settle an important bet

...in a theorycraft dispute

what would win

10k marines vs 1k cannons

assume that the cannons are played dimensionally at 100x10 and for whatever reason dont need pylons
the marines come in waves of 250x40, maybe each line is 2 seconds behind the line before. the marines are also 0/0 and not upgraded in any way


Poll: what wins
(Vote): the cannons
(Vote): the marines



SECONDLY - and less importantly...if for some reason 2 ghosts were hanging out in the 1k cannons and each could nuke once, assuming they nuked at different times and then the ghosts disappear, what would win



Poll: what wins
(Vote): cannons and 2 nukes
(Vote): marines

10K MARINES DOZER

cannons and obv cannons w/ 2 nukes

i dont think its very close tbh

10k marines would absolutely rape both scenarios

2 nukes wouldnt do shit... they would kill like what a maximium of 120 marines per nuke

like doomer said

yeah the cannons will repair shields and the cannons inthe previous rows will shoot as well when the marines show up making it extremely hard to kill any cannons, id think if they were stimmed the marines might win though

also u should paint how they would be arranged and how the waves would go

On December 01 2008 19:44 RaiNKhAN wrote: yeah the cannons will repair shields and the cannons inthe previous rows will shoot as well when the marines show up making it extremely hard to kill any cannons, id think if they were stimmed the marines might win though

repair what shields dude... we are talking about thousnds of units, cannons and marines alike are going to die in a fraction of a second, shield regeneration is not important here at all

this thread sounds really intelligent if read by someone oblivious to starcraft imo lolol

no range and no stim and cannons will win

I once played a game on some fastest map possible where I decided to beat some guy who only built cannons using only unupgraded marines, had like literally 70 barracks blinking at some point and I built thousands of units
not even sure I could do it until i got upgrades although I might have, was years ago anyway ;(

like if it were somehow possible for all 10000 marines to attack at the same time and surround all the cannons then they would win but if you have them attack 250 at a time after eachother against 100 lines of 10 cannons where the marines are only targeted by like 20 cannons at a time, cannons are gonna win.

someone make to settle this shit

cannons would annihilate the marines

2 seconds until the next wave? that's a long time.


also the second part i don't understand -.-

Cannons easy.

Baal is noob.

it's simple, it's because the cannons far outrange the marines, that's all there is to it

On December 01 2008 19:46 Baal wrote: Show nested quote + On December 01 2008 19:44 RaiNKhAN wrote: yeah the cannons will repair shields and the cannons inthe previous rows will shoot as well when the marines show up making it extremely hard to kill any cannons, id think if they were stimmed the marines might win though repair what shields dude... we are talking about thousnds of units, cannons and marines alike are going to die in a fraction of a second, shield regeneration is not important here at all

cannons won't die that fast at all

and yes shield regeneration is important. a hundred cannons all regenerating, they will all regenerate some, the lucky ones will regenerate a lot. that hundreds to thousands of hp of regeneration

Cannons. EZ PZ.

the cannons arent gonna regenerate much if all marines come from the same area cuz it'll just be the same cannons getting hit until they die
if marines come spread out across the entire area 250 at a time with 2 seconds inbetween they literally wouldnt kill 1 single cannon and if all 10k marines came spread out across the entire area theyd prolly win and there wouldnt be much regeneration anyway

also note that it will be one line of marines shooting at a time with a few about half a second break inbetween the marines dieing and the next ones coming up.

Do 10 cannons vs 100 marines on a map and test it out.

depends on which side Chuck Norris is : /

that doesnt really test it though. 200 marines vs 20 cannons, cannons lined up 10x2, marines attacking from whatever direction leaves the cannons all firing at once will almost do the trick. the more cannons the better they perform though, as is always the rule with longer range units vs shorter range units.

that being said I think if marines dont attack from the sides, I think 20 cannons would beat 200 0/0 no range no stim rines. but its gonna be a way bigger difference if you have 1000 vs 10000. I honestly think if this was 500 cannons it would be close.

250 each wave, do you seroiusly think a 250 wave will die faster than 2 seconds? the marines will quickly reach its "peak" and then the game will reach its balance, actually this is quite easy to know who will win, it probably wont take over 30 seconds to realize who is gong to win and its easy to test.

if 250 attacked from all sides theyd die almost instantly as theyd get shot at by like 400 cannons.

if they attack from 1 side against 10 at a time then there would be a constant stream

I would assume cannons win, just get someone to make a UMS map and test it!

On December 01 2008 20:09 Tien wrote: Do 10 cannons vs 100 marines on a map and test it out.

that will come down to whether the choke is as wide as the cannon spread

if the marines can't spread around they can't win

even if the marines had the 2 nukes on their side they wouldn't win vs the cannons.

oh gosh, i hope PU didnt put to much money on this one..

errrr i think we are assuming that the 250 rines are spread out 2 per cannon.

If 250 rines came up they would all probably be as wide as like 15 cannons =[ the other 30 on each side wouldn't barely hit them.

If you consider 250 rines spread out in a width of 2 rines per cannon cannons win pretty easily.

This can be solved though with a UMS test for sure though.

stim or no stim?

On December 01 2008 21:17 Red9 wrote: stim or no stim?

no upgrades at all.

300 marines vs 30 cannons

Do you guys want me to play this out?

WHAT IS THE RESULT.

at least 50 marines left, lmk if you want me to send you map, all you do is bring the civ to the beacon.

Why did I just waste 30 minutes of my life making this map lol

cannons should win cause 2nd row of cannons will attack (i beleive) as the marines get in range to attack the first line of cannons

On December 01 2008 21:22 Subaru wrote: 300 marines vs 30 cannons Do you guys want me to play this out?

visually it looks like the marines should rape the cannons here ;0 weird maybe im just used to marines always being stimmed/ranged?

Fine, I'll make a replay lol

http://rapidshare.com/files/169387452/300marines_vs_30_cannons.rep.html

This is the replay

even though this is a good starting point in this simulation. You can not make final conclusion from the results of this.
This is not a linear problem. Simulating this can not be achieved this way. If you happen to get to a steady state from the simulation then its GG marines. But Im almost positive this wont happen. Then you have to basically do this a bunch of times and compare the result and hope that they converge.

I dont know anything about starcraft but I actually like to try this out in the christmas break. But looking at the result of this simulation will be a good start of an investigation.

300 marines.. 68 survived

I wonder what would it would be like if there was more of an open choke.

so subaru
is that screenshot the entire height of the map?

On December 01 2008 22:02 asdf2000 wrote: so subaru is that screenshot the entire height of the map?

in the replay he blocks the bottom with CCs. So in that screenshot the rines only allowed that shown space. I thought in a choke situation the rines would get crushed, apparently not so in an open area i'd assume they would do better.

^^ got it

i still think the cannons would destroy the marines

the more lines of cannons the greater the cannon advantage since that is more movement that the marines have to do, more shifting forward, and a longer period of time with 2-3 rows of attacking cannons

not a big enough sample size imo

Meh, I'll make more lol

lol

if u do like, 100 cannons vs 1000 marines
and compare the results

we may be able to figure out the answer

does it have to be 100 cannons vs 1000 marines?

won't 50 cannons vs 500 marines be fine?

53 cannons vs 530 marines

was worried there might be a difference of like 5 marines or somethin and it won't be enough to tell if it was just variance or what

YER GONNA PAY

I CALLED IT YOU DONKEYS!

10k vs 1k wouldnt change a thing, actually marines would rape even harder given that when the 300marines go to a low # of marines, the firepower is greatly reduce, while in a 10k vs 1k fight, the firepower will remain constant from start to finish.

im willing to bet on this, any takers?

what's the bet on exactly?

to simulate it as big as BW maxed units allow it

On December 01 2008 22:38 Baal wrote: to simulate it as big as BW maxed units allow it

Unhacked clean version of Campaign Editor will allow only 7 players x 200 units each so we could do 1400 marines vs. 140 cannons.

the 10k marines will win imo

although cannons have more range marines are a lot smaller. Enough more marines will be attacking than the cannon range can make up for and thus do higher DPS overall while also starting with higher total HP

also if the marines had enough space to move into the concave arc it would be more over

WTF? I wished I had posted earlier so that I could say 10k rines would win SOOOO EASILY

The thing about cannons is that they have an annoying habit of firing at one target until it dies. That means many of its shots would be totally wasted on dead rines.

Marines on the other hand deal damage instantly. If a marine could even get 1 shot off and die the next rine could do the same without any hesitation. How is this even a thread?

wow so many ppl think the cannons would win lol

the cannons are clumsy, the slow fire rate and the clumsy animation, they will miss an incredible ammount of shots on already dead marines, probably up to 15% of their power or more will be wasted.

rofl this is not close canons kill marines really easy(2 hits if i remember correctly) and their range is far greater than marines also marines will tend to obstruct other marines and such and canons regenerate shields

pd: baal shhhh

and if they come wave then a brake then another wave etc canons will rape them even more because they will regenerate their shields in that time

you guys are noobs. rines ez. the nukes might make a difference so i voted for cannons+nukes tho.

On December 01 2008 23:43 Darki wrote: rofl this is not close canons kill marines really easy(2 hits if i remember correctly) and their range is far greater than marines also marines will tend to obstruct other marines and such and canons regenerate shields pd: baal shhhh and if they come wave then a brake then another wave etc canons will rape them even more because they will regenerate their shields in that time

read the thread u fish....

first scale test = marines won by a land slide

lol, even after big results, there are still ne sayers

Well you're just talking about computer AI right? So with 10K marines, there's a very large chance that the AI will mess up and the marines will start streaming in one at a time.

And, if a human is controlling the marines, he wouldn't be able to get them all to charge at the same time, no matter how fast he is, so they'd be getting owned one control group at a time.

but we agree baals still a newb tho right? : ]

cannons would win if Elky micro'ed them manually

for people who are kinda blindly choosing one side or the other based on intuition...here's the theory.

if you send 10 marines vs 1 cannon it is very clear that the marines will win, and a certain % of the marines will survive on average.
but if you send 20 marines vs 2 cannons, it is once again very clear that the marines will win, and a certain % of the marines will survive on average. BUT, the % of marines that survive might be lower in this scenario.
and as you continue to increase the number of cannons (maintaining the 10:1 marine:cannon ratio), the % of surviving marines can continue to decrease, possibly to the point where the cannons actually win. but the % of surviving marines may also reach a level that stabilizes as you continue to increase the numbers, so that for any arbitrarily large n number of cannons, 10n marines will defeat them.

finally, the spatial arrangement of the cannons and marines becomes an issue as the numbers grow large. when the marines are constricted to attack in waves, this can be advantageous to the cannons. or, when the cannons are restricted to have less than 100% of them firing at all times, this can be advantageous for the marines.

given the results of the trial runs with 300 vs 30, i would estimate that there's a very low probability that 10k marines won't destroy 1k cannons

Exactly, its a standard hyperbole that will quickly flatten out or "stabilize"

20 vs 2 50% survive

40 vs 4 40% survive

100 vs 10 35% survive

500 vs 50 33% survive

10k vs 1k 29 % will survive

100k vs 10k 28.9% will survive

and so on, the line in the % of deaths of marines as the # of unit increases quickly flattens out

lol next thing ur gonna tell me is there is math in poker amiright?

LOL

dude on a perfect scenario cannons win ez first 3 lines is flawed because a 4th cannonw ill shoot imo second if the cannon only need 2 shots to kill a marine and asuming each cannon will shoot the marine infront of it the the marine will only shoot once before it dies, lather rinse repeat it takes more than 10 shots to kill a a cannon so rines ez lose IMO. ( UNDER PERFECT CIRCUMSTANCES I REPEAT PERFECT!)

- Jorge

btw i just talked to jorge about this and we agree that a simulation has to be run with more lines of cannons, atleast 5 rows for it to be more accurate because once u get down to the last line of cannons, but with the other cannons aiding as there will be 2 lines of cannons shooting getting more shots in, i think that it will be more close and the cannons should win

I think the cannons would win because of the range if they come straight up with no flanking.

On December 02 2008 01:11 [vital]Myth wrote: for people who are kinda blindly choosing one side or the other based on intuition...here's the theory. if you send 10 marines vs 1 cannon it is very clear that the marines will win, and a certain % of the marines will survive on average. but if you send 20 marines vs 2 cannons, it is once again very clear that the marines will win, and a certain % of the marines will survive on average. BUT, the % of marines that survive might be lower in this scenario. and as you continue to increase the number of cannons (maintaining the 10:1 marine:cannon ratio), the % of surviving marines can continue to decrease, possibly to the point where the cannons actually win. but the % of surviving marines may also reach a level that stabilizes as you continue to increase the numbers, so that for any arbitrarily large n number of cannons, 10n marines will defeat them. finally, the spatial arrangement of the cannons and marines becomes an issue as the numbers grow large. when the marines are constricted to attack in waves, this can be advantageous to the cannons. or, when the cannons are restricted to have less than 100% of them firing at all times, this can be advantageous for the marines. given the results of the trial runs with 300 vs 30, i would estimate that there's a very low probability that 10k marines won't destroy 1k cannons

All this was said before in post #39 in one sentence ----> "This is not a linear problem"

On December 02 2008 02:15 capaneo wrote: Show nested quote + On December 02 2008 01:11 [vital]Myth wrote: for people who are kinda blindly choosing one side or the other based on intuition...here's the theory. if you send 10 marines vs 1 cannon it is very clear that the marines will win, and a certain % of the marines will survive on average. but if you send 20 marines vs 2 cannons, it is once again very clear that the marines will win, and a certain % of the marines will survive on average. BUT, the % of marines that survive might be lower in this scenario. and as you continue to increase the number of cannons (maintaining the 10:1 marine:cannon ratio), the % of surviving marines can continue to decrease, possibly to the point where the cannons actually win. but the % of surviving marines may also reach a level that stabilizes as you continue to increase the numbers, so that for any arbitrarily large n number of cannons, 10n marines will defeat them. finally, the spatial arrangement of the cannons and marines becomes an issue as the numbers grow large. when the marines are constricted to attack in waves, this can be advantageous to the cannons. or, when the cannons are restricted to have less than 100% of them firing at all times, this can be advantageous for the marines. given the results of the trial runs with 300 vs 30, i would estimate that there's a very low probability that 10k marines won't destroy 1k cannons All this was said before in post #39 in one sentence ----> "This is not a linear problem"

yes but it seems people are not getting that the curve flattens quite quickly

On December 02 2008 01:11 [vital]Myth wrote: for people who are kinda blindly choosing one side or the other based on intuition...here's the theory. if you send 10 marines vs 1 cannon it is very clear that the marines will win, and a certain % of the marines will survive on average. but if you send 20 marines vs 2 cannons, it is once again very clear that the marines will win, and a certain % of the marines will survive on average. BUT, the % of marines that survive might be lower in this scenario. and as you continue to increase the number of cannons (maintaining the 10:1 marine:cannon ratio), the % of surviving marines can continue to decrease, possibly to the point where the cannons actually win. but the % of surviving marines may also reach a level that stabilizes as you continue to increase the numbers, so that for any arbitrarily large n number of cannons, 10n marines will defeat them. finally, the spatial arrangement of the cannons and marines becomes an issue as the numbers grow large. when the marines are constricted to attack in waves, this can be advantageous to the cannons. or, when the cannons are restricted to have less than 100% of them firing at all times, this can be advantageous for the marines. given the results of the trial runs with 300 vs 30, i would estimate that there's a very low probability that 10k marines won't destroy 1k cannons

lol i love you corwin

it seems OP is biased because the second poll has no meaning if the first poll is actually debatable

baal your curve is entirely wrong
300 vs 30 and only 22.6% survived, 500 vs 50 would have a smaller percentage than that
I have to admit that the result surprised me some and that my human experiments with this just dont matter at all cause I cant attack with 250 marines at the same time
but there is a huuge difference between 10000 vs 1000 and 300 vs 30. cannons have 3 range longer. in the 300 vs 30 scenario, once the marines have killed 10 cannons, only 20 are firing. once theyve killed 20, only 10 are firing. whereas normally, between 20 and 30 will fire at any given moment. i mean basically, if you made a scenario with 300 vs 40 however where the 10 cannons behind were invincible and the marines only had to kill the 30 in the front, it would kind of illustrate what I mean.. marines would also follow up behind but its just a tactical fact in starcraft that the bigger amounts of units you have, the larger the advantage of the longer ranged units becomes. rines have 4, cannons 7. 20 goons will absolutely slaughter 5 tanks. 200 tanks vs 800 goons and I actually think tanks will win ;p
but yea its very possible that I was wrong from the beginning, when I see 300 rines lined up like that makes them look stronger than "300 marines" ;(

those numbers were totally made up and i never intended them to be remotely accurate, only to illustrate what i meant with an hyperbola


about tanks vs goons i dunno, probably the tanks win cuz goon AI fucking blows and the Tank attack has no animation delay so there are no missed shots and there will be many missed goon shots

mmmmmm this could develop in really nice prop bet thread.

Cant someone just use some hacked map editor or something to make a map closer to OP post (i can't btw)

decent sized map with more cannons in rows, and then use beacons or sth for getting marines in since then you don't have to have them all standing on screen and that will get somewhat around the # of units limitations?

i believe it takes four shots to kill a marine from a cannon because of the armour class, cannons only doing 1/2 damage on "light targets" (could be wrong though)

you are wrong, you would be correct if we played starcraft 1.00 or 1.01, so like, 10 years ago
cannons deal normal damage

is it goons that deal 1/2 damage?

yes

Cannon defense!! Just stack the cannons for an easy win. I still voted for cannons btw even without stack. Because all the marines will die before even getting off a shot.

In the Replay provided the dimensions are 10X3 cannons and 30x10 marines. Giving the marines 3 times more length coverage. In OP its laid out as 100x10 cannons and 250x40 marines giving then only a 2.5:1 ratio advantage in area. therefore the marines going to be more sorrounded by the cannons.

You're right about one thing eri, 300 marines looks insane when you actually see them on a map haha.

that's what i'm thinking about when we talk 10k marines.

Marines easily for one reason that nobody stated

In starcraft, there exist what we call "firelag"
anyone ever played "ultimate canon defense" or some shit like that. when too many canons fire at the same time(over 200 approx), some of them wont shoot because in starcraft there is a maximun of things that can shoot at the same time

and for some reasons, units doesnt all have the same firelag. most recents units shoot first but it also has something to do with the unit type (for example do a test, build 100 scouts, and then, a second player build 98 scouts. if they fight, the 98 will win ez due to firelag cuz more recent unit fire more when there is firelag)
some units, such as dragoons, canons, battlecruisers, etc have very high firelag
but some units like marines have nearly no firelag

in the above scenario, 100 scouts vs 100 scouts, if if you had to choose between 20 marines or 15 goons to help out your scouts, you probably should pick the marines cuz goons will nearly never fire at all.

Btw in the test, subaru probably didnt experience this, because 30 canons is not enough to produce firelag

there would need to be at least 200 canons firing at the same time


The nuke wont change anything, it could kill 200 marines being VERY generous and 200/10000 that doesnt change anything

The only way the canons could win is if the marine are sent very slowly so that they never experience firelag (aka there is never 200 canons firing at the same time because there is never that many marines to be shot) or if canons are all stacked together lol

there would never be 200 cannons firing at the same time because theyre getting attacked where they are thinnest.

do u rly think theyre never gonna be more than 200 shooting?

if for some reason the terrain was made in a way the marines cannot attack the sides, there probably wouldnt be firelag, but pooruser never stated this in his original post?

I've tryed to make a map close enought to the OP statement and i got this results.

1.- I hate pooruser.
2.- Make the map its harder that it seems.
3.- It may seems that in a perfect scenario as said by capaneo and baal 10k marines could easily own 1000 cannons but its not like that in practice, in my test high volume of marines become EXTREMELY dumb, the AI mess everything up and get killed like ducks.
4.- Its imposible to get marines to attack in a straight line, they confuse, stuck and act like goons on stairs.
5.- 3 Rows of cannons are always shotting killing marines very fast as they aproach.
6.- More than 400 or 500 marines at a single time make the map become bugged, u can see the animation of units shotting but they're not making any damage.
7.- In practice 958 marines got killed and only 25 cannons were destroyed mainly because 100 cannons take a very wide portion of the map while 240 marines doesnt.
8.- At least 40% of the marines die without making a single shoot.
8.- I'll try to make a better map later cause yesterday i got tilted and erase all the triggers >_<.

On December 02 2008 11:11 Funktion wrote:

nice one ahahha. cannons are defective tho TT

On December 02 2008 02:15 capaneo wrote: Show nested quote + On December 02 2008 01:11 [vital]Myth wrote: for people who are kinda blindly choosing one side or the other based on intuition...here's the theory. if you send 10 marines vs 1 cannon it is very clear that the marines will win, and a certain % of the marines will survive on average. but if you send 20 marines vs 2 cannons, it is once again very clear that the marines will win, and a certain % of the marines will survive on average. BUT, the % of marines that survive might be lower in this scenario. and as you continue to increase the number of cannons (maintaining the 10:1 marine:cannon ratio), the % of surviving marines can continue to decrease, possibly to the point where the cannons actually win. but the % of surviving marines may also reach a level that stabilizes as you continue to increase the numbers, so that for any arbitrarily large n number of cannons, 10n marines will defeat them. finally, the spatial arrangement of the cannons and marines becomes an issue as the numbers grow large. when the marines are constricted to attack in waves, this can be advantageous to the cannons. or, when the cannons are restricted to have less than 100% of them firing at all times, this can be advantageous for the marines. given the results of the trial runs with 300 vs 30, i would estimate that there's a very low probability that 10k marines won't destroy 1k cannons All this was said before in post #39 in one sentence ----> "This is not a linear problem"

yeah i know but lol if you think people know what that means

Marines ofc

but with nukes cannons may win

and stacked cannons will have a SHITLOADS of missed damage, like the first marine will take damage from 500 cannons nearly

and the marines come from everyside and make a circle around the 1k stacked cannons will be hugely easy for marines imo

i've got a friend that used to mapmake i will ask him what he can do ^^

I doubt the sprite limit would allow you guys to actually really test 1k vs 10k.

(who would call that "firelag" though, what a noobish term )

It would also depend on the position: Like if the cannons form an arc and the marines would move into the "belly of the beast" by default, cannons' job would get a tad easier. I'd bet on them anyway, as an active SC player.



Figures...after playing SC so much, my first post here is in a topic about SC.

Greetings fellow poker players at LP.net...I've just been a huge lurker for years as I wasn't playing poker myself, I just started recently.

P.s.: Props to anyone that recognizes me, good to meet you here too off the SC field.

Oh, and good run at the tables, or whatever should I say

Hello, welcome to the forums !

I was going to make a long post about how I thought the marines would rape, but this is actually close to call, imo.

There's one thing that I don't think has been considered yet:
1 cannon is like 4-5 rines side by side. Assuming the rines are spawned and sent right into the middle of the cannon formation, it would look a lot like this:

oooo


mm
/

As the front cannons get destroyed, a concave will form and likely continue to exist for the rest of the battle. That would be such a weird "Cannon/Rine Ratio - Rine Survival" graph.

This is like chess, a waste of human intellect.

you can definitely try this out by using team melee mode and then save game when you have 250 marines and a line of cannons, maybe 40tall 5-10 deep and get an average result of how many cannons die per wave of 250 rines.

empirical method imo.

On December 02 2008 22:56 Tien wrote: This is like chess, a waste of human intellect.

lol