements as the polite and constructive critique they are meant to be, not as a flame

I thought of a couple things that would be useful to the Supreme Commander regarding the construction of units and buildings, and the automation thereof. The first is the ability to have a factory continuously produce units according to an abstract rule. Instead of queuing up a specified quantity of units, which will obviously not work for a game like SupCom, you instead queue up a percent. That percentage is the odds that that factory will randomly choose to build that unit. So if you have two units (A and B) each set to 50%, then the factory will produce A, B, A, B, etc. on into infinity, or until you change it. More complex ones might be A, B, C, D, and E, where A is 40%, B is 35%, C is 15%, D is 7%, and E is 3%. You get the idea.

Basically, you're saying, "instead of having a precise percentage of a given unit being built, to my specifications, when I've asked it to be, I want to leave that to a random number generator".

This is not a good idea.

Let's look at a typical example from NanoBlobs, where Spring's UI gives players the power to precisely regulate unit production in ways that make sense.

In mid-game NanoBlobs, players are usually attempting flank-sweeps, and are gearing up for late-game production of fortress-cracking units (i.e., Demons and SquareRooks, in the current version). A player needs to build up their total resource pool, but needs to continue to project offensive power. A great solution, purposefully included in the game design, is to mix MegaSheep (which produce resources, as well as being functional attack units) with other units in a way that gradually ramps up production, but doesn't starve them during the attrition warfare that is a central element of the game design.

For example... 4 MegaSheep + 1 TriRook, per AutoFac build-cycle = +100 Resource (Metal and Energy are merged, for all practical purposes in the game design) while giving players an offensive combo that is useful and multipurpose (for example- send the TriRooks in a flanking manuever through water, while hitting the enemy head-on with MegaSheep and spotters to blunt their thrust).

If it was randomized... it'd completely RUIN THE GAME DESIGN. Taking away precise control from players would just make the game much more flukey- sometimes I'd come out at a net resource gain, but without enough TriRook production to hit hard offensively, sometimes I'd come out at a net resource loss- bad for future production runs. Either way = bad. That's just one, small example from NanoBlobs- the ways to mix MegaSheep with other unit streams is a very, very important part of the game's total balance

Building on that, you might have the ability to set up different plans and projects for factories (or construction units). For example, you set up the project of a preset squad of units. Perhaps 10 Flashes and 5 Stumpies. You then queue up three of these projects. Or, you can assign this project a percentile just as if it were a unit. A different possibility is the ability to create plans, which are percentile sets which you preset, and can switch between. A factory with three plans might have one for 100% A, one plan for 40% A, 40% B, 20% C, and a third for 50% C, 50% D.

This just adds more statistical uncertainty on top of the previous randomness

A modification to the system previously outlined which I feel is superior is to instead assign point values to different selections, as opposed to percentages. The factory can figure out what percent each item is of the total points issued to it. A factory with 99 points on A and 1 point on B is the same as 99% A and 1% B. A factory with 1 point A and 1 point B is the same as 50% A and 50% B. Adding a point to C reconstrains its output to 1/3 on each. If you were to then add a point to A, A would have 2, with B and C having 1. So A is 50%, and B and C are each 25%.

How's about just letting players choose?

I realize I'm stretching this post out. Bear with me. So what about creating plans and projects for construction units building facilities? It can work. You outline a sequence of facilities to construct and create a plan of it. This plan can then be assigned to any construction unit. It can be assigned a priority, percent or points, which causes construction units to go do the tasks with varying probabilities. Once again, the points proves itself superior. When a building project is finished, its points are simply removed. If there were three items at 1 point each, the remaining items each have 50% priority.

Wheeeeeee... now you've taken the simple, straightforward micro of, "build three factories, give them all identical build orders, set on Repeat, feed Move-->Patrol paths" and made it a nightmare of complexity

Why do you want to make something that is straightforward, simple, and easy-to-control in Spring and make it worse?

As an additional building tool, the power to control the flow of resources would be extremely useful. In TA, you had no control over the amount of metal and energy used by nanolathers, other than adding or subtracting nanolathes. In SupCom, you should probably be able to declare that "I want this project to halve resource use" so that you free up resources for another project. Direct control over quantity would be excellent, of course with a maximum. A rush production with added cost might be cool, but is wholly unnecessary.

That would imply that it wouldn't just be simpler to have nanocasters of various sizes, which a player could build given their resource-output desires, and then let Spring's economy code just continue to regulate things. The only thing stopping the idea I just expressed here from being in mods is that nobody's put this idea into their mod yet. I should probably put it into NanoBlobs, just to show how it can work. There is NO REASON WHATSOEVER, from a game-design standpoint, why an OTA-based mod could not have two Factories, both with identical buildtrees, which differed only in Workertime. There's no reason to have "worker units" assisting nanolathing, for example- that was just a game-design concept from OTA that was kept in most mods, because it rewards skillful manipulation of a base design (i.e., making sure your factories can have a few Farks nearby). It's not mandatory- there are a LOT of different strategies a game-designer could use.

In short... these are all really bad ideas. I'm sorry in advance- I don't mean to flame you, and it's not meant personally- for all I know, you regularily help cripples learn to walk, give your allowance to the poor, and make your momma breakfast on a regular basis But these ideas basically either take control away from players, which mainly would screw the skilled over the weak... or are totally un-necessary, from a game-design standpoint, and reflect an OTA mindset about what's actually possible.

Thankfully, "Thomcaster" and not "Carnagal" actually made that post.

I think it shows maybe a tiny bit or promise, but not for this game... Units are far too clearly defined in Spring and most of its mods, to the point where 9.9 times out of 10 the player NEEDS to build a certain amount of a certain unit.

I think what this guy is missing is the repeat setting. If I want a 5:2:1 blend of Predator, Falcon and Eagle drones (example from CvC, laser, missile and sniper, respectively) I set repeat to on and queue up 5 Predator, 2 Falcon and 1 Eagle.

Let's hope that this Thomcaster is actually influential over in GPG-land, then, because if these are his "great ideas", Spring is going to wipe SupCom up as it matures further

Wow very long answer.I do not have he time to read and understand it ( my english is not very good ). I will read it better later

From the short time I played I noticed it was difficult to change the unit production if the queue is long. I have to reset all and create a new queue of unit ?( it is awkward ) .

Maybe I miss a useful thing here ?

I want to add that I simply hate all thoses players who are strong because of micromanagement skills.

It is for kids. I want a game where a strategy is THE skill andavoid theses painful micromanagement games.

I will read what you wrote, but I keep my point of view that theses ideas are good ( Thomcaster ideas )

I will see later

You can't insert stuff into the queue but you can add to the end or remove from anywhere.

KDR_11k wrote:You can't insert stuff into the queue but you can add to the end or remove from anywhere.

Yes ...

Ty for your answer, but it is not good. It would be nice to change that.

I want to add that I simply hate all thoses players who are strong because of micromanagement skills.

It is for kids. I want a game where a strategy is THE skill andavoid theses painful micromanagement games.

No offense... but that speed factor is what defines true greatness in an RTS player. Speed of scouting. Speed of reaction to an enemy's first probing attacks. Speed of defense of a vital point. Speed in setting up a base. Speed of recognizing a feint, creating a feint, or coming up with a flanking strat on the spot to save a unit or two during attrition gameplay.

It's about SPEED... and if you don't like it... tough. Go play something turn-based, like chess, contract bridge, or BattleTech. RTS stands for "Real-Time Strategy", not "Strategy as defined by people with poor motor control"

Competition players do things like map out a standard opening, practice it until they're super-fast, then have pre-built expansion and attack strats that they use to defeat an opponant. Sheer speed is what seperates the mediocre from the good in an RTS, just like a FPS. If you don't get that... sorry. And don't say "it's for kids" like that dismisses the basic truth- if you watch world-class adult players from the serious tournaments (y'know, for money and stuff), you will see that sheer speed is one of the most important factors.

Still don't believe me? Go watch IMGRUNT play ... whatever RTS he's playing right now. There are quite a few replays of him and other world-class players out there... and they don't sit around.

And, in the end, the ideas expressed in the above post detract from strategic play. You think Eisenhower said, "well, I hope that 50-1000 LSTs will show up in time for D-Day, and 100-10,000 soldiers"? No... that's not how it works in the real world, and I darn well wouldn't want it to happen as a player, either

And if you think that SupCom will somehow be different... um... you're going to be soooo disappointed when you're wrong...


Bah... why am I bothering with this... if he thinks the above is really a good idea, then he doesn't know enough about game design to bother hammering home the point

I guess you are respected here, and I will tryt to not be rude.

I think you don't know what strategy means. Obviously the speed factor is important in a war but the problem is that the speed factor is compared to the speed micromanaging. Here is your mistake.

I wish a game where your Macro-management is the biggest factor.

For your information I do not have poor motor control. The problem with almost all the actual RTS communauty is they think they know what the skills should be in RTS.
Do you think Eisenhower had big Motor skills ? That is my point.

And, all what I said should not be taken with such an arrogant ton, yes it was ...

When I say it is for Kids, well, I am sure of it, it is a fact. Do you t hink there is no top players 30+old because their brain is damaged, consequently their stratey skills are poor ?
RTS games as they are, our day, are played essentially by Kids, because of the fact that RTS are conceived in a such a way that motor-micromanagement skills is the main factor.
I do not accept this idea, and it isn't because it was like that in the past, that we should accept that. We need some creative spirit to change things, unless, of course if you venerate theses micro-management skills.

I think our points of view are so different that we can't understand eachother ( and most of the actual RTS communauty, I am afraid, is agree with you on many points ).

My dream would be a game where I could play vs Eisenhower and I wouldn't be sure to win. ( sorry, but it is LOL here ...)

Carnagal wrote:I guess you are respected here, and I will tryt to not be rude.

I think you don't know what strategy means. Obviously the speed factor is important in a war but the problem is that the speed factor is compared to the speed micromanaging. Here is your mistake.

I wish a game where your Macro-management is the biggest factor.

For your information I do not have poor motor control. The problem with almost all the actual RTS communauty is they think they know what the skills should be in RTS.
Do you think Eisenhower had big Motor skills ? That is my point.

And, all what I said should not be taken with such an arrogant ton, yes it was ...

When I say it is for Kids, well, I am sure of it, it is a fact. Do you t hink there is no top players 30+old because their brain is damaged, consequently their stratey skills are poor ?
RTS games as they are, our day, are played essentially by Kids, because of the fact that RTS are conceived in a such a way that motor-micromanagement skills is the main factor.
I do not accept this idea, and it isn't because it was like that in the past, that we should accept that. We need some creative spirit to change things, unless, of course if you venerate theses micro-management skills.

I think our points of view are so different that we can't understand eachother ( and most of the actual RTS communauty, I am afraid, is agree with you on many points ).

My dream would be a game where I could play vs Eisenhower and I wouldn't be sure to win. ( sorry, but it is LOL here ...)

Play Turn based strategy games. Essentially that's what Eisenhower was playing, he didn't get an overview of the battle screen in real time. Turn based strategy games are entirely about the macro strategy you want to play, real time strategy games are just about entirely tactical. Most players play RTS because it's a fast genre, you can't even make the game you are talking about without alienating RTS players. Additionally, I'm not even sure you can make the game you're suggesting without it becoming incredibly boring, real strategists are presented with massive maps and move amounts of troops around, and that's it. They sit for weeks doing nothing and trying to learn about what the enemy positions are like for 1 or 2 large scale assaults. RTS is not a game of strategist, it's a game of battle master. There is strategy required, but it's the underlying theme, not the meat of the game.

Your post is interesting, nevertheless I am not entirely agree.

It is possible to create a good RTS where the " spam-command " processus is not the way to win.
To say the least I have a concept, but even in my main language I have a problem to explain it (while it seems easy to understand for me oO, but not to explain ).

Beside this " concept ", the way to avoid that is to improve the UI micro-management. It is as simple as that.

When you talk about real strategist with incredible map waiting for weeks, I am not agree.
Transform these weeks in 40 mn. ( not a big deal ).

Do you think the chessboard is an incredible map ? The point is that the map do not need to be complexe at all. That is all prejudices.

Springs system for queing orders is fine.


In the average comet game my T1 Car lab will go through a cycle of 5 Insitgators then one construction vehicle.


Enemy has fighters? Sure i'll just take away one gator from eacho cycle and add a slasher.



I really cant see what the need to show this as a percentage. Whether i get 1 con to every 5 gators, or I get 16% cons dosent really affect anything.

As far as micromanagement. Spring's interface is about as intuitive as you can get. You can order a queue of orders crossing the whole map with real ease.

I played Rise of Nations recently and found all my 'micro' was simply having to scroll around the map in horribly close zoom to make sure nothing was idling.


rANDY isnt a excellent player because he clicks twice a second. He is a good player because he outsmarts his opponent and makes wise descisions.

Ishach wrote:rANDY isnt a excellent player because he clicks twice a second. He is a good player because he outsmarts his opponent and makes wise descisions.

And clicks twice a second to do it

I kindof agree with your basic idea Carnagal, but not with the implementation of it. Making it more complicating makes it much easier to lose the overview of the game.

Ofcourse the top players like rANDY do all the stuff you're supposed to do in strategy, but he'll win a game against someone with the same level of tactical and strategic skills but with slower clicking and mouse coordination.

A way you can make such a game in realtime imo, is not allowing direct unit control at all. More implicit unit control... the settlers series for example have very indirect unit control, but it takes the action a bit away from it. Basically I don't have a solution

Since this post in initiated from another post till alive. Here is the link http://garage.gaspowered.com/?q=node/6653#comment-64397

Discussing in two threads at the same time is seriously not practical :/ At GPG you link to here and here you link to GPG. Also, this forum is not specifically for AA

I have idea!!

how about, go with me here...

HOLD DOWN SHIFT AND CLICK!! make it a PRIORITY QUEUE!!.. put it at THE FRONT OF THE LIST!.. end of argument.