« first day    last day (32 days later) » 

01:46
I think the rules on fungi are confusing and it doesn't seem to me like farming fungus is very effective. I think it would be better if they had a shorter timer (maybe 1024?)
Maybe they should also spawn spores in the adjacent cells when they turn into food.
Also: do the ants have a sense of direction?
Cause the original randomizes facing direction
Also, I'm a little confused about payloads. If I understand correctly, ants can drop food they're carrying onto adjacent cells or they can spontaneously produce a fungus spore on an adjacent cell.
Does food carried by ants trigger fungal growth?
02:08
Question 1: You can surround fungi wih food to make it grow faster. Give it 32 food in a vicinity, and growth happens in 2048 turns. Sounds like a long time, but I wanted a game where farming food isnt the obvious best stratey
gy.
Ants have a sense of relative direction in the short turn, and you can give them long term direction by keeping track of turns in internal state
Fungus and food are separate entities. Fungus grows in the presence of food, but unless planted, it simply doesnt have a presence in the game.
02:20
Food carried by ants does not factor into fungal growth.
Is facing direction randomly rotated, as in the original challenge?
Only for the queen first turn
Workers are oriented so the space the queen was when spawning them is "behind" them just as if they moved from that space
After first move, both queens and workers are oriented so that the cell they moved from is "behind" them, and remains so until they move again
 
3 hours later…
05:06
If I choose to stay still, does my facing direction remain what it was in the previous turn, as there is no orientation that indicates "stayed still"?
05:17
> Ants can see the color of empty cells, and the properties of objects in filled cells, but not the color of filled cells.
Can an ant see the color of its own cell?
 
4 hours later…
09:43
@trichoplax My guesses are 1) yes and 2) no. I'd argue it's not even vague - just not explicit.
a thing that worries me is that nothing is permanent in this QotH. It'll be pretty damn hard to defend a farm from an enemy avalanche
oh there is no stopping a queen just instantly raiding a whole farm..
If the priority was reversed then a wall of workers could defend a farm
Not sure if that would be more or less interesting
@trichoplax it'd also make queens able to die
10:08
> If the number of [unladen enemy workers] is equal to or less than the amount of food carried by [an] ant, then the ant loses food equal to the number of adjacent enemy workers.
Does that mean that a laden-unladen pair of rival ants will constantly swap food?
@dzaima You can try to steal food from her... At a fairly large risk.
Well, obviously she can just choose to ignore workers, at which point you're screwed.
An idea would be to make static workers (or perhaps even static laden workers) have higher priority precedence than a queen. That way, entering a cell with a worker in it could be made illegal for a queen. (not sure if this is what @trichoplax meant)
If the answer to this is "yes" and only laden workers may block a queen, then it would be possible to penetrate such a worker wall with the help of an unladen worker (which would literally steal the food from the workers forming a wall, making it penetrable).
10:23
oh there is no FTL info transfer so you can't even make a system of carrying the info of enemies to the queen so it can collect all the food too
You can carry information at the speed of 3 cells per step though (with the use of fungus and a vision range of 2)
With the proposed precedence change, I think queens may get permanently trapped regardless of their actions if an attacker is persistent (and rich) enough.
Yup, because laden workers may murder stationary unladen workers and food leeching occurs after collision resolution.
Do newly-spawned workers take part in food leeching?
And does a newly-spawned worker count as a stationary or moving one in the precedence ranking?
Oof. Important edge cases.
10:54
with no rule changes the time of no disturbance required, after which all everything can be lost, becomes very important.
For 1 food surrounded by 80 fungi that's 829 frames (but in the 30k frames there'll be absolutely no food gain)
For a 16x16 checkerboard it's 2209 (1 food/17.3 frames)
For a 16x16 checkerboard surrounded by 3 layers of fungi - 1353 frames (1 food/10.5 frames, quite an improvement from above)
@Alion This is probably incoherent to anyone but myself. Assuming the proposed changes, the only real way for a queen to tear down (or at least get through) a wall of laden ants is to steal food from them. The only way to do that is via an unladen worker made to steal food from the wall, but since laden ants have higher precedence than unladen ones, and because of the action resolution order, laden ants will be able to kill the unladen worker before it has a chance to steal anything (cont...)
Therefore, a queen surrounded by laden ants has no way of escaping (assuming a thick enough box so that no holes will ever appear due to attempted assassinations of enemy unladen workers). Also, a queen won't be able to do anything at all once she is perfectly surrounded by 8 laden worker ants.
Heh. Nevermind! I just remembered that she can kill enemy laden workers by putting food under them :D
This is glorious.
@dzaima How about 4 layers of fungi?
@dzaima Oh wow, a simple 16x16 checkerboard farm is really inefficient, huh? 1 food per 17 frames is almost doable in the old Formic... This is just plain sad. Considering how hard it is to defend, I'm starting to think there should be some shroom growth speed buffs.
@Alion the 4th layer only makes one food per 16k frames which is bad enough for me to not consider. But anyway, 1293 frames, 1 food/10.1 frames
A'ight, yeah.
@Alion mhm..
Maybe not as drastic as @Beefster suggested (1024) - that would be plain insane (and would break the nice evenness of the timer). But 4096 frames per shroom seems... Alright, maybe?
It would definitely become more efficient than simply collecting food from the map pretty quickly, though.
11:06
for a more practical, but crazy, idea like this, it gets 20 food/frame with 1471 no intrusion frames (with the current rules)
@dzaima What am I looking at? A farm design?
@Alion mhm
oh wait that thing has many flaws
Is fungus a diamond or a red square? In any case, there are some problems indeed.
@Alion fungi are the small squares, food are diamonds just like in FF
Yeah, then some fungi are going to get trampled nicely...
11:11
@Alion a 16x16 checkerboard then gets 1 food/1.07 frames with 138 no intrusion frames. That feels a bit too OP but it's hard to say
Oof. A piece of food per frame seems insane. 30k turns and you've officially collected more than 11x the food on the map :P
(and that's with a 128 food investment; we're not talking about expanding the farm!)
@Alion I'd make a guess and say that it wouldn't be too hard to make a bot that collects all the food in the map in ~10k moves even without using fungi for farms
@dzaima Right, probing.
Wait, can the queen pick up more than one food per turn..?
No.
> If multiple of an object type are put into a cell simultaneously, the duplicates are lost, and resolution proceeds as if only one of that kind of object was placed in the cell.
Well, I guess transfer via suicide is also possible (although lossy)...
@Alion yeah, now entries are gonna have to be super careful to not accidentally destroy their own ants or their food
Oh, transfer via suicide makes no sense (because you have to respawn the suicidal ant, which takes one turn (and one food!) anyway). Whoops.
So that means that under the faster-fungi rules a 16x16 farm saturates the max food intake of a queen.
So how large would a farm need to be under the current rules to saturate the queen?
64x64, correct?
So actually, I think it's possible that even under the current rules the queen will be the bottleneck. Wow.
(64x64 is only a 2048 food investment)
11:30
@Alion that gives 1 food / 0.85 frames and a 1757 no intrusion frame requirement
And what is the "no intrusion frame requirement"?
@Alion it can get 100% raided every 1757 frames with nonnegative gain (assuming the raid is instant & repair is too)
Uh-huh..
Well, anyway, I think something needs to be done about the max queen food intake. This seems like a pretty bad problem...
that thing's also gonna be taking up 6.4% of the map vertically so collision with enemies is gonna be very often
Right, and depending on the ecosystem that may or may not be horrible.
Oh, right. Food farming is going to be a requirement anyway because food is relatively scarce in this edition. Mmm...
11:42
so if I'm understanding the queen moving logic, the queen (Q) here can't move anywhere because of the enemy queens (E) in the way?
Yeah, that's one example. Though I think cell 7 is protected by 2 queens, right?
> a queen may not move ... to a cell that her view shows is in the von Neumann neighborhood of an adjacent queen
Oh, well that's clearly an oversight by @eaglgenes101, isn't it?
@Alion why would it be? without that rule a queen would always have to move away from an enemy queen (or stay still)
Both the queens on cell 3 and 4 may move to cell 7, only one object is allowed on a cell and both queens are immortal. Whoops.
Am I missing something?
Oh yeah, I am. Right. Nevermind then.
You know what would be fun? Spawning queens.
11:55
@Alion of course it'd be fun but making that not be over/underpowered is pretty hard
I'm forced to agree.
> If food is within a 9 by 9 cell area centered around the fungus, then the fungus will grow.
Maybe elites at least, then? Can collect any amount of food, high in precedence, all food stored by them counts towards victory, but really costly and can't spawn workers.
This would neatly solve the food throughput problem.
Spawning queens is a separate contest I'd be interested in
But then the scoring probably wouldn't be food related
 
5 hours later…
17:14
I'm thinking of making it so redudant pickups result in just lost food rather than dead workers, and make it so multiple food collapses into a single one only on a cell with just food
This should make it somewhat easier to defend, and remove the throughput bottleneck
@eaglgenes101 Well, shift it from 1/frame to 8/frame, at least...
@eaglgenes101 You might run into a certain issue, and namely a single worker raiding entire farms (well, until it gets stopped by presumed guards at least)
Ehh, I guess that's also possible without that change, just slightly less efficient and/or destructive.
 
2 hours later…
18:57
at first I thought that parallel execution would open interesting options of play but it seems to just be making queens way too OP and distracting from the goal of farming
Farming, raiding, and defending against raids
@eaglgenes101 the problem is that defending against raids is impossible if a queen is participating in the raid
well unless you have workers constantly standing by each piece of food ready to pick it up in case an enemy is approaching, which doubles the starting investment, decreasing the throughput at the same time of the already slow farms
and even if you do that, in the case of a queen coming by you have to pack up & find a new home anyway, and even still the enemy queen could just follow the workers moving
oh you could defend against queen following by leaving behind workers that see a queen, but that immensely complicates settling down
for a 16x16 farm the worker approach ups the frames/food from 17.2 to 38.3 - >2x decrease..
19:23
assuming it takes only 10k moves to get the starting 171 food, this setup would get ~522 food in the remaining moves, but minus ~63 food per raid due to the resetting of the fungi growth
19:46
Yeah I'm going to have to come up with a mechanism that can prevent indefinite tailing of enemy queens
Wait the queen exclusion rules already do this Apparently
Hmmm
request: either pin the sandbox message or edit a link of it in the description
@eaglgenes101 what do you mean by queen exclusion rules?
Well the rules saying that a queen cant get too close to anoher queens
I might need to think on this
@eaglgenes101 that only applies to queen-queen interactions, not queen-worker
@eaglgenes101 workers risk following the enemy queen though, as they might get stepped on & killed
20:25
Pins can fall off after a while, but a link in the room description is forever
20:40
I think the growth zone for fungus might need to be shrunk, but the growth time should be decreased.
Those farms are insane
@Beefster I wouldn't call them insane, as they are about par to SlM in the original FF, with the minus that they are easily destroyed
What if it used a 5x5 zone instead, with immediately adjacent food causing 2 steps of growth per turn?
and then drop the timer to 50000 frames
I know we're programmers here and we like powers of 2, but I see no real reason for it here.
@Beefster The really nice timer. Check out the fungi section. Well, I guess you could split it pretty evenly anyway, but...
> we're programmers here
@Beefster without the adjacent = +2 rule (harder to implement), the 16x16 checkerboard gives 1 food/37.9 frames
Also, fungi can be used for probing for food from a distance. If we drop the zone size to 5x5, this mechanic is gone.
20:47
@Beefster immediately adjacent = the 4 in a + formation or the 8 surrounding the fungus?
That's actually kinda cool
^^^
I didn't think about that
It was mentioned in the spec even. I missed it the first time around too :P
Obviously, this changes the dynamics a bunch. I don't think the challenge is based around that really, but it's a nice-to-have that is quite difficult to make use of (and very effective when used properly).
Something like sliding miners or wildfire would absolutely destroy farmers
You'd have to have a lot of guards
20:50
@Beefster the problem is that guards can't defend against enemy queens, so enemy queen = pack bags & go find a new home
You could probably space them out somewhat since they can see ants coming
It's statistically unlikely that a queen even runs into you though
> statistically likely
FTFY
Not if you keep your farm small
Considering even in the old Formic entries tend to sweep the whole board at this point...
If you have a 50x50 farm, each other queen has a 5% chance of running into you horizontally
It's less vertically
The queens didn't sweep the board
20:54
@Beefster a 16x16 farm takes up 1/62th of the map vertically, a queen can wrap around the map 12 times in a game, there are 16 queens, 12*16 = 192 passtroughs vertically in a game, way more than the 62 that the farm can not get hit by
They did in wildfire, but most of the rest of them went in a straight line
Yeah, you might want to revisit that challenge...
I have an entry up that collects about 2400 (300 turned into workers) food on an empty map. There's 2500 available.
Sliding miners?
Nope, Hyperwave.
Though Sliding Miners isn't much worse.
the Sliding miners queen moved in a straight line
20:56
@Beefster Hyperwave doesn't though
I don't recall that entry
'coz it's new
like less than 2 weeks new?
Yea.
@Beefster That's not what I mean anyway. Even if workers bump into a farm they can inform the queen. They have 8 internal states and 8 colors available, for Pete's sake. Of course they can invite Their Majesty to a feast.
Point being, farms will be raided. A lot. Even by a single entry designed just for that (which we can safely anticipate). And there's no way to defend against that. And there probably should be, even if prohibitively expensive.
Then again, @eaglgenes101 said explicitly that he doesn't want to make fungus farming = free food. It's supposed to be a risk...
I thought hyperwave was monorail for a second
21:04
my suggestion would be to calculate worker outputs first, and make queens unable to move onto workers that won't move in the current frame
then I realized it wasn't moving in a straight line
@dzaima Well, do you really need to do that first? I mean, you can just say that "Queens may not move onto cells containing workers" and hey presto. Is there anything wrong with that?
And if you want to chase that worker, then tail him 2 cells behind. No serious side-effects to this rule (well, aside from being able to trap queens now (which may or may not be considered a positive) :P)
@Alion that would work too, but I wanted stationary workers - guards - to have more privileges than moving ones, making it harder to trap queens
@Alion maybe allow a queen to stomp enemies if it's completely surrounded by them?
@dzaima Seems too arbitrary...
@Alion true, but being able to trap queens feels evil
21:13
@dzaima How would additional privileges for guards make it harder to trap a queen?
@Alion if the queen is moving around, you've got to move together with it, and while doing that, you're not blocking it. With your suggestion moving workers can block the queen at the same time
@dzaima I guess that under the current ruleset you can still spoonfeed workers to explode them :P
@Alion oh that also makes my rule idea pointless
@dzaima But they risk colliding with her upon movement anyway.
@dzaima Not unless you have tons of workers to spare...
I think it is in eagl's vision to make fungus farming prohibitively expensive and very risky...
@Alion infinite workers to be precise. The queen can kill the worker which leaves food behind, and the queen can collect that food in the next move.
21:18
To make defending somewhat easier, I'm thinking of adding (Diplomacy)[en.wikipedia.org/wiki/Diplomacy_(game)] style supports to the challenge to replace the current precedence rules, and limiting leeching to queen->worker
here's a working link of the above
@dzaima Wait a sec, you can just funnel away the food, right? Because placing food is first and then picking up second. So you can't actually explode workers if the cage is properly maintained.
Ah, right, but you risk leaving yourself vulnerable.
No? Because ant leeching occurs last right?
Ahhh, this challenge is an edge case parade!
2
@Alion right, if all the workers continuously place down food they can not explode (and waste the enemies queens food at the same time)
@Alion currently all fungi are really useful for is keeping track of time
21:33
And probing. And farming when the map is already zeroed.
The idea of diplomacy-style support is that it makes the defender's job easier, as defending ants can be supported from behind, while attacking ants have only the comrades on their sides to help them
(you can't lose anything if you're really careful)
Should I go ahead with this?
Mmm I don't know if it's a good fit. It might be? I'm not sure.
The worker-leeching mechanics makes my head spin, now that I think about it
21:38
This entire challenge makes my head spin, now that I think about it.
The basic idea is sane, but actually working one's way through how to go about doing a raid is less so
@eaglgenes101 what of that are you planning to implement?
Remove worker-leeching as the means of breaking through walls of laden workers. Instead, an additional pseudo-payload of Support can be "dropped" instead, and if done, will reinforce allied ants that move into that cell, giving them higher collision precedence.
@eaglgenes101 Well, under the current rules, raiding is as simple as a) find farm, b) go back to the queen while trcing the path to the rally point, c) get the queen, d) go ham.
Support mechanics will allow walls of defenders that are robust but not impenetrable, while under current rules it's much harder to do such a thing
21:43
@eaglgenes101 I don't see how that rule change makes walls robust
now all it takes is an enemy queen & worker working together and the worker is in, quickly followed by the queen if I'm understanding the idea right
On the other hand, defenders can supply support of their own, and create a whole wall of increased-precedence defenders
it's either walls are penetrable or are not. And if they're not, farms are very ineffective.
how does the "support" work? is it a boolean of an ant? a counter? how long does it last? who can create it?
A counter
It applies only during the immediately following collision resolution phase
I'm hashing out the exact words right now
what happens if a queen doesn't have enough precedence to move trough a worker? Does it just stay in place?
in any way, I don't see any way that makes a queen (+5 its workers supporting it) not able to easily penetrate a 2-wide wall of workers
21:58
Collisions of ants are resolved by precedence. A single unladen ant moving into a cell has zero precedence, the lowest amount one can have. Each of the following actions contributes one precedence to an ant in a collision:

  - Being a queen (supplies one precedence)
  - Holding food (supplies one precedence regardless of food held)
  - Having an allied ant supply support into the cell (supplies one precedence for each such neighbor)
  - Being still in the cell (supplies one precedence)

If the collision involves a queen, then she alone survives the collision regardless of precedence, and g
(not included: a worker can drop a "payload" of support. Ants can "drop" support on themselves too.)
Wording might need some work, but that's the initial idea.
@eaglgenes101 so queens just survive always and don't have to work at all to penetrate a wall?
At her own risk. She will lose food if the ant she stepped on was higher-precedence, and will probably be looted mercilessly by defenders
And the rules should make it difficult for an empty-handed queen to get food and escape with her loot
@eaglgenes101 but she can still easily get in the farm and loot the 70+ food in there (at worst, destroying it so the entry doesn't have it anymore (or, at an even worse case, just disrupt the farms ecosystem & resetting the fungi timers))
Right, I might have to think about it
Scratch that, I probably will have to think about it
just resetting the fungi timers = a loss of 1638 moves in a 1:1 fungi:food farm
22:16
Time to finally pull out placeable barriers?
Ones that the queen alone can't chip through if properly placed?
so here's my playground thing for this. 1-8 - colors, qwertyuio - ants, f - food, ` - fungi, = for stats in the console, s/l - save/load + name + enter. drag - select, c - copy, v+click - paste. Made in Processing
and I'm gone for today
22:39
See ya while I brew up barriers
Or... a similar idea that is similarly impenetrable to queens
23:00
Any other ideas?

« first day    last day (32 days later) »