| Forum Index > SC2 General |
|
palanq
United States761 Posts
perhaps you should make separate columns for cost-effectiveness when dealing full splash damage vs no splash, and cost-effectiveness vs armored vs light (for appropriate units) also the disclaimer about shields isn't really necessary - shields aren't necessarily better than terran/zerg hp because of the way P upgrades work, T repair/medivac heal, and Z regeneration | ||
|
Deckkie
Netherlands1595 Posts
1 supply depot costs 100 minerals and gives 10 food. 1 marine costs one food, so thats a costs of ten minerals. 1 marauserd 2 food = 20 minerals 2 BC costs 6 food is 60 minerals. | ||
|
Tiptup
United States133 Posts
On July 24 2010 23:04 palanq wrote: also the disclaimer about shields isn't really necessary - shields aren't necessarily better than terran/zerg hp because of the way P upgrades work, T repair/medivac heal, and Z regeneration Protoss shields recharge fast outside of combat in SC2. That can be a big deal, particularly in early game. | ||
|
Nihility
Canada72 Posts
On July 24 2010 23:12 Deckkie wrote: I dont know if this will make any difference, but you could think about the food cost. 1 supply depot costs 100 minerals and gives 10 food. 1 marine costs one food, so thats a costs of ten minerals. 1 marauserd 2 food = 20 minerals 2 BC costs 6 food is 60 minerals. Except supply depots actually give 8 food. So marine=12.5 minerals marauder=25 minerals and 1 BC is 6 food=75 minerals. | ||
|
Blacklizard
United States1194 Posts
- buildings usually have 1 armor, meaning subtract a little from your low dmg slow attack rate units. - can the units get into the enemy's base? Cloaked units like the dark templar or banshee for instance. Blink stalkers + observer. Quick hellion run by or reapers. Air units like phoenix. - how much movement do they need to kill workers or buildings (i.e. marines are way more effective than zealots due to range, focus fire, and later on stim + medivac)... demonstrated nicely by game 2 in TLO vs Tester in Day9's king of the beta. - how good the units are at killing static defenses, like photon cannons. Ranged is better due to less static defenses reaching you at any given time. Survivability is a big deal here, like zerglings vs photon cannons or a bunker being repaired. Planetary fortresses get repaired. But if you bring High Templar you can psi storm the workers or an Infester and you can stop them moving to go repair. - how easily the unit can escape when defensive units show up in numbers. Fast movement and cliff walk or hop as reapers and stalkers. Flight like banshees, voidrays, or vikings. Several fit into convenient transports like hydras in overlords or marines in dropships. Can the transport get away? That's easier with more vision, more units to scout, or ranged units to attack from a distance and put the transport less in harms way. Hell, after all of that it's easier to just say harass with: - reapers, hellions, dropship marines/marauders, siege tanks/thors/colossi/vikings on cliffs, vikings, banshees, auto-turrets - stalkers, dark templar, voidrays, phoenix, high templar drops on workers, immortal drops early on, forcefield drops (block ramp or mineral line) - mutas, infestor (infested terrans), burrowed roaches, baneling drops on workers, nydas, hydra drops Kill buildings with: - marines, banshees, battlecruisers, marauders - voidrays, immortals) - hydras, zerglings, roaches, ultras, broodlords, banelings - whatever is convenient  If you have your whole army in their base from a recall, use the whole army. Just hope you can get out alive or nearly finish them here. ---out of time sorry for any typos, incompleteness, etc. | ||
|
randomnine
United Kingdom56 Posts
| ||
|
carwashguy
United States175 Posts
On July 24 2010 23:04 palanq wrote: perhaps you should make separate columns for cost-effectiveness when dealing full splash damage vs no splash, and cost-effectiveness vs armored vs light (for appropriate units) I've done armored/light/etc already, but I'll look into splash. On July 24 2010 23:25 randomnine wrote: Except that you only need to factor in the food cost fully when expanding your army at the start. If you're replacing stuff that's just been destroyed, which happens all the time after the first few units, the food is already there. I'll make a totally separate chart to factor in supply. | ||
|
lololol
5198 Posts
Viking air attack cooldown is 2 seconds, not 1. Hellion vs light is not 18 damage, it's 14 without the preigniter upgrade and 24 with. Thor ground attack cooldown is 1.28 You're rounding too much. For example: the SCV has 54 in the last field, but should have 60, which is more than 10% difference. Marines/marauders with stim/shield are still missing. | ||
|
carwashguy
United States175 Posts
On July 25 2010 00:20 lololol wrote: Speedling refers to zerglings with move speed upgrade. Hellion vs light is not 18 damage, it's 14 without the preigniter upgrade and 24 with. Thor ground attack cooldown is 1.28 Thanks, I confirmed this with Wikia. Apparently Liquipedia is slightly out of date. On July 25 2010 00:20 lololol wrote: Viking air attack cooldown is 2 seconds, not 1. You sure? Liquipedia and Wikia say otherwise. On July 25 2010 00:20 lololol wrote: Marines/marauders with stim/shield are still missing Adding to todo. On July 25 2010 00:20 lololol wrote: You're rounding too much. For example: the SCV has 54 in the last field, but should have 60, which is more than 10% difference. Let me get back to you on this. | ||
|
Ndugu
United States1078 Posts
| ||
|
lololol
5198 Posts
On July 25 2010 01:17 carwashguy wrote: You sure? Liquipedia and Wikia say otherwise. I just checked Liquipedia and it's correct? Cooldown: 1 (Ground) 2 (Air) Last edit was 19 July 2010 | ||
|
carwashguy
United States175 Posts
On July 25 2010 01:38 lololol wrote: I just checked Liquipedia and it's correct? Cooldown: 1 (Ground) 2 (Air) Last edit was 19 July 2010 Haha, I actually made the same mistake at first, too: AM means “Assault Mode” (ground). FM means “Fighter Mode” (air). I know; it's confusing. | ||
|
lololol
5198 Posts
On July 25 2010 01:51 carwashguy wrote: Haha, I actually made the same mistake at first, too: AM means “Assault Mode” (ground). FM means “Fighter Mode” (air). I know; it's confusing. They are switched(by "it's correct" I meant that liquipedia is correct, not the spreadsheet) Assault Mode/Ground/Machine guns should be 1 second(instead of 2). Fighter Mode/Air/Missiles should be 2 seconds(instead of 1). So actually you didn't make a mistake at first :D | ||
|
Makh
Canada143 Posts
My summary is also very similar to yours. Zealots are amazing, Tier1 units are by far the most cost-effective units. Terran units are squishy. Mutalisks are actually not good at all. Of course these numbers are inherently meaningless because they ignore AoE, Armor, Micro, etc. But I still think there is solid value in at least understanding how good units are relative to one another. One thing you might consider doing next is a more complicated analysis of performance based on HP of a unit. 10 Marines v. 1 Thor, for instance. Each Marine dies quickly, gradually lowering their output overtime. The thor, however, keeps shooting 'till it dies. If you visualize this on a graph, The marines' output over time would be a descending staircase. The thor, however, would be one big box. If you analysis 'performance' as a function of the area under the line, your numbers will start to favor high HP units (Thor, BC, Ultralisk) over Tier1 units - regardless of things like Armor and AoE. Nice work! | ||
|
lololol
5198 Posts
On July 25 2010 02:03 Makh wrote: One thing you might consider doing next is a more complicated analysis of performance based on HP of a unit. 10 Marines v. 1 Thor, for instance. Each Marine dies quickly, gradually lowering their output overtime. The thor, however, keeps shooting 'till it dies. If you visualize this on a graph, The marines' output over time would be a descending staircase. The thor, however, would be one big box. If you analysis 'performance' as a function of the area under the line, your numbers will start to favor high HP units (Thor, BC, Ultralisk) over Tier1 units - regardless of things like Armor and AoE. This makes less and less of a difference the bigger the armies are. For example: 1 carrier beats 3 vikings, but 2 carriers lose to 6 vikings, 3 lose to 9, e.t.c. | ||
|
d3_crescentia
United States4054 Posts
Dividing by R^2 should come naturally as you're calculating each in terms of resources. I see the equation as (DPS/R) * (HP/R) - it seems more reasonable to think of it that way, but the problem is that in the end you get some nonsensical units. I'm uncomfortable with is articulating HP as "more time for the unit to actualize DPS," especially as the reasoning for multiplication. In the end you get some sort of Ambiguous HP^2 value (who knows what Damage*HP is supposed to mean) per second per R^2, that is: AHP^2/(S*R^2). I don't think this expression makes any sort of sense logically. Here's my counterproposal: define unit HP in terms of time as well against some other unit X and determine individual cost-effectiveness unit versus unit. Then the equation should turn out to be D/R, where D is the damage it inflicts over its entire lifespan (continuous fire). There's a few more kinks in this system what with the inherent flaws of using DPS as a standard of measurement overall (damage is done in discrete chunks, etc.) but I think the numbers will be more accurate. It should tell us certain things we already know (marauders bad against immortals, etc.) but in the end we can also do some analysis on the numbers as a whole for all of the different unit matchups to determine overall cost-effectiveness. It will take longer, but with the marvel of modern programming not too long. Let me see if I can't get started on this. | ||
|
HubertFelix
France631 Posts
Mutalisks appear rather costly, throughput-wise. With their very high speed and very low range, their are very good for harass et just bad (compared to banshee for example) for big fights. In game experience is the only valuable ressource. There are just too much parameters to do this theory craft. | ||
|
kcdc
United States2311 Posts
| ||
|
Pinkie
United States145 Posts
I just have some suggestions. Since we are dealing with cost effectiveness of units maybe you can use a separate chart to compare the cost effectiveness of upgrades. For example, Stim Pack, Zealot Charge, Storm, and Hunter Seeker Missile(I especially want to see this one) Should be pretty easy since you already have the data for units. | ||
|
carwashguy
United States175 Posts
On July 25 2010 02:03 lololol wrote: They are switched(by "it's correct" I meant that liquipedia is correct, not the spreadsheet) Assault Mode/Ground/Machine guns should be 1 second(instead of 2). Fighter Mode/Air/Missiles should be 2 seconds(instead of 1). So actually you didn't make a mistake at first :D Eh, liquipedia says fm = 1 and am = 2. And I'll work on that rounding. Google's doing some weird stuff there. On July 25 2010 02:08 d3_crescentia wrote: Dividing by R^2 should come naturally as you're calculating each in terms of resources. I see the equation as (DPS/R) * (HP/R) - it seems more reasonable to think of it that way Agreed. Fixed. On July 25 2010 02:08 d3_crescentia wrote: Here's my counterproposal: define unit HP in terms of time as well against some other unit X and determine individual cost-effectiveness unit versus unit. Then the equation should turn out to be D/R, where D is the damage it inflicts over its entire lifespan (continuous fire). There's a few more kinks in this system what with the inherent flaws of using DPS as a standard of measurement overall (damage is done in discrete chunks, etc.) but I think the numbers will be more accurate. It should tell us certain things we already know (marauders bad against immortals, etc.) but in the end we can also do some analysis on the numbers as a whole for all of the different unit matchups to determine overall cost-effectiveness. It will take longer, but with the marvel of modern programming not too long. Let me see if I can't get started on this. Personally, I don't have time for a project of that magnitude, but for you: I say go for it. That definitely sounds like a great way to evaluate units (on a matchup basis). I will have been glad to have prompted you into thinking about it. =] | ||
| ||
|
|
Online Event
Korean StarCraft League
Afreeca Starleague
Rush vs Soulkey
CranKy Ducklings
AI Arena Tournament
BSL Open Qualifier
Sparkling Tuna Cup
OSC
IPSL
BSL Open Qualifier
[ Show More ] |
|
|
EPT
