|
On April 28 2010 07:28 Kantutan wrote:
If you start cloak the same time as you start making a banshee, it's finished what... 30 seconds after that banshee is made? Then you force your opponent to have to use scans over mules then build a raven, etc, etc.
Banshee build time=60 seconds
Cloak research=110 seconds
Viking build time=42 seconds (and Reactor-able, and cheaper on gas.)
Looking at this, one might not ask, why not just open with Banshees and Vikings? 1 techlab starport and 1 reactor port.
I think this might be safe just in the context of this particular decision matrix, but IMO you may not get enough Banshees out to beat any kind of ground-based Marine/Tank/(Turret) push.
|
Wow. As an econ senior, I can ALREADY tell you you've made 3 pretty ridiculous assumptions.
1. No other units exist (there obviously are)
2. It's a game of perfect information (it isn't)
3. You randomly assigned values to the outcomes of the game. (damages will be different every match)
Take an actual game theory course please, then come back.
|
lol I see ECO TAs teaching 1st year ECO student here.
|
Nivra:
I appreciate your criticisms. For [B1] I would argue that non-turrets/vikings really don't pose that much of a threat to Banshees at the juncture of decision making we are discussing. For [V1], vikings are quite good at harassing mineral lines and being ground units, so they can never be worse than what they come out at, in my opinion.
I think the proper way to do this would in fact be to run the expanded simulation of Vikings vs Banshees vs Banshees w/ cloak vs 1 raven into vikings and see what we get. Right now, there are just few ground units that seem to offer anything relevant to the air battles, that I think this simplification will suffice.
ZapRoffo:
While SC is zero sum per se, game theory payoffs don't operate in the same way, particularly because the game doesn't end once you choose Viking and he chooses Banshees. By assigning weighted points for a possible outcome, you can predict what role it will play in a game overall. If we can decide upon the strength of a unit in different scenarios, SC can be boiled down to calculating all the relevant values up and comparing them.
Cartoonx:
There is some truth in that, but not for our purposes. We are trying to make informed pre-emptive decisions regarding basic strategies a matchup revolves around. The purpose is to say, all other things being equal, how should I approach the use of air units? What takes skills is understanding that a) things will rarely be equal and b) properly accounting for that to modify your decision making paradigms. Thus, we don't need to work through all the variants, just the key principles of what is critical to each decision, and consider them in real time as needed. As I mentioned explicitly in my post, these three things are NOT of equal value, which is where actual playing comes in.
EDIT:
Sadistx:
Sorry chump, the use of game theory in economics is more recent than its use in most other fields. In fact, game theory offers some of the weakest success in properly predictive economic models. The most success occurs under biological scenarios. Speaking from three years of active research, I can safely say, simplifying assumptions are the manner in which you approach a problem. The way in which to improve the model is to assume the simplest case and then build up as data becomes available. Stop trolling.
|
Sleight,
3 years of research means you are in graduate school at best, or just starting. The other poster is a senior, which means he's not even in graduate school. Both of you need to drop the personal credentials war and just discuss the topic at hand imo 
|
I realize that you're trying to break this down in a formulaic way, but I feel that you're oversimplifying something that is, in general, a bad line of strategy.
Pardon my bad manners, but KawaiiRice played poorly and mindlessly, there's just no other way to look at it. His early Starport was scouted, giving his opponent ample time to counter it, yet he commited to it. Twice. A Banshee rush is like a Wraith rush in Starcraft 1. Against a competent opponent, it won't work without the element of surprise. In fact, it's a colossal loss to the player doing it if it's at all countered or stopped. The issue is as simple as that. KawaiiRice showed no adaptability beyond a static strategy he had memorized, and embarassingly, he commited to it in his second game as well, with even worse results than the first.
Secondly, Terran players have it wrong in thinking that they rush or be rushed by air. There's a third option: get an engineering bay and play your own game, not an entirely reactive one to something that's easily countered by something that will eventually be in your build anyhow. One or two scans, which you would use on Banshees, more than cover the necessary Turrets and Engineering Bay that you would get otherwise -- doesn't this just more logical, solid sense? I just can't wrap my head around people's lack of solid builds or drive to formulate solid builds. Terran players will be playing like this once they stop getting so wrapped up in this one-base all-in mentality. It's the equivalent of what Day9 used to talk about when he began his career; he used to do lurker drops on his enemies, and one day, it stopped working when, gasp, he ran into a few turrets, and he had no strategy, he didn't know what to do. The same holds true for most "top" Terran players today in regards to air tactics, as is evidenced by KawaiiRice, unfortunately.
I apologize if I'm missing the point of the more precise analysis you were trying to make with this thread, but I can't help but feel that people are missing the bigger picture here, and that this type of analysis simply isn't necessary.
|
terran air mechanics is real simple. if enemy goes banshee, you have no other choices but to go air too, i.e. vikings. mass thor wont do well against mass banshee in late game (try this if you dont agree)
the key to dominate TvT is to make the right amount of vikings. if you have too many vikings and enemy switches to pure ground, you are done. if you mass ground units i.e. tanks+marauders plus a few but insufficient vikings and enemy goes mass air (vikings+banshees), you are in danger, too. banshee kill ground units as fast as vikings kill banshees. resourced used in both build? equal.
best way is to keep a reactor-port just for quick transition and check out his unit combination every now and then. cos in TvT surprise is fatal.
|
On April 28 2010 05:58 avilo wrote:
sometimes there is no intriguing or massive science behind why players do things.
... and having the "correct counter unit" doesnt automatically make you win. Playing skill (scouting the Banshee AND controlling your Vikings) is required too to be effective.
|
|
|
I am a COLLEGE STUDENT, hold the applause, and I'm here to say that game theory has nothing to do with anything, especially games.
Genetic algorithms will solve everything anyway. EVERYTHING.
|
On April 28 2010 12:42 RatherGood wrote:
I realize that you're trying to break this down in a formulaic way, but I feel that you're oversimplifying something that is, in general, a bad line of strategy.
Pardon my bad manners, but KawaiiRice played poorly and mindlessly, there's just no other way to look at it. His early Starport was scouted, giving his opponent ample time to counter it, yet he commited to it. Twice. A Banshee rush is like a Wraith rush in Starcraft 1. Against a competent opponent, it won't work without the element of surprise. In fact, it's a colossal loss to the player doing it if it's at all countered or stopped. The issue is as simple as that. KawaiiRice showed no adaptability beyond a static strategy he had memorized, and embarassingly, he commited to it in his second game as well, with even worse results than the first.
Secondly, Terran players have it wrong in thinking that they rush or be rushed by air. There's a third option: get an engineering bay and play your own game, not an entirely reactive one to something that's easily countered by something that will eventually be in your build anyhow. One or two scans, which you would use on Banshees, more than cover the necessary Turrets and Engineering Bay that you would get otherwise -- doesn't this just more logical, solid sense? I just can't wrap my head around people's lack of solid builds or drive to formulate solid builds. Terran players will be playing like this once they stop getting so wrapped up in this one-base all-in mentality. It's the equivalent of what Day9 used to talk about when he began his career; he used to do lurker drops on his enemies, and one day, it stopped working when, gasp, he ran into a few turrets, and he had no strategy, he didn't know what to do. The same holds true for most "top" Terran players today in regards to air tactics, as is evidenced by KawaiiRice, unfortunately.
I apologize if I'm missing the point of the more precise analysis you were trying to make with this thread, but I can't help but feel that people are missing the bigger picture here, and that this type of analysis simply isn't necessary.
If your opponent goes banshee vikings, you're stuck in your base unless you want landed vikings to demolish your towers and let the banshees in.
|
I'M THE PRESIDENT OF THE UNITED STATES AND I HAVE 100 NOBEL PEACE PRIZES
|
United States32976 Posts
I've never seen a useful application of game theory on a video game forum ;p
Here's some advice from a former econ major: If you learn econ well (in undergrad), the biggest thing you'll learn is how LITTLE you know, and how silly it is to try and apply your rudimentary theoretical knowledge to real life things  . But it will get you a job easier than art history ^_^
|
On April 28 2010 15:19 Waxangel wrote:I've never seen a useful application of game theory on a video game forum ;p Here's some advice from a former econ major: If you learn econ well (in undergrad), the biggest thing you'll learn is how LITTLE you know, and how silly it is to try and apply your rudimentary theoretical knowledge to real life things . But it will get you a job easier than art history ^_^
This is the truth, basically, although it improves your thinking skills and gets you thinking by default in terms of opportunity costs which is a pretty valuable concept that lots of non-econ people just totally botch.
Although I did take one seminar course where we just analyzed the shit out of data about recessions, now I know quite a lot of patterns behind that.
|
On April 28 2010 15:13 Odge wrote:Show nested quote +On April 28 2010 12:42 RatherGood wrote:
I realize that you're trying to break this down in a formulaic way, but I feel that you're oversimplifying something that is, in general, a bad line of strategy.
Pardon my bad manners, but KawaiiRice played poorly and mindlessly, there's just no other way to look at it. His early Starport was scouted, giving his opponent ample time to counter it, yet he commited to it. Twice. A Banshee rush is like a Wraith rush in Starcraft 1. Against a competent opponent, it won't work without the element of surprise. In fact, it's a colossal loss to the player doing it if it's at all countered or stopped. The issue is as simple as that. KawaiiRice showed no adaptability beyond a static strategy he had memorized, and embarassingly, he commited to it in his second game as well, with even worse results than the first.
Secondly, Terran players have it wrong in thinking that they rush or be rushed by air. There's a third option: get an engineering bay and play your own game, not an entirely reactive one to something that's easily countered by something that will eventually be in your build anyhow. One or two scans, which you would use on Banshees, more than cover the necessary Turrets and Engineering Bay that you would get otherwise -- doesn't this just more logical, solid sense? I just can't wrap my head around people's lack of solid builds or drive to formulate solid builds. Terran players will be playing like this once they stop getting so wrapped up in this one-base all-in mentality. It's the equivalent of what Day9 used to talk about when he began his career; he used to do lurker drops on his enemies, and one day, it stopped working when, gasp, he ran into a few turrets, and he had no strategy, he didn't know what to do. The same holds true for most "top" Terran players today in regards to air tactics, as is evidenced by KawaiiRice, unfortunately.
I apologize if I'm missing the point of the more precise analysis you were trying to make with this thread, but I can't help but feel that people are missing the bigger picture here, and that this type of analysis simply isn't necessary. If your opponent goes banshee vikings, you're stuck in your base unless you want landed vikings to demolish your towers and let the banshees in.
to counter: make no more than enough vikings to handle banshee and use the remaining resources to build a ground army to finish off the opponent. because your opponent needs resources on both mass vikings and mass banshees, and you only need a little more vikings without banshees. you have the econ advantage for ground units turret, and expansions. Vikings are purely for map control, whereas ground units are the key to bring a gg to him.
|
On April 28 2010 12:42 RatherGood wrote:
Secondly, Terran players have it wrong in thinking that they rush or be rushed by air. There's a third option: get an engineering bay and play your own game, not an entirely reactive one to something that's easily countered by something that will eventually be in your build anyhow. One or two scans, which you would use on Banshees, more than cover the necessary Turrets and Engineering Bay that you would get otherwise -- doesn't this just more logical, solid sense?
Your basic idea is right, but Banshees are really strong air-to-ground and enough of them will easily kill turrets. (Banshee+Raven > Turrets, unless you have lots and lots.)
A few turrets can help as as do-not-die-right-now cludge, but I think in a longer game Vikings are the best bet.
|
On April 28 2010 16:23 ComTrav wrote:Show nested quote +On April 28 2010 12:42 RatherGood wrote:
Secondly, Terran players have it wrong in thinking that they rush or be rushed by air. There's a third option: get an engineering bay and play your own game, not an entirely reactive one to something that's easily countered by something that will eventually be in your build anyhow. One or two scans, which you would use on Banshees, more than cover the necessary Turrets and Engineering Bay that you would get otherwise -- doesn't this just more logical, solid sense? Your basic idea is right, but Banshees are really strong air-to-ground and enough of them will easily kill turrets. (Banshee+Raven > Turrets, unless you have lots and lots.) A few turrets can help as as do-not-die-right-now cludge, but I think in a longer game Vikings are the best bet.
turrets are a terrible counter to fast banshees. Their range is so short, the banshees can always find a way to sneak in and do damage.
|
On April 28 2010 16:46 Luddite wrote:Show nested quote +On April 28 2010 16:23 ComTrav wrote:On April 28 2010 12:42 RatherGood wrote:
Secondly, Terran players have it wrong in thinking that they rush or be rushed by air. There's a third option: get an engineering bay and play your own game, not an entirely reactive one to something that's easily countered by something that will eventually be in your build anyhow. One or two scans, which you would use on Banshees, more than cover the necessary Turrets and Engineering Bay that you would get otherwise -- doesn't this just more logical, solid sense? Your basic idea is right, but Banshees are really strong air-to-ground and enough of them will easily kill turrets. (Banshee+Raven > Turrets, unless you have lots and lots.) A few turrets can help as as do-not-die-right-now cludge, but I think in a longer game Vikings are the best bet. turrets are a terrible counter to fast banshees. Their range is so short, the banshees can always find a way to sneak in and do damage.
One in each mineral line is definitely a good investment, they will hold banshees off until there is 3-4 and give you time to react and pull the neccesary units to actually defend properly. Not to mention detection before ravens and without wasting scan.
|
On April 28 2010 19:02 sob3k wrote:Show nested quote +On April 28 2010 16:46 Luddite wrote:On April 28 2010 16:23 ComTrav wrote:On April 28 2010 12:42 RatherGood wrote:
Secondly, Terran players have it wrong in thinking that they rush or be rushed by air. There's a third option: get an engineering bay and play your own game, not an entirely reactive one to something that's easily countered by something that will eventually be in your build anyhow. One or two scans, which you would use on Banshees, more than cover the necessary Turrets and Engineering Bay that you would get otherwise -- doesn't this just more logical, solid sense? Your basic idea is right, but Banshees are really strong air-to-ground and enough of them will easily kill turrets. (Banshee+Raven > Turrets, unless you have lots and lots.) A few turrets can help as as do-not-die-right-now cludge, but I think in a longer game Vikings are the best bet. turrets are a terrible counter to fast banshees. Their range is so short, the banshees can always find a way to sneak in and do damage. One in each mineral line is definitely a good investment, they will hold banshees off until there is 3-4 and give you time to react and pull the neccesary units to actually defend properly. Not to mention detection before ravens and without wasting scan.
one in each... so you fast expanded? And also built an ebay and two turrets? Then you're dead. You won't have enough marines, you won't have vikings yet, and all your unit producing buildings will be undefended. With 2 fast banshees he also has the option of just going straight for the cc.
|
On April 28 2010 19:07 Luddite wrote:Show nested quote +On April 28 2010 19:02 sob3k wrote:On April 28 2010 16:46 Luddite wrote:On April 28 2010 16:23 ComTrav wrote:On April 28 2010 12:42 RatherGood wrote:
Secondly, Terran players have it wrong in thinking that they rush or be rushed by air. There's a third option: get an engineering bay and play your own game, not an entirely reactive one to something that's easily countered by something that will eventually be in your build anyhow. One or two scans, which you would use on Banshees, more than cover the necessary Turrets and Engineering Bay that you would get otherwise -- doesn't this just more logical, solid sense? Your basic idea is right, but Banshees are really strong air-to-ground and enough of them will easily kill turrets. (Banshee+Raven > Turrets, unless you have lots and lots.) A few turrets can help as as do-not-die-right-now cludge, but I think in a longer game Vikings are the best bet. turrets are a terrible counter to fast banshees. Their range is so short, the banshees can always find a way to sneak in and do damage. One in each mineral line is definitely a good investment, they will hold banshees off until there is 3-4 and give you time to react and pull the neccesary units to actually defend properly. Not to mention detection before ravens and without wasting scan. one in each... so you fast expanded? And also built an ebay and two turrets? Then you're dead. You won't have enough marines, you won't have vikings yet, and all your unit producing buildings will be undefended. With 2 fast banshees he also has the option of just going straight for the cc.
yes you're right! only counter to banshee - viking. turrets only buy you sometime.
|
|
|
|
|
|
EPT
