Quick Start Development

[prime socket]

2 😄 3 😮

[stiff hazel]

Depends on what start i want

I could full eco and have a 3min eco, but very limited ability to produce units. Or I could go half and half and have 2min economy + units

[worn torrent]

Remind me, what is the starting M and E? Does it depend on if there are any leftovers on the budget?

[stiff hazel]

This is also on good wind maps btw, I havent been able to test in on a shit wind/ solar map

[prime socket]

200 metal is leftover, 800 energy leftover

[stiff hazel]

Usually I queue my lab last so I only get a partial benefit and the ability to spend more on eco with only a partial benefit for the lab

[bronze tree]

In 2 minutes if I'm opening greedy (no units), I have 10 wind, 5 mexes, botlab and 3 constructors out on the map, with an energy storage on the way with 400 metal still in the bank. With default quickstart, I have: 8 wind, 4mexes, botlab, and no constructors. Would say it puts me at around the ~1:30 mark. given that the constructors are still important. Note: Normally I would be building units as well, slowing a down a bit more, but I would also have units at the front already

[storm moat]

I have finally been able to figure out and test what i was trying to do like 8 hours ago. jc. all over 1 mistake i kept making like a moron. i am sad.
however. The walloff I mentioned is exactly as possible in the old iteration as it is in the new one. Except now you get 5 winds as arm instead of 4.
Or 4 winds and 3 converters. which is probably better anyway.

[prime socket]

Walls are cool.
Any problematic cheesy exploits?

[storm moat]

only most likely to occur on very specific maps with many chokepoints, or on maps like like that one long rectangle ice map

but theyd be there for sure

[worn torrent]

Walls wont be a problem

[storm moat]

losing 3 windmills to start in a way that completely blocks access to a potential raid path seems pretty decent to me. its a common tactic in 8v8s on chokes in places like glitters canyon, and while ya it gets busted the point is slowing that down and denying pure scout rushes. Ticks dont do much if they get slapped by a lone thug guarding a wall.
but, on the only two maps that matter in 8v8, it is unlikely they will make a huge difference anywhere on a spawn point that's true.
And on small teams/1v1 maps its more likely to be a non-issue

[prime socket]

What happens if you build dragonsteeth and I don't and instead I go eco
I figure I got an advantage

[worn torrent]

No one builds walls seth

[storm moat]

3 winds is the advantage.

[prime socket]

Risk vs reward.
I suspect it'll be more useful to build some scouts of your own and try to goalie the enemy some, or harass of they don't do the same

[worn torrent]

Walls are frontline, not a first two minutes thing

[storm moat]

you can draw a line across a full radius of the building circle, in just dragon teeth, for 3 winds, and have enough left over for converters.

[prime socket]

Yeah, me neither. They're annoying and time consuming to build. The time my cons spend building walls is time not building eco. And likely exposed to getting bonked from raids.

[storm moat]

again, not saying it would be op on its own. it would only matter on specific 8v8 maps i imagine.

[prime socket]

Howeverrrrr building them instantly sounds a lot more convenient. I might wall off a choke. But more likely I would build a home base llt or just build an extra pawn at the beginning to guard

[storm moat]

yes, the build time is their biggest detterent. Which is why i imagine building it on spawn would be ideal and done somewhat regularly on those maps

it would make more sense to build the walls in spawn and the llt after. less bp/time consumed. that's kinda the point im making

but anyway, just throwing it out there, idrc anymore. was good for learning how to do prs though

[worn torrent]

There are no enemy units at the start anyway, why are you even thinking of walling your base? Make units and use them with your apm

[storm moat]

you dont get to make units in prebuilding. And 4 winds, 4 mexes(assuming you can only reach 4 in prebuild) , 3 solars, and 3 converters is plenty eco to make most units on

[safe pier]

Probably going to install and test tomorrow. But one potential issue I see is with the build range. On some starts it could allow you to place buildings and radars in weird elevated positions. A potential example of this issue would be on Charlie In The Hills, with the back spawn.

[sour spear]

Putting walls and a sentry at a specific chokepoint to stop leaks there early on would be a strat that helps on AI games and maybe in PvP games. Is it a "efficent" strat probably not but you dont need to wall it off to slow the leak enough to get entirely killed by a single sentry. It is probably better to just eco scale with the extra eco but thats greed vs defence on niche situations.

[prime socket]

That can be fixed. I can calculate the angle between where the object is being placed and where the commander is and disallow it if it's too high up.
But I'm also feeling kind of like YOLO "it is what it is". Dunno. Community feelings will sway me.

[sour spear]

Build radius for everything else is currently a cylinder. Its a edgecase situation to make pre spawn a sphere where the current widget doesnt adjust for height with its range showing for the client.

[worn torrent]

If you want it to potentially be used for serious games you are gonna have to limit build area to the area where a commander would have moved to on normal games depending on the budget size - budget meaning aproximate game time.

[prime socket]

aight, three votes so far. How would you expect that to be detected?

[prime socket]

sphere isn't the ideal solution. More punishing on the edges.

[worn torrent]

You would measure the standard average distance away from the commander start point at X time depending on budget.

Sphere is good, coms struggle to build on highground.

[sour spear]

So basically adjusting cost based on commanders move speed from the center?

[prime socket]

too unintuitive

all the costs indicated in the buildmenu would be false

[sour spear]

without a tooltip showing its price changing at the mouse yeah.

[prime socket]

sphere relative to what? That's the issue.

[worn torrent]

No, cost are fine-ish as they are. Just need extra tax due to E cost.

Start position

[prime socket]

if the commander stands on a slanted hill, it must be able to build in places it would normally be able to within range

and I don't have any good pathfinding methods at my disposal to check if you can get to a spot

If I have a sphere? Up close, I can reach waay high things. on the edges, the height is pathetic.

If it's a squat cylinder? It's going to get wonky as you try to place things along a slope.

[worn torrent]

Well, a cilinder is way worse than a sphere

Unless you can restrict the upper and lower limit

[sour spear]

Its a specific map edgecase situation which is relevant to smaller more terrain maps correct? Can the radius of the gadget be adjusted based on the map size as the most simple way to control it?

See DSD for a better example of it

[prime socket]

Yeah, but that just makes bases more tight which messes up auto generated layouts somewhat and makes starter builds more difficult and less predictable.

[prime socket]

use noob speak plz

[sour spear]

Delta Siege Dry

[prime socket]

Well. I could create a mesh and treat it like the bucket fill tool from ms paint. That's an algorithm.

I would treat angles > 20 degrees or whatever as a "wall"

Then sort the points in a random order and check proximity of the currently selected things to the mesh points

but that would be extremely expensive sounds like

yeah

[worn torrent]

No need to do map specific anything, building buildings is a standard procedure.

• Document what time your budget is exactly referring to
• Measure average commander travelled distance away from start position. Add com buildradius (cilinder diameter buildbox)
• Add a sensible enough higher and lower cilinder limits based on the standard commander reach.
• Remain conservative on the lower side.
• Keep start building area rather small and extend reasonably with increased budget options declared X ingametime.

[sour spear]

If you can use the traversibility mesh and give a radius based on only where you can move in there it should work better but would be costly for low end CPU players.

the commander build height is unlimited fyi if it can walk beside something to build above it can. Same for all cons.

[prime socket]

checking that shit every frame is oh my god

[sour spear]

Yeah its not performant likely unusable I bet

[worn torrent]

False

[sour spear]

Then you best tell the developers that.

[prime socket]

somethting like this?

this is doing angle checks at commander/maxBuildRange coordinates, whichever is closer. If angle > 20 degrees no buildy

[prime socket]

NOTE TO SELF: Fix fadout animation for base budget popup

[prime socket]

EEHHEHEHHE

[bronze tree]

The angle approach seems pretty reasonable. You might want to test it on the industrial map (6v6 I think?) That has the steepest terrain I can think of that is still walkable

[worn torrent]

It isnt false, you just cannot allow that to be done with quickstart because there is a price to pay for travel distance.
The buildcilinder is too big.
If you want to build over an edge you can do it after game start.

[bronze tree]

https://www.beyondallreason.info/map/industrial-revolution

Alternatively, do you have access to the movement terrain mex? The: This is where a commander can move to (pressing f6). Not sure how accessible that is by code. You could use this +the build range of the commander to get the exact tiles

[prime socket]

meh, it's wonky. I'll have to stew on it some more but this angle check thing has potential.

[prime socket]

the tile search every frame is incredibly expensive. Think about how many points there are to check for each build and you can now drag-build many many things at once.

[bronze tree]

Hmmm.... could be coded efficiently by storing it based on commander start location, but that extra implementation work to store and keep updated

[prime socket]

That's what you're seeing in the screenshots and video above. I'm checking the commander's build range distance away in the direction of the commander. Measure the height differential angle between the build, and that arbitrary point build range distance away

But that creates weird situations where opposite of the commander is another weird terrain elevation

[worn torrent]

Those building cilinders are unreasonably large. You are not making people pay for travel distance in your costing so the cilinder must be small

[prime socket]

That's a separate issue. Even if the cylinder was tiny you could still stand right next to a canyon cliff and build stuff on top of it

Unless it was unreasonably tiny

[bronze tree]

I meant something slightly different:

1. Generate the "standard traversability" matrix of the map for the commander.
2. Overlap this with cells within the circle.
3. For cells within X distance of a buildable cell, set them to buildable.

This would get you everything the commander is normally able to build (within the circle range) if they were placed there.

Still an edge case if there is a map with a looooooong path to the high ground above them, but for that you would need to find the distance to the cell, and this might be unintuitive.

[prime socket]

@cedar wren is there a more sane way accessible both in which it and gadget space that we can check if a commander can reach a build at a significantly different terrain elevation?

We are trying to avoid this https://cdn.discordapp.com/attachments/1413936809980858408/1434863689944793201/image.png?ex=6909e072&is=69088ef2&hm=5bb9e3b7d85752145d495e2ba23581e088d1e1662be7048c12b2f918d519fdc7&

[prime socket]

If the terrain grid points are reduced to the ones immediately around and within range of the commander, that could be formant enough. It'll be much code though

[sour spear]

If it changes to that a type of Gui showing the buildable area will be needed before game start.

[bronze tree]

I'm happy to help take a stab at it tonight. Can definitely write the pseudocode for it, but still need to get around setting up a bar developer environment

[prime socket]

Slightly outside of scope. I hate shaders.
Currently, valid places are purple for quick start otherwise they are green

[prime socket]

See here

[prime socket]

Oh man why pseudo code you going to make me work aren't you
Lol

[worn torrent]

Coverters/storages will need to be significantly more costly if you can afford that many 😂
Really gonna have to crank up on the extra Mcost based on Edemand

[sour spear]

Valid reasoning with the shaders.

[bronze tree]

Mostly because the last time I touched c++ is a decade ago ^^ I'll try and actually set it up though

[prime socket]

My thoughts on the best Lua approach is to create a widget that scans the entire map once and performs bucket fill type algorithm marking each point on the grid as a different number according to what group they belong to

So if you can't get from point A to point b, the section A belongs to is marked as 5 let's say and b it's marked as section 6

[bronze tree]

That would be even better. You only get 1 call at game start, instead of every move of the commander

[prime socket]

Then it's a matter of trying to provide a API function to get the nearest grid points and check what group of points you are in

That gets complicated when you consider that different units have different traversibility. Some can move through shallows of different depths than others. Some can't touch water at all. Some fly

Some are spiders, some are tanks, some are bots.

Too complicated for me, it will be a month or more of development and bureaucracy to make.

[bronze tree]

Should be good enough to just use the commander traversability right? It's about quickstart, movement type of the commander should be leading

If you are happy with the idea itself, I'll try and take a stab to implement the api that says: You can or cannot build on this cell

[prime socket]

@mossy lichen I want to know if you're going to freak out if we try to generate a traversibility map at the beginning of the game to help us determine if you can get from point A to point B

This is a problem scavengers has as well, but our interest in it here is in detecting if something is on a unreachable cliff

[prime socket]

I don't want to say do it until I know the more technically minded aren't going to freak out about it

But I can't stop you. If it sounds like fun and you don't mind the risk of throwing away work take a crack at it

[bronze tree]

I have definitely thrown away more work than I kept 😅

[mossy lichen]

Spring.TestMoveOrder

[prime socket]

This tests if that exact position could be stood upon by the unit, not whether or not the unit can reach it in practice

There is no path in checks there

There is a spring function that returns a series of waypoints from one position to another, but I've never been able to get it to work

[mossy lichen]

Spring.GetUnitEstimatedPath

[prime socket]

That's the one. Couldn't get it to work

[mossy lichen]

Generating a traversability map is the exact thing a game should not do. It is the engine's specialty.

Especially because deformable terrain means the map can change.

[prime socket]

Looks like that function was removed which explains why I couldn't use it

Looks like it was made into another function. Unfortunately it requires a unit ID which does not exist in the pregame

https://recoilengine.org/docs/lua-api/

Forgive me for making the AI do the research, texting and driving lol

[prime socket]

@floral shard you could probably use the move goal thing tho

So could zombies now that I think of it

[storm moat]

Uhm. I assume that plasma artillery has a widget that measures height to increase its range down slopes, correct?
Is it possible to use that in reverse, and only for positive elevations?

[prime socket]

Not in the way you think, and I'm not sure that we want to use a arbitrary weapon definition to do free line of fire checks

Maybe a combination of tracing a ray between the commander's position and the position of the build would work. Probably with some arbitrary height above ground between the two spots as well

Combine that with allowing up to your build range to build above and below the commander's current position

Since most desirable build positions are flat, you wouldn't have to deal with this most of the time. If you spawn in canyons or something, then it's whatever your commander can see or whatever is at a similar elevation

[storm moat]

Hmm. Well, I trust you lol

[prime socket]

I wouldn't lol I know me

Spring.TraceRayGroundInDirection is a thing

[bronze tree]

Using the commander vision line? That should cover most cases as well. As long as trees and rocks don't count for intercepting it.

[open oak]

all of this might mean your build-radius is just too big, no?

it's like 2-3X the com's buildrange now?
if you pull it down to 1-2X there's a lot less space for shenannigans

[prime socket]

It is not.
If the build range was 200, and there was a cliff within 200 range, you would still be able to build on top of the cliff.

This isn't an issue in the vanilla because build orders you can't read are attempted and fail.

[open oak]

but how can you reach a spot in quickstart but not ingame? 2d distance vs 3d sphere or smth?
i mean, if you can just spawn and build a thing over a cliff ... that's just the map, right? i build observers on unreacheable cliffs all the time

[prime socket]

Because we are not measuring to the commander's current position build range, we are measuring to where he could be standing to build that thing

[red parcel]

Yeah. Sound like a fucking headache

[prime socket]

This is my hacky and cheap workaround.
The black dotted lines are Ray's drawn that return true if there's no terrain blocking
The yellow highlighter zone is + or - build height
If either condition is true, the ray Trace has direct line of sight or the build is within the highlight zone then you can build there.

[red parcel]

Calculating all the possible position of commander and therefore all the position it can build in

[open oak]

is this not exactly the same as within build range? your logic smells

i'll invert the question
is there a raycast position that's valid, and NOT in the buildzone? i don't think there is

[prime socket]

Depends on how tall the build zone is.

[open oak]

i think you want raycast AND buildzone
right?

[prime socket]

Yeah, but I agree something smells about it

Example if the green marker was a little impassable hill, it would still be able to build there on the other side

[open oak]

(raycast AND range) would still let you build a thing in a through, but i think you can do that in-game anyway.

[prime socket]

Version A

Version B

Which one do you guys like better? @open oak @red parcel @bronze tree @storm moat @worn torrent @odd loom @sour spear

[red parcel]

what are the difference in specific? B is your raycast-based solution right?

[prime socket]

it's all floodfill. The difference is how precise and picky they are about what is "reachable". One is basically only things you can walk to, the other kinda has some fudging to allow building up some reasonable cliffs.

[storm moat]

b is probably more competetive
did not see that you could build up and down with option b, i thought at first the difference was to do with walking distance.
since that doesnt seem to be the case? A is better

[red parcel]

I think I prefer being more strict about allowed build location. Erring on the side of "not allowing dumb exploit" is the better call imo

So A is what I would pick

[safe pier]

I like B for some cool starts on DSD.
But I think A is more clear and consistent

[prime socket]

I think this is my preferred one.

lol ram warning

[worn torrent]

This seems ok
But do give the circle a trim, about 20% less radius should do

[open oak]

This tbh, but seth wants to emulate walking around a bit, hence the large radius
Also, @prime socket A seems logical, B ... If you hadnt told me it's a looser fill alg, i wouldn't have known. I'd prefer more "intuitive", whatever that means

[worn torrent]

You cant walk around "a bit" and also build the same ammount of stuff as the other player on the same ammount of budget time.

@prime socket I suppose you could purchase additional building range for a price to simulate the cost of walking. Could that be a thing code wise?

[bronze tree]

Those look great!

I don't mind a larger circle, but trimming it would be fine for me as well.

[prime socket]

600>500

[cedar wren]

I'm not aware of an easy way to ask if the unit (x) at point (a) would be able to build (y) at point (b).

[prime socket]

https://github.com/beyond-all-reason/Beyond-All-Reason/pull/6044 OK GUYS
I feel it is ready.

• budget reworked to include buildtime. 1 second commander buildtime = 1 budget.
• the energy conversion cost for budget is 60:1 instead of the old 10:1
• budget amounts reduced to accomidate these budgetary cost adjustments
• Custom rounding. If budgetCost < 10, it is what it is. < 50, round to nearest 5, >100 round to nearest 10.
• build range reduced 600 -> 500
• you can no longer build stuff in places your commander can't walk to from where it spawns within the instant-build range.
• you can now drag-build with grids
• you can now "Ready" when your budget < cheapest eco building (metal maker)
  Please test it out and tell me your thoughts. I'll mark this for merging as soon as i get some verifications it's not busted.

[prime socket]

You know, I haven't seen @eternal flower in a while

This feature seems like something that would be right up your alley

[prime socket]

@iron wraith why is it that other examples you named of games with pre-built bases suck?

I'm looking them up, but, I'm unlikely to hear a prospective from a high level RTS player such as yourself about them

[prime socket]

lol you can set some bullshit and then press g to cancel them as soon as game starts to get a prebuilt autogenerated base.

[prime socket]

really hard to debug without being able to recreate the bug myself.

[floral shard]

Maybe because all the good ones just make you start with a central building and a bunch of builder units

Or a mobile unit in case of SupCom and other TA-likes

[prime socket]

the base budget display for spectators needs to do something else

[prime socket]

while dragging a build, deduct the total amount from the budget bar for all the builds in the drag

[vale idol]

Sorry I'm late!

Vision range, not radar range. One factory only. If there is a metal spot in vision range, one mex on the highest value spot in range, AND one solar OR one tidal; or if there are no metal deposits in Commander's vision range, one metal maker AND four solar OR four tidal (or whatever makes sense mathematically for balance)

Unless you want to continue coding super complicated map detection algorithms and no-go masking + for when people inevitably grief each other during spawn, for a feature that only happens once per game and most of the time will not even be used, I don't think it is possible to have a simple function that works for all maps and game formats that is also not very strict.

To be clear, I think it is extremely worth looking at the demographics of who would be using this feature. Can't say for certain because I haven't collected any data on this, but I would suspect that only people brand new to the game, and also people who have a very limited amount of time to play, would even use it.

The stated goal is to skip to 1:30-2:00, I think building defenses is right on the precipice of that. To me, that's about the point where my first scout is making contact unless the map is very big, and am just finishing the first expansion point mexes after having made only one LLT there first, if the map is on the smaller end.

But this system doesn't put units on the field. If we go with my top paragraph, all we are trying to do is get units cooking right off the bat with minimal eco to sustain it. Now the player makes the choice whether to tap into his starting reserves and immediately begin building defenses, or begin building surplus eco. No tactical decisions until the timer is ticking.

The odd thing to me is that I don't ever find myself just waiting for constructions to finish unless I am pausing the game to issue build orders, which is obviously either impossible or bad behavior for pickup play. When I do this it is an attempt to study perfect strategic theory, because it's too easy to mess something up in real-time. This is a non-factor when I play with other people, but I don't think most people playing for the first time, or are playing on a time budget, would do this. But I still see the value in skipping this boring and repetitive part of the game, which then frees up some mental bandwidth to use for things that actually matter.

As far as AFK handling...I think it is considered rude to start a game when you know someone is AFK. At the same time, life happens to all of us. Therefore I think it is better for the game engine to err on the side of patience. Do not force starts, do not automatically pick factories, do not allow the game to begin until everyone has chosen a factory. The potential for abuse to grief people via time wasting by not submitting here is something that the normal report and ejection systems can handle.

With so few structures, it is also not really necessary to worry about space constraints, unless the degree of player spawn overlap is severe, which at that point may be a mapmaker issue.

[worn torrent]

People should be considerate of other player's time. The countdown should happen as normal, the games can be forcestop or started anyway as the lobby requires

[prime socket]

The game design team wanted quick start to allow for fully custom base building. That constraint means not taking the one of this, one of that, one of this other thing approach.

It's a bit late for a retooling - come the next update the feature is going to be near complete and polished.

The approach to economy you describe sounds like about 45 seconds to me. Have you tried quickstart as it is now? There is a smaller budget amount version that may be more to your liking. I haven't decided which budget amount is best yet.

The structure placement algorithm for AI (and afk players) was a challenge. I've tried at least 5 different methods. Weighted random, deterministic weighted selection, true random, some attempted twice. Ultimately what I went with is a simple list of structures queued in the order they appear in the table. https://github.com/beyond-all-reason/Beyond-All-Reason/blob/9ed2af8b8b12920c3ec0198202babaf798a3a2b7/luarules/configs/quick_start_build_defs.lua#L4 mostly so that future non technical maintainers can edit it to their liking.

What is com vision range again? We ultimately went with 500 range for the base build radius. This is because if it's too small the structures start spawning in weird unpleasing arrangements if auto generated.

[prime socket]

I designed it so if you get force started your base budget gets spent on structures. If your queue was empty you'll have a mighty fine base. If it was partial you might find your start a but wind heavy.

[prime socket]

It's live

My PR for projected quick start costs accurately reflecting drag builds didn't make it in this time. Probably next week.

[warm sinew]

Hey, I played a few rounds with this mode and I think it's quite nice. However, we played on a metal-arm map and some players chose Legion. The advantage here is having a lot of energy right away for the more expensive tier 1 metal extractors. Of course, this isn't a direct problem of the Quick Start, but maybe it could be balanced a bit, making wind more expensive for the faction so it's not so easy. I think this is an advantage because otherwise Legion usually takes a while to get the extractors on the field and with the energy, they basically only need the build power, which is cheap for that advantag. In the Quick Start setup, it's like you fight tier 1 first, so the intermediate stage provides an advantage for longer, than in the casual setup, since there its shorter (the time you have advanced extractors) and legion gets less metal, to build the energy it needs for those extractors.

[prime socket]

Although quick start is busted right now (hopefully by the end of the day it won't be) the budget formula has been adjusted in this most recent update like this morning

1 sec commander build time = 1 budget
60:1 energycost to budget
MetalCost = budget
Total that up and round it to nearest 5 or 10 and you got the new budget cost.

Should be more fair

[warm sinew]

Gonna try it. You are doing a great job.

[prime socket]

It's busted right now 🙁 I'll post here when it's fixed.

[warm sinew]

I gonna try it when its fixed ;D

[glad lintel]

https://github.com/beyond-all-reason/Beyond-All-Reason/pull/6134 its waiting

new gameversion is being deployed

[prime socket]

It's fixed (probably)

[warm sinew]

Very good :)

[prime socket]

Thanks! I promise to never make a mistake ever again. 😉

[glad lintel]

thats a mistake if you ask me

[prime socket]

@stiff hazel I think the rounding to nearest 5 might not be the bestest thing

I remember there's a subtle difference in cost between the windmills of the different factions. Competitive boys make a big deal out of it. @worn torrent would probably >:{ at me bout cortex and armada having the same cost for windmills in quickstart

I like how the numbers are pretty and it makes budgeting slightly easier to parse mentally decreases the chance of weird numbers... But... I think our community likes accuracy more. Nuanced values.

[worn torrent]

leg windmill shouldnt be 5 more if cor and arm are same price
I just would like further extra E cost beyond the 60-1 cause E is very limited early on

[stiff hazel]

The main reason I pushed for "nice" numbers is because gdt also wants "nice" numbers.

[prime socket]

without custom rounding

[prime socket]

Do they? I don't remember hearing that.

[stiff hazel]

Let me see if i can find the message

[prime socket]

I wonder why the dragons teeth go up to 9 without the custom rounding. Weird.

do we feel better about having nearest whole number below 100 budget cost, and nearest 5 above 100?

https://discord.com/channels/549281623154229250/1437173642814292059 I'll leave it to the community to decide.

[stiff hazel]

I didn't think that would look good at first, but its growing on me

[stiff hazel]

After looking through 2000+ messages, I couldn't find it ):

I don't have access/no clue where it is, to the game design document but it might be under what I interpreted as arbitrary numbers

[prime socket]

Ok lol

Thanks for trying

[stiff hazel]

It was about something causing stuff to cost decimals, but when I thought about it, arm wind is the only thing that isn't divisible by 5

[vale idol]

I wouldn't think there would be faction differences here, but I looked at Armada quickly to confirm this. This is more than enough area to encompass the small kickstart setups I described.

I am making these predictions under the assumption that people will be using Quick Start with otherwise default game settings, meaning 1000 mass and 1000 energy and build rate multiplier of 1. Also kind-of assuming a short walking distance to the first mass expansion point, and what that means also kinda depends on the map; some maps have them in clusters of 3 while others are a bit more evenly distributed, especially Cavedog maps, and there may be other patterns of mass distribution in circulation that I haven't seen. The first expansion point would be the ones where you send either your first built constructor, or the commander to (or both I guess), depending on how you prefer to play, not the mass very near what one could consider "the starting location."

Normally I start my games with 1 mex, 1 solar unless wind is really good, and then factory which queues up constructor first, if I do anything else then some kind of tradeoff. The constructor rolls off at 0:55 with a healthy reserve of mass and energy and already growing. 0:50 with critically low mass and energy if the commander makes one additional solar and then assists the bot lab, and also 0:55 but trading some mass for energy in reserve if the commander makes 2 additional solar before assisting.

Can hold off on the mex first until after the bot lab to save a bit of time and be right on the borderline of mass a bit earlier than I would like, and the constructor rolls off as soon as 0:48 with a just a bit more mass and energy in reserve than I would consider critical, if commander builds additional solar after the mex and then assists the bot lab.

Starting factory first, then two solar to minimize the energy deficit prior to the constructor rolling off, has the rolloff occur at 0:48, with minimal energy in reserve.

That leaves at best, 1 minute and 12 seconds for my constructor to travel to the expansion patch of metal, put up a LLT, and make more mexes, and build more solar with Commander and/or put the LLT somewhere else because it doesn't make sense to put it by the expansion patch. It's a pretty tight window to get it right, that I'm also trying to be generous with for approximating how much eco can be built by the 1:30-2:00 mark. I think I actually wouldn't quite be done with a 3 patch by then, but getting close to it. Really depends on travel distance.

And this is all just about whether or not the preplacement of structures would allow only eco buildings. I think it makes sense, but that's only my opinion.

I am merely reviewing what is here according to the original scope proposed in the OP. It could certainly be possible that the scope has changed since then, and to me, is lost in the giant text wall sea above, but if that is the case then all I can really add is to ensure that you are still solving the original problem. I don't really have a dog in this race, I haven't tried Quick Start and likely would not use it personally, but again reiterate that I think it is a good thing to have some sort of system like this.

The tooltip in-game says the Commander's LoS radius is 450. It's more than enough space to fit the small amount of structures in the hypothetical snapshot I'm talking about, unless the player chose to drop in an extremely confined location, but I think that's an issue you're kinda working through anyway yes?

I hope you can see my concern. I don't think anyone is doing anything wrong here, just seems to be doing a lot more than you originally set out to

Honestly at this point what you have now could work as a more advanced extension, but I feel it's potentially fast forwarding to much later in the opening

[prime socket]

So far the default budget has done this. The smaller budget might be closer to the original goal.

[vale idol]

I'll check it out now before I do anything else

[prime socket]

The 1 of each opening you described is unlike any I've spectated so far. Most dudes start 2 Mex, couple wind, factory, third Mex iirc

Or was it third Mex then factory

Most of the popular maps have 3 Mex openers

[vale idol]

So everyone is running complete deficit at start?

[prime socket]

I think so

[vale idol]

Strangely when I assisted bot factory with commander during deficit, it took even longer to build

I'm not sure why, putting all generated m+e into a single build should only result in the minimum possible time, not extend it

It could be that the build priority system makes the deficit viable, but that still should not happen I think

Hmm well the normal budget is about what I was thinking, I suppose I just overestimated how much it would be seeing you guys showing edge case handling with flood fill and reachability exclusions.
The fact that we can continue queueing into the Commander's build order before the timer has started ticking is doing a lot of heavy lifting too

I think maybe some playtests are needed for someone who rushes units from the instabuilt factory(ies) vs someone who spent all his budget on LLTs

I can see someone leaving a cluster behind as a surprise while building from scratch in a different location, to confuse the enemy opening

[prime socket]

I feel that would be super risky because you're dumping all of your starting resources in a dead end location and trying to catch up in a secondary location.

That entire time that you're not set up in the second location you are losing out on the income differential.

Another thing, if you are not assisting your factory with your commander or don't have a factory you will quickly start wasting resources. You can convert the excess energy until your metal storage fills up but 800 metal fills fast at 10/sec

[vale idol]

Yeah but if you can time a counterattack on the rush force, you get the flanking bonus

[prime socket]

So what's that... 80 seconds?

[vale idol]

I never let my mass fill, supcom has trained me to not care about floating energy, that may be something I need to work on here but tbh not really since it seems like deficit is less forgiving overall due to priority

Wait are you saying wasted energy turns into mass now? So running surplus energy is a thing everyone should always should do?
Sorry I guess that's kinda getting off topic though

[prime socket]

You would plop down a few extra metal converters to account for the fact that you're going to walk for a while to a second location to prevent wasting energy and get some converted metal while you walk

[vale idol]

Oh. Yeah I could be better about using those but don't think I would normally consider building them at the beginning of the game when e is so scarce.
It's not just traveling to the expansion point that happens, for sure, if it's the constructor traveling then the commander is building something, if the commander has to travel that early then the first constructor builds solars right after rolloff, and the factory continues building the whole time

[dusty marsh]

I haven't kept up on this thread but I've played a few rounds with quickstart enabled and I have to say this is probably the coolest quickstart solution I've seen so kudos on that.

My only issue so far is it seems that you have to spend your initial budget before you can ready. At least with the settings I've seen the budget is big enough that I would actually prefer to save some resources to invest in either units/cons/buildings cons can build (geo/nano) but I have to spend the entire budget on whatever the com can produce.

This isn't a big issue and is likely intentional but I feel it does potentially cut out a few potential opening options you might have on some maps.

[prime socket]

This was a design constraint handed down to me from the powers that be. I'll be talking to them on Sunday. I can bring this up.

After watching brightworks get very upset about having to spend his last 8 base budget to ready, I made it so that you can ready if you have < Metal Converter budget remaining (cheapest eco structure)

Is that good enough or would you prefer more leeway? When you start with a small amount of budget remaining, it is refunded as metal

[dusty marsh]

My personal preference would be giving power to the player and allowing the ability to ready at any budget but maybe with a big fat warning if you have >5% left or something?

I think if this does become allowed some indication/preview of what resources you will start with if you ready with current spend would be good/needed.

(an example of what I wanted to do is start with a legion bot lab + winds and save resources to immediately boost out a con and start upgrading mex's to t1.5 with those saved starting resources)

[prime socket]

Quick Start deducts a mere 800m/200e. If I refunded the budget as metal or energy or whatever - what good is 1000 energy/metal vs a fully built base? The only use case I can think of is intentional funneling nonsense for coop cheese. Or, if it wasn't rewarding enough, unaware players who just hit ready or were forcestart'd would be at a huuuuge disadvantage to the point of calling stop.

[prime socket]

So like, maybe a specific threshold like 500 or so you can opt not to spend on structures for a refund. I would need someone big brain to help vet this.

[dusty marsh]

Yeah that's kinda what I expected initially.

Another reminder, I went in totally blind and haven't been keeping up with this thread so this is all just off the cuff remarks and I don't feel that the current implementation is bad in any way just that these were my "I wish it worked that way" impressions.

[prime socket]

The best kind of impressions. Thank you.

[dusty marsh]

Thanks for putting in the work! Way more fun than the supcom quick start where you have no customization for your starting structures.

[prime socket]

Maybe something like this.
@weary turret Increased refundable amount

little indicator

I could theoretically increase this up to the deduction amount but the players will be boning themselves pretty hard losing out on the build time

[weary turret]

So I watched that video and I can't discern what that thing is supposed to be telling me. It looks like a button I push if I want to receive a refund... What am I supposed to understand from that label?

[prime socket]

What it's supposed to tell you is that the remaining amount will be refunded to you when you start the game

[weary turret]

HMMMMM. This is raw off the dome, so it might not be right, but here is what I'm thinking.

We disable the Ready button to force the player to start placing buildings. My eye is looking at that ready button as a hint for when I'm done. When that button flips back to green, I'm thinking, "ok I can start now."

Since my eye is going to that button already, it seems like that's where the "Ready and get 500m refunded" should go.

If we can't do that, then rather than just a flashing button looking object, can you say it more plainly? "500m will be refunded" or something that looks more informative and less interactive?

[prime socket]

Probably option 2. It is not trivial to integrate rmlui and GL UI elements.

[prime socket]

Given that we are going for option two, what is the most aesthetically pleasing way of doing so?

For a small little button like thing the Blinky background might be okay but it's if it's wide enough to accommodate all that text it might be very annoying looking

[weary turret]

Something similar to the wasting energy label. "Unused resources will be refunded as metal" or something?

[prime socket]

"Remaining" and you got yourself a deal -
No pulsing? Just a solid color background label? How do you feel about the pulse I'm using now?

[weary turret]

Imo it’s drawing too much attention. It’s not a button. So just treat it like a label

[prime socket]

@weary turret

I think the static color looks a bit 90's learning game ui

but I need it to be a banner to grab attention, but if I make it see-through it doesn't look good with the background

I think I wanna go back to a more subtle ease in out

[weary turret]

Why does it need to grab so much attention though? The mechanic will happen whether the player does anything or not. It’s an “inform” not a “take action”

[prime socket]

this isn't meant to grab attention so much as shake The Sims 1 UI feels

[weary turret]

I think a little banner off the bottom of the main UI container would be fine.

Also, I don’t think we need to worry about the “sims 1 ui” feel. It’s all getting reskinned with the rmlUI update. I think we need to aim for great UX even if the UI isn’t perfect. Is it driving the right behavior and is it informative?

[prime socket]

I think it tells people what they wanna know.

If you tell me to kill the animation dead, I will.

thats as good as I can make it while fitting your design criteria.

for the blinky

[prime socket]

the rest is mutable

[weary turret]

Personal opinion, it doesn’t need to be a purple background. However if we are getting feedback that players are confused and we need them to see this, this it’s fine for the next update 🙂

[prime socket]

Would you prefer some other kind of thing?

I don't have an artistic bone in my body

[weary turret]

I mean, I think the default container as the background would be fine

[prime socket]

So like a second container with the same stripey gray background with the white text? Or are you saying expand the panel for the budget display to also include this banner when it appears?

[weary turret]

I’m not at my computer. But you know how we have those labels we attach when you’re wasting energy or metal?

[prime socket]

Yes, they are overlaid over the top right corner even clipping over the max energy or Max metal amount

[weary turret]

One of those, placed in the position you have the label now. Black or (80% transparent black) background

[dusty marsh]

Yeah mechanically that's good and is what I meant, I know you and Scopa continued on discussing further but I actually think this looked pretty solid.

I think maybe it would be more clear if instead of a separate button the ready button could turn yellow maybe with a pop-up that says "remaining budget will be converted to metal at a 1:1 ratio, note you are giving up free build power would you like to continue? " (possibly with option to toggle off confirmation?

Full spend of budget would be green ready button as per usual.

(but tbh everything you cooked works just as well, the mechanics are what I care about I'm not picky on ui)

[prime socket]

"it's much more efficient to place builds before the game begins." Or some such. People tend to tldr

[dusty marsh]

Yeah, I do technical writing sometimes so I'm overly verbose

[prime socket]

I'm hesitant to further clutter up the code base with quickstart stuff but it's not a terrible idea. I'll think on it.

[prime socket]

Would it be sufficient to simply make the button yellow with a tooltip warning saying such?

[dusty marsh]

Imo yes it would

[prime socket]

I can do this. Or I can do the tooltip thing with yellow button. Which do you prefer?

[dusty marsh]

I think quick start will quickly become defacto mode for less standard/competitive game modes (it seems to already be on the way).

[weary turret]

Tooltip thing?

[prime socket]

I have a method to inject tool tips into the ready button. So when you hover over it, it'll tell you why it is its color (grey, unselectable currently)

(upon hover-over)

[weary turret]

I think that’s great.

[prime socket]

So, I can make it yellow instead of green and clickable. If you hover-over it, it'll tell you why it's yellow: Your base budget will be converted into metal.

[prime socket]

Or, I can do dis

[weary turret]

I don’t like a yellow ready button. You’re ready or you’re not

But you can still have the tooltip over the green button

[prime socket]

Well, the button would have be something ... weird. To get your attention. If it's green you'd just turn your brain off and click without thinking to hover-over first.

people click and often move their mouse off the thing they clicked.

[dusty marsh]

can we add yellow stripes to border or something but leave button green?

[prime socket]

ew sounds like work

unfortunately that wasn't made in rmlui which makes it non-trivial to do any changes to the texturing, borders, etc

I'm a bit spoiled

[weary turret]

Look, this isn’t a game breaking issue. You get your money back. Pleasant surprise. We are just adding some additional context

We don’t need to be too heavy handed

I like the tooltip on the grey and the green buttons. I like the label attached to the top bar

I’d do all three things

[prime socket]

https://tenor.com/view/why-not-both-why-not-take-both-gif-11478682

[dusty marsh]

https://tenor.com/view/the-road-to-el-dorado-both-both-is-good-gif-8304204

[weary turret]

But to be clear, no yellow or middle state for the ready button. That button is binary off/on

[dusty marsh]

I'm happy with whatever the result, thanks again for the mechanic adjustment.

[prime socket]

not sure if I'll get the greenlight 100%. I'm pretty confident with an explanation, there's a really good shot.

I'll know sunday.

@weary turret thanks for going back and forth with me

you da man

I'm thinking on the ready button thing -
I don't wanna just make an exception just for quick start because it's just so special. What would game modes WANT the ready button to do?
I've seen other games and websites have a counterdown start upon hoverover the ready button with a popup saying "Hey did you forget this? You sure you're ready?" and then allow you to click the ready button. Something like that sounds more useful than an easily-missable tooltip in isolation.

When it comes to touching vanilla widgets I bigtime prefer creating generic functions future generations can use instead of per-feature exceptions.

[weary turret]

The disabled button tooltip is great though. Add that for sure

[prime socket]

I have good news! It's already there. And it's generic no less!

Call a function, add a reason to block, it's concatenated into the tooltip as a newline. Once all reasons for blocking are gone, button turns green. 🙂

[worn torrent]

Well, yes, for that not to happen you would need to price the stuff much more expensive on budgetBPtime extra cost

[prime socket]

One budget per second of Commander build time is not enough?

[storm moat]

Could make the button pulse between yellow and green if full budget is not spent

[prime socket]

The boss says no third state

[storm moat]

Pain. Oh well

[worn torrent]

Commander got 300bp, if you are now gonna let people not finish their budget intentionally then that is gonna bite them excesively.
So you will have to increase bp related budgetcost to make up for the "time" advantage somehow

[sour spear]

If that can be set as a seperate option of excess turns to metal I am all for it.

[prime socket]

What's your concern?

[prime socket]

I suppose mine is that it might be discovered that there may be a rock paper scissors arrangement that happens.

• all in spending base budget
• saving some for early defense
• saving max for max offense
  Where max offence strat might work out in metal starved environments like greenest fields or in naval battles where the winner gets the reclaim and the reclaim snowballs. If it's enough to make up for the buildtime, oooof.
  The defense against that is to save a little extra base budget yourself to make just enough defense then out scale your foe.
  This gamble at the beginning could decide the game before it really begins. May feel bad because it's effectively luck if you guess right.

[dusty marsh]

Isn't that similar to the decisions you have to make at the start of most games depending on map mex/energy/rush distance/reclaim? Do I go 1/2/3 mex, do I go solar or wind, do i go con first or pump a few units. Depending on the map you might lose by going 3 mex + con first on some maps if your opponent goes 1/2 mex and spams units. (alternatively there are maps where with long rush distance/safe bases rushing units first would put you so far behind someone with a greed open that it's also really bad)

Having the tradeoff of losing out on some free BP to make the choice to get more starting bank is better than trying to account for the bp in the budget somehow imo. I personally think it's okay to have the negative to not using your whole budget be losing out on that free BP which also gives some small incentive to 1: not always go all offense open 2: not save as much budget as possible to commie

[prime socket]

Because one budget cost equals one second of build time, that means not building structures gives you that much metal in exchange. Not sure how close to fair that is

I fear that it'll either be super underpowered and thus be a trap, or it'll be potentially overpowered leading to cheesy strategies

[dusty marsh]

I think that's probably under costed just in refund per BP personally. A com BP is 300 and a t1 lab is ~5000 BP. That means you get refunded 16.67 metal for losing out on 5000 BP if I understand it correctly which isn't nearly enough to pump BP to catch up on that loss quickly.

Note that I think under costed is good in this case even though I'm the one who wanted the ability to refund more, I think it would be too good if you got a bigger refund.

[prime socket]

It might be best to think of it in terms of the question "what should the resultant balance be?"
Projected on specific types of maps

• resource rich
• resource starved
• popular/common

[dusty marsh]

I'm heading to work now I'll pop back by later. I'm open to running some different open tests too if someone would like to.

[worn torrent]

This seems like a good thing to me, it imitates the investment in early units vs eco-time greed.
But you have to make people pay for bp-eco-time more as eco-time already pays for itself overtime in a cumulative way.

[prime socket]

Can you translate this into layman's terms?

[worn torrent]

More BP and E related cost beyond what you already have, start by x2 I guess, probably will need to be more than that

[prime socket]

That would put the energy cost conversion at 30:1
That would make one budget equal 2 seconds com buildtime

That would increase the budget cost of energy converter from 30 to 60

Cost of windmill goes from 47 to 60 (cortex)

Solar goes from 160 > 168

[worn torrent]

building area still massive and not budget referenced

[prime socket]

I think a greater objective is to have rough parity between the in-game costs of things with metal and energy and the budget cost

[worn torrent]

what would be the estorage and econvertor cost is the question

[prime socket]

You can't cheese terrain elevation differences anymore

[worn torrent]

basic budget is fine, but the building area still is too much regardless of the cheating terrain elevation, could use the trim

[prime socket]

Might go down to 450 but beyond that you're going to have to burn me at the stake

Currently it's 500

[worn torrent]

with basic budget you would only be building little more than what your starting building radius is

it can definitely go down

with higher budgets you should reflect that by having increased building ranges

[prime socket]

People are spread out quite a bit by the two minute mark that's what the range is simulating not how much crap you can pack inside of it

[worn torrent]

people cant spread that much and also build same as the other player who did not move

and being unable to price higher (buildings placed further away) from the exact start coordinates the only way to reflect that is by having a reasonably smaller building area

[prime socket]

Until I start seeing some degenerate gameplay I'm more interested in getting the refund amount right to allow for starter variety

[worn torrent]

this is definitely cheating the terrain buildable area (of a any normal game)

[prime socket]

WTF
I thought I fixed that.
You're not in an old developer version of the game are you?

[worn torrent]

maybe xD

let me go on multiplayer

[prime socket]

Just check the developer tab in the chobby menu

[worn torrent]

Cliff building is fixed, do trim the area though. Particularly would be good if you can bind a different building area to each different budget to reflect that the biggest of the areas still needs a trim to 450 as you said, the other ones trim a bit each) It would also help discourage the undesirable battery mentality.

It is fixed, though it isnt accurate

No need to change anything if you actually restrict the building area

[worn torrent]

All those cost would be reasonable, as I said, making Eexpensive stuff is basically a nogo in the early game cause there is no E, you have to make eproduction before you get to even think of doing estorage or conversion

if you build estorage and conversion it basically means you are a battery and need no lab

[prime socket]

This is a technical limitation. Is extremely expensive to iterate over every point in the grid checking if it is close enough to the one that's off limits, for the possibility of being reachable with build range.

Implementing that I suspect would cause a very very noticeable hiccup at game start. Very sad times.
So you have to put up with simple traverseability checks.

[worn torrent]

fair, just cut the range xD

[prime socket]

NO

[worn torrent]

you have to xD

[prime socket]

https://tenor.com/view/my-precious-gollum-lotr-rings-gif-26678083

[worn torrent]

otherwise you get the com walking over (around) the cliff at the start of the game

[prime socket]

I'm okay with this. That's why it's green.

[worn torrent]

what else can I say xD

can you make the size budget dependent?

or thats a no go (code wise)?

[prime socket]

No, I'm really averse to doing that because I want the commander to be in between the base and the middle of the map. The base arrangements start getting really ugly automatically generated when the range gets too small

Having Dynamic build ranges means I have to handle balancing all of them. On all possible maps. All possible budget configurations. Triples the workload

That isn't it though, I like having the larger range to allow you to have more freedom to arrange your base how you like. The more restrictive it is the less fun it is and the less likely it will receive adoption

[worn torrent]

can you not cut 50 range each?

[prime socket]

:{

[worn torrent]

I would say bringing in the competitive side in will bring you the adoption

[prime socket]

What is the downside to having this wide area? Why is this bad for competition again?

[worn torrent]

cause it doesnt imitate real game, you have to simulate it in cost and distance, just try it xD

[prime socket]

I don't think there's anybody who's tried it more than me lol

However you do make a good point. The range is based on assumptions on my part. It would be good to look at various early game builds but that will take a long time that I don't really have.

[worn torrent]

do small budget with the increased cost and (450 - 400 - 350this one for small budget) area, just a screenshot in a couple of regular maps like canis or theta

[prime socket]

You tend to be on the extreme conservative side. I would be interested to hear more perspectives from other people who have observed lots and lots of games.

I also need to observe more games myself honestly but I won't be able to get up to speed enough to be an expert.

[worn torrent]

let it autogenerate also in one of the screenshots to see how it would look like

[dusty marsh]

Is the intention for this to even be used in competitive games?

In my experience quick starts typically tend to be used in RTS more for fun games like metal maps/raptors/scavs/lava/etc which is where I've been seeing this used very often already. In this light I think the more fun option (bigger build radius) beats out the more competitive option imo. Ideal would be an adjustable size but I'm sure that's work to implement.

I think aiming for balance is still good but I wouldn't want to see quick start become part of the competitive environment personally (I'm still salty about starcraft adding the extra builders)

[prime socket]

My priorities
Primary:

• for use in competitive time limited game modes to provide more t3 time (territorial domination)
  Secondary:
• for use by casuals who either don't like getting bumped in the early game or find the early game boring.
  Tertiary:
• be palatable to competitive players such that it can be used in a competitive way for those tired of the slow beginning or grinding 1v1 practice.

[open oak]

do people really enjoy the first minute of the game?
you're literally waiting for your com to poop out bp

[safe pier]

The main thing is that a lot of creativity still happens in the first min.
Just look at omega series

[open oak]

literally the fastest route to something resembling interaction is straight into lab, which still means you're sitting on your ass staring at your com for 30 seconds

[prime socket]

I have a feeling that quick start might actually make for a more interesting broadcast. What you decide to do with your starting budget is immediately revealed at the beginning of the game giving the broadcaster a moment to theorize what they are going for until the action happens.

[open oak]

there's a reason zeroK made firstlab free

[prime socket]

Coop, mostly.
Speeding up the early game was a happy coincidence.
That's what I remember from the blog post anyway

[open oak]

ye

[safe pier]

Using omega as an example of potential opening diversity.
Not casting
-# but yes, it is more fast paced

[prime socket]

This was a while ago, but the most recent interesting change being explored:

[open oak]

quickstart doesnt dimining opening versatility tho
just cuts down on ass-sitting

[prime socket]

Make a larger amount of the starting budget refundable as metal.

Normally you are deducted 800m, 200e at start but this gives you back up to 500m if you don't spend it as budget.

[open oak]

UI comment btw: if you're refunding, it might be nicer to just show how much metal you start with
instead of magical resource units

[prime socket]

I hate how right you are.

I can't think of a way to present that in a cohesive fashion right off the top of my head. I will think on it for a bit though.

Because if you don't know that you start with 200 metal base, then you don't exactly know how much metal you are going to start with.

[worn torrent]

It is a gif, it works if you press on it 😂 I swear

That makes much better sense yeah, the quickstart budget should be like real game, whatever you dont use it remains in the resource bar and the best indicator is to tell you how much resources you are actually gonna end up with.

I do worry about massive coop issues but then again if there is anyone bother about coop issues they can pay a visit to the meowcore.

I do also worry about how veh is gonna be just so much better than bots with quickstart and that the balance team is not gonna move a finger about it

[prime socket]

I'm kinda stuck thinking about the visuals of it - how to communicate "It's a budget! OH ACTUALLY it's gonna be your metal now!"

I'm not so concerned about the coop potential.

• first factory discount (my preferred method) is still there, waiting to kill COOP dead when it rears its ugly head.
• you're sacrificing both com walk time and buildtime and a measly 1 metal per second of buildtime is not very good. Even 2 isn't great.
• at the beginning of the game, you NEED the com's buildpower at the factory to pump out units quickly. If you don't, you'll waste resources.
• stacking coms nearby risks random bullshit happening due to overlapping base builds. Jank, but, there's no technical way to prevent this easily anyways.

Like, with the default budget, oh my goodness. You're already teetering on overflow with the base it can buy you.

[vale idol]

If it ruins co-op we will just not enable it for co-op, what's the issue?

[prime socket]

No no, the opposite, the concern is that it will make Co-op trivially easy to execute

After seeing it in action, I don't think that's likely

[vale idol]

Can't speak for everyone of course, but I think the majority of those players are not looking for a quick game anyhow

[prime socket]

Survivorship bias. People who can't stick around for potentially 45 minutes don't play until they can stick around for 45 minutes.

I'm not sure what the upper limit should be to a Time limited game mode. Territorial domination does so with an objective. If you don't want all those pretty colorful squares, something else more suitable for vanilla is necessary IDK what that should be yet

[prime socket]

Maybe part of the bar is grey like the metal bar

Maybe the whole bar turns grey like the metal bar once you are getting into the remaining budget.
Maybe base budget gets renamed "Starting Metal" when it gets low enough.

Maybe "remaining budget will be refunded as metal "additional base structures will deduct metal."
Wherein the bar empties at 200 metal to prevent noobs from doing stupid shit like starve themselves of budget needed to pump out first con

The number to the left of the budget bar bottoms out at 200.

I might be able to apply some kind of purple tinged gradients to the base metal color to imply it is still usable as a base budget

[prime socket]

@weary turret bat signal this^

With this being normal display before dropping below the Starting Metal threshold.

[open oak]

Just dumping my opinion: i'd remap it to metal, instead of purple budget.
Thing cost gets remapped to how much starting metal you get for them.
Instead of going through "abstract budget" just remap it to metal. It's 1:1 with metal anyway, no?

[open oak]

So this direction, my gut says

[prime socket]

But the budget costs of items are not == to metalCost alone. It's a player-unknown formula.

[open oak]

Does that matter? It's on the buttons, no?

They just cost more metal pre-game

[prime socket]

The purple bar is to create a mental causal link between the purple squares around valid build locations and green invalid ones for base spawning

And the purple numbers in the build menu for unit costs

[sour spear]

Not a concern just a way to expand on its variability for games. (Also that bug with people on low end machines might still exist and be really problematic still)

[weary turret]

Or maybe

[prime socket]

@weary turret rats observation is that what the player will want to know is how much net metal they will have at game start.

Currently if you use all of your base budget, you start with 200 metal. With this rework it can go as high as 700 (and maybe in future 1000 idk)
That's the number they wanna know

Maybe that banner says "you will have X metal when the game begins."

[weary turret]

Oh I see

Hmm

[dusty marsh]

For me this would be the most clear/succinct way to present the information rather than just showing what the budget refund amount would be. The only other potential addition I could see would be maybe having the portion of the budget that is "forced" to be spent in a slightly different shade or with a small dividing line between the "forced spend budget" and the "optional spend budget"

[prime socket]

like this?

[prime socket]

To resolve the above
Disallow building past the red lines where the circles corners intersect

[rugged robin]

i can't really see the reason you want to jump to a 2:00 min buiild immediatly. but here is a voronoi library. if you scroll all the way down you cann see the function you might be most interested in. checks a position inside a polygon.

https://github.com/TomK32/iVoronoi/blob/working/voronoi.lua

dont know if a 1000+ line lua script is too much?

[open oak]

Very tldr: most of those first 2 minutes is sitting on your ass while your com builds your queue

[prime socket]

Thanks for the link

[prime socket]

Ratty said one of the reasons. Here are mine in descending order:

1. Time limited game modes that last 20 minutes for example suffer greatly from a slow start. Territorial domination benefits from taking the first 2 minutes of the early game and instead putting them into the late game for more T3 time.
2. Some people are tired of the repetitive first 2 minutes of the game.
3. It let's players skip ahead to the T1 main battle phase, the fun part, skipping past the super early game where the enemy can raid your early expansion potentially gimping your whole game.
4. Practice aid if you're grinding out 1v1's for a tournament for example

[quiet gale]

4. would save/load game just work for grinding a build order?
   #1414560036595826789 message save/load just has this bug atm

[open oak]

i think quickstart's bugged now, idk if there's a seperate bugreport thread
game_quick_start.lua:668 atempt to perform arithmatic on field ? (nil value)

repro:

• place things
• click ready
• game stutters half a second, nothing happens and throws error in chat

[prime socket]

Does this happen every time or sometimes? Is there specific maps? Configurations?

[open oak]

I think always? Did 3 games monday evening on 3 different maps, and once in a skirmish this morning to check the exact error it throws

[prime socket]

Yeah, I bricked it with my last "fix". I have a fix for the fix that will fix it for sure.
Probably be fixed by weekend.

[prime socket]

The fix is in, please test it out and make sure it really is fixed

[glad lintel]

https://tenor.com/view/fixed-gif-14953349

...this is how I feel like adding nil checks to 'fix' things/other peoples code

[prime socket]

Facts

[sour spear]

If there a possibility of making it so a AI or player with a bonus has cheaper structures for quickstart?

[prime socket]

What's the appeal?

[open oak]

i think he means "multiply quickstart budget with the bonus a player has"

[worn torrent]

On a normal game you dont get extra initial budget, just future i come

[prime socket]

Yes but why would you want this?
It would be a instantaneous advantage on top of the income multiplier advantage. Let's say it was 50%, you're not getting 150% income at the very beginning now, you are getting 150% x 150%

That stage of the game is the most difficult one to survive when the enemy has a resource bonus already.

[sour spear]

A discount scaling to some degree via the bonus might work but not a exact percent discount. Aka a partial discount.

[prime socket]

Anybody using legacy layout? Can you test and verify the budget costs of things (purple) in the build menu are there not metal/energy cost in the pregame?

[dusty marsh]

no it's metal/energy

[prime socket]

it'll be fixed when next stable releases

sorry about that

[prime socket]

@worn torrent had a chat with the gdt, the majority of the competitive players on that committee agreed with you that distance from the commander should factor into the cost of the build

I'm thinking I'll keep up with the tradition of one second of Commander time equaling one budget. Which means, since the commanders walking speed I think is 30? Every 30 Elmo's of walking distance costs 1 budget

I think a simple distance calculation from the commander's spawn location isn't right.
Measuring distances deducting Commander build range in between builds is too complicated and expensive and unfriendly to noobs doing layout out of order

So I think the best option is to assume a circular walk around the commander's start location just far enough to reach the outer builds

Also, further reducing the default Quick Start build range too 435

[worn torrent]

Still think that the most reasonable thing you can do is
"do small budget with the increased cost and (450 - 400 - 350this one for small budget) area
Not sure what the size of those areas are IG, just make it gradual reasonable for each budget. Again, it depends on what timings you are aiming to replicate, if you state the timings then it is clear how much people normally spread.

[worn torrent]

This is clearly not intuitive by any means, and will mean that people will build on a small area anyway to max the budget so what is the point? Go for the sensible building area sizes

[red summit]

Can you make like 2 or 3 total circles, where the outer 1/2 just increase the cost by 10%/20%? I think that should be enough and its simple to understand. Also shouldnt be that hard to convey

Or a flat cost increase would work as well. I dont think it has to be anything too complicated

[quiet gale]

just keep the build area reasonably small, no one should have to pull out a calculator to say this building costs X here but Y there

ez to understand > trying to mimic vanilla balance

[worn torrent]

Actually both things are the same in this case, it is just a matter of accepting that being reasonably sensible is good for all publics

[prime socket]

Also @dusty marsh your request for refundable budget as metal is approved as I implemented it

[prime socket]

@sudden oracle re: #1445197631222452226 message

The build range currently is 600. I thought I nerfed it to 500 but I was a dirty liar. It is still 600. Lol.

The next stable release will bring that range down to 435 (llt range) for Medium/Small/Default

Large will retain 600 range to provide more space for the giga-sized budget.

The commander's default the build range is a mere 250, I didn't want people to squeeze all their base into such a tiny area. Additionally, the objective of the feature is to skip ahead to the 1:30-2:00 mark which includes a little bit of base expansion. You usually have four or five metal extractors up at that time

[worn torrent]

Finally you came down to a sensible size 😂 it should be quite good

[prime socket]

Okay a new big update happened. @dusty marsh 's suggestion to allow you to refund budget as metal is in the game now. Reload and try it out and let me know if you see anything sus
Also range reduction 600>435

[stiff hazel]

Yay, I like the metal refund

[foggy compass]

fyi i used this a few days ago and it ended up being quite intuitive and easy to use. i initially had skepticism about this feature, but now i think it’s quite nice (in the right lobbies of course)

[prime socket]

Please try to find disgusting Co-op abuse. They will come a day of reckoning when it will be decided whether to keep the default quick start or factory discount version

[stiff hazel]

I don't do coop 😭

But i will try

[prime socket]

Wait a minute I know some guys

[red summit]

If the first factory has a discount, then what immediately comes to mind is: immediately after the game starts, rush a con out, eats the lab for the extra metal and give it to somebody who still has a lab.

Perhaps if 2 peaople cooped a bit by one person reaclaiming the discounted lab as well has having some surplus metal by not spending all their starting fund and gave ot to the other Player, that other Player could immediately start a t2 lab, maybe blow a com and have a t2 lab in under a Minute

Would have to test that, but should be possible

500m refund + 1000m reclaimed labs + 1250m com reclaim = 2750m, just enough for a t2 lab. The only bottleneck is bp.

Sure, it takes 2 players, with 1 having almost nothing, but most units need a lot more then 1 minute to traverse most maps, at ehich point the enemy will already have t2 available

[prime socket]

compare that to just having the full cost of the lab to funnel wherever instantly

also you lose the e cost

[foggy compass]

also there’s the BP cost

while you’re using comm and con for eating lab, they’re using their BP to boost out units or scale energy or smth

so i suspect that the discounted labs at 0min for everyone means that trying to eat the lab and funnel to a t2 player just results in pawns in your base at 30-60s into the game

in other words, commie no longer meta (but idk if this intuition holds up in a real game, someone would have to test)

[worn torrent]

Dont worry too much about coop, quickstart wont be able to prevent it due to its nature.

Doing what he described will get the coopy people killed in that scenario.

[odd loom]

tbh this is probably my favorite feature added to the game in quite awhile. Im seeing it used quite a bit in custom lobbies of all kinds.

[prime socket]

Thanks for helping shape it

[odd loom]

Thanks, yeah I hit my busy season a few weeks back which is why I just kinda disappeared Ill be fully back after the 1st

[prime socket]

keep an eye out, i've seen several people complain of lost build queues. I suspect it might be related to accidentally pressing the stop key

[storm moat]

Ive noticed it happen somewhat often in vs ai matches while testing. some ais will just have a total of like 3 mexes, 4 winds, and a solar, with no lab or other buildings constructed. Might be more than just a stop command there

[prime socket]

yeah, I think that'll work

[vague obsidian]

I stumbled upon this thread because I found a lobby that was using it, no idea it existed!

I'm really interested in the factory discount development because I had had a similar thought to try making a "first lab free but reclaims for 0" as a way to mitigate the amount of starting resources that contribute to funneling early t2 lab timings

i see that some of those concerns are discussed, and i was wondering if theres been any testing with making the first lab not give metal when its reclaimed (to encourage all players to make and maintain control over their first lab and have less early metal to sling to break timings in 8v8 that would otherwise not be possible in 1v1)

[dusty marsh]

@prime socket FYI, I've had a few times where I have buildings dropped from my pre-built base.

Usually it's a mex on the edge of the build radius although it's definitely purple and fully inside the build circle and isn't going over budget since I make use of the extra starting metal option almost every time.

[vague obsidian]

I see that in the original post, it is called out as a potential coop solution and i REALLY like this

[vague obsidian]

I submit the APEs for COOP abuse testing

[dusty marsh]

Fwiw I love quickstart for fun games but for core vanilla gaming I'm opposed to it ever being the default.

I like starting with one unit and building from there, it gives a lot of room for micro skill expression in the opening minutes even if currently it also allows for the opposite of that with full coop.

I also didn't like starcraft 2 making every game quick start though and most people do so maybe I'm weird.

[prime socket]

I've seen it too. I'm investigating why. Something gets desynchronized between the GUI and the game side of things. Maybe I need a different inheritance mechanism.

[prime socket]

Glad you do. Reclaiming lab wastes the energy and takes time and also wastes the buildtime it takes to build a new one - perhaps that gimps early rushing strategies because I'm the beginning you only have your com's 300 BP to do things with.

[prime socket]

I've thought about it a lot. I told the GDT I don't want to see this in the vanilla game - let's call it classic. It absolutely should be the default in any time limited game mode (such as territorial domination) because more lategame time is better than early game time.

[dusty marsh]

Agreed, hmm I really need to give territorial domination a try, thanks for the reminder!

[solar lynx]

Yo this is sick! I played with the base building construction turret in the games on new years eve and that was sick. i think anything that can progress the game faster is a great option - not that I am at all bored by the early game, it's just better from a retention perspective IMO.

[prime socket]

Funny thing, it was your videos that got me into this game to begin with. Now I made a feature you think is neat. Butterfly effect eh?

Feature is done. You can find it in the extra's tab.

[solar lynx]

That is so super cool. I love that. Validates why I make content. Gotta help contribute to growth!

[vague obsidian]

https://tenor.com/view/now-kith-mike-tyson-gif-13249709

[solar lynx]

i pay the same amount for winds as arm?!? HELL YEAH BROTHER

oh god what have i done GO SOLARS GO SOLARS

god honestly this feels so much nicer playing ngl.

[prime socket]

The people spoke and I listened
https://discord.com/channels/549281623154229250/1437173642814292059

[prime socket]

For advanced players in a tournament? Maybe not feel good. For a new? Not dying to early game raiding before you even can get an army going probably feels pretty good

Nobody has really figured out the actual optimal starts given the budget amounts yet.

[solar lynx]

That's where you're wrong kiddo.

[velvet widget]

hi, I just want to say that this feature is awesome, thank you

[solar lynx]

weeeeeeeeeeeee

idk if walking or labbing is better on front, instant lab + units immediately seems CRAZY strong

just like being able to do this at the start and immediately get into the game is so fkn nice though

[prime socket]

I wonder how risky rushing T2 is if the enemy is going all in on T1 units immediately. It could mean immediately losing a lane because you're down one teammate worth of units. Probably map dependent.

[weary turret]

Maybe feature this option in a video or two so more lobbies try it out? I’d like to see some high level lobbies genuinely give it a try. My casual friends love it when we play barbarians. I don’t know how it would change higher level PvP.

[prime socket]

Hey did you guys notice you can move around your friendly AIs by default now? 😮

In the cheats tab, there is a toggle that lets you move around hostile AI as you want too

[weary turret]

Pfft we don’t play with FRIENDLY AI 😉 but yes, when I play alone with ally AI I love that I can move them around now. It’s great. Now just get them to stop building random shit in my base or going air for no reason!

[open oak]

ngl this is one of the best QoL features lately
also giving hostile AIs a set spawnpoint is great for PvE

[prime socket]

Meanwhile

[storm moat]

atm, with no coop restraints, you just build a bunch of eco, give it all to one player, then comm walk for bonus mexes. They can get their t2 up astoundingly quickly. You might need a couple players to go t1 to prevent too much damage being done, but ya it's not amazing

[prime socket]

You waste resources immediately because it's throttled through that one players 300BP commander.

[storm moat]

fair point, but just comm walk the nearest two comms to the lab that is placed as communally as possible

[foggy compass]

although to what extent is this balanced out by the fact that the opponents can start producing pawns and grunts to kill you with at 5s into the game instead of the usual 1m+?
i.e. the anti coop “feature” might just be that the enemy who uses their discounted lab instead of doing t2 commie will be able to attain and abuse army advantage faster

[prime socket]

https://private-user-images.githubusercontent.com/165520713/534277242-fbf1cda2-18df-4c9b-a147-7eaf5a89405f.png?jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3NjgwODkyNjMsIm5iZiI6MTc2ODA4ODk2MywicGF0aCI6Ii8xNjU1MjA3MTMvNTM0Mjc3MjQyLWZiZjFjZGEyLTE4ZGYtNGM5Yi1hMTQ3LTdlYWY1YTg5NDA1Zi5wbmc_WC1BbXotQWxnb3JpdGhtPUFXUzQtSE1BQy1TSEEyNTYmWC1BbXotQ3JlZGVudGlhbD1BS0lBVkNPRFlMU0E1M1BRSzRaQSUyRjIwMjYwMTEwJTJGdXMtZWFzdC0xJTJGczMlMkZhd3M0X3JlcXVlc3QmWC1BbXotRGF0ZT0yMDI2MDExMFQyMzQ5MjNaJlgtQW16LUV4cGlyZXM9MzAwJlgtQW16LVNpZ25hdHVyZT0yMzA5YjZmZmQzMWRlODZiMzBmOWQzNTFhMGI4YWFhODI1ODFmMDlmNzE4ZTY1YWFlZjgxODY3ZGZjYmMwYThkJlgtQW16LVNpZ25lZEhlYWRlcnM9aG9zdCJ9.FLTXskQeljlR7ni4YB8phNyuBauQ0TTKHWzdzF5X6lQ

https://github.com/beyond-all-reason/Beyond-All-Reason/pull/6520 This PR needs stress testing. I think it's ready for primetime.

[prime socket]

@weary turret yo check this out

I don't wanna do super complicated stuff for overlapping ranges. Right now I draw a line from corner to corner where the build ranges overlap. That's what the red line is.

Should it be a different color like purple or green?

This whole pregame handling of spawn locations, build distances, minimum placement distance, and interactivity with moving friendlies could use some proper design work tbh

[weary turret]

I’m on my phone but red shouldn’t be the color. That’s attack range. Green or purple are my first thought

[prime socket]

You got it boss

I'm thinking people because of quick start association

[weary turret]

What is happening lol. Again, I’m on my phone but that looks really confusing to me

Maybe I need to spend some time think about it to form a good opinion

I thought we were just doing merged circles

That top one looks like the beginnings of a summoning circle 😄

Are we trying to delineate where two commanders can share a resource? What’s the problem we’re trying to solve with the additional lines?

I’m wondering what decision we want the players to make based on these new lines

[prime socket]

if you're too close to a friendly, we need to show what part is yours and what part is theirs so we don't have conflicting build spots.

This is effectively a fence.

@weary turret

[weary turret]

So even though I have my circle there, I can’t build over that purple line?

Hmmm. Do we need to block players from overlapping circles? Why not allow players to place wherever, but if two buildings overlap we just remove them from the board? Just like how if you place a blueprint and then place it again on top it cancels it

This would require everyone to see everyone else’s pre-game builds before the game starts though

[prime socket]

Because you can't see what everyone else is trying to build, and it would be extremely bad performance if I tried to show what everyone else was trying to build

[weary turret]

Hrrrmmm ok

[prime socket]

A 8 player game would likely be noticably laggy to display all the build things people have queued up for potato pc's

We do need to handle this because tiny maps with many players and tiny start boxes force players close to each other. Normally that isn't a problem with quickstart big problem

[weary turret]

Can you turn off the green circle, make it purple, and actually reduce the shape?

[prime socket]

i hate it when you're right

[weary turret]

Ya I think that looks a lot better to me

[prime socket]

Totally. I was being a bit lazy lol

[prime socket]

whenever you get time, review the whole pregame ui including quickstart, minimum distance between spawn locations (lil faint red circles around ai's) etc and let me know if there's anything stinky

[prime socket]

Aight, it's in the game now. With that I think QuickStart is a finished feature.