EPTThe only drawback to this, is that it only measure the quantitative data. There is no way you can include how much you may throw a player off balance into a mathematical equation.
Economic Damage Calculator - Page 2
| Forum Index > SC2 General |
|
Najda
United States3765 Posts
The only drawback to this, is that it only measure the quantitative data. There is no way you can include how much you may throw a player off balance into a mathematical equation. | ||
|
Cascade
Australia5405 Posts
On August 30 2012 12:17 Bored wrote: I think that in the long run I can probably do that, it'd be nice to see how delayed your push is if you decide to go for an attack and how much it'll set him back. I'll give it a go, although I feel like this'll take a bit longer than a few hours that the current one I did. Thanks for the idea! yeah, it's a quite big project I think, I'd be impressed if you manage to go through with it. You could even look at a specific game, as use the builds they used in that game. Then compare to what would have happened if he had not built the earlier factory and the starport, how much earlier would his first push after that be, and how much earlier would the zerg have the units he used to defend that push. That would be really cool, if you could say that MVP needed to kill at least 9.3 drones, or take at least 94 drone-seconds of mining, or a corresponding combination, to make up for his investement. Ofc then you will have issues like, without the hellions his creep would be much further, no safety from ling all-ins etc etc, but that will have to be another story I believe, as it is very hard to transform into economy... | ||
|
Bored
Canada11 Posts
On August 30 2012 12:25 Cascade wrote: yeah, it's a quite big project I think, I'd be impressed if you manage to go through with it. You could even look at a specific game, as use the builds they used in that game. Then compare to what would have happened if he had not built the earlier factory and the starport, how much earlier would his first push after that be, and how much earlier would the zerg have the units he used to defend that push. That would be really cool, if you could say that MVP needed to kill at least 9.3 drones, or take at least 94 drone-seconds of mining, or a corresponding combination, to make up for his investement. Ofc then you will have issues like, without the hellions his creep would be much further, no safety from ling all-ins etc etc, but that will have to be another story I believe, as it is very hard to transform into economy... Well I'll probably grab a friend of mine so I don't have to do it on my own haha. It'd be a fun project actually. | ||
|
Xapti
Canada2473 Posts
On August 30 2012 10:59 Maxd11 wrote: How do you modify the equation for each race to factor in chrono boost/the facts that zerg can (theoretically) replenish all their workers at the same time using larvae? This is really important — namely for zerg. Each drone kill that an opposing player gets on a zerg isn't really worth much more than 63 minerals worth of killing anything else (spine crawlers, zerglings, etc.), while tat number would be considerably higher for other races since they can't produce workers as fast (produce workers instead of army) | ||
|
DangerAl
Australia88 Posts
- the 50 minerals spent on the worker is a sunk cost. So whether the worker gets killed or not the player has already spent 50 minerals on it, and it should no longer be taken into consideration. It is true that the player will have to spend 50 extra minerals on an additional worker, but because most of the time the player will be constantly producing workers anyway, these 50 additional minerals do not actually come into effect until the player would have had all the SCVs they need for the game (eg. if you're terran and you stop at 65 SCVs, you will have to build a 66th SCV to replace the one you lost). That means if you lose a worker at 6 minutes you won't feel the cost of replacing it until 12-13 minutes (depending on your build), by which time minerals aren't worth nearly as much due to the time value of minerals. - What you do feel is the lost income, and think this is a far better indicator of economic damage. A worker mines approximately 40 minerals per minute (less if the base is fully saturated). This means that if you lose a worker at time 0, you have 40 minerals less than you would have at minute 1. At 2 minutes you would have 80 minerals less, and so on. Then you should be adjusting for the time value of money. For example, losing 40 minerals in 1 minute might only be as bad as losing 30 minerals right now, and losing 40 minerals in 2 minutes time might only be as bad as losing 23 minerals right now and so on. If you add all these discounted future mineral flows, you get the true present cost of losing the SCV. There are problems with using this method of course, the most obvious being that it is practically impossible to work out the time value of minerals. Also, once you get to your full worker count, the lost SCV has essentially been replaced, so you should only discount up to the time that full workers are obtained. tl;dr: workers are worth way more than 50 minerals. They are a stupidly good investment and that is why players get so many of them. I hope that makes sense. What you have done is quite good, certainly better than nothing, but I feel it doesn't paint the truest picture of the economic damage done. | ||
|
LloydRays
United States306 Posts
Aside, I do think there should be some kind of derivative which we could use to measure differences at resource collection rates at 2 points in the game, say before and after harass. | ||
|
hellokitty[hk]
United States1309 Posts
Me too and then again after I read this. | ||
|
teamamerica
United States958 Posts
On the premise that you only look at the number of bases your opponent has, I can't take your formula seriously at all. I mean, killing 8 drones with gas first hellion rush vs 3 hatch before pool Zerg has a drastically different effect. In the same vein of Cascade arguing that minerals at some point are worth more than minerals at another point, and the different effects different builds have, delaying your expansion also costs you money. When I play 1rax expo before depot vs reactor hellion, I think my orbital is done about two minutes later. So I'm putting myself 2x worker production for two minutes down in workers. Otherwise killing 10 scvs with 4gate vs 1rax expo would make it seem sick. I understand that this strays toward build specific differences but would you take into account how long you delay your expo (minute wise) also. Anyway, different builds make this formula pretty neat but I think too shallow to use to actually develop strategies (although of course, I just steal pro builds, so maybe I'm wrong). I mean TvZ revolves around marine tank pushes to shut down the 7-8th gas of Zerg while you grab your 4th. If you do a ton of hellion harrass but have no follow to stop the Zerg from redroning, or let the Zerg reach infestor/corrupter/broodlord because you delayed you attack too much because of hellions, the huge number this formula gave me is worth nothing. And hellion harrass does delay any attack I do, the only question being how much. I mean I saw Guhmiho go CC first into 3rd cc before has into tanks and did marine pressure on the 3rd with medivacs, and still was able to hit a 2/2 1/0 180 supply tank push around 12:30. Can any formula quantify that by hitting around then, there's almost no way at all you'll have to face hive tech (I've seen 11 min hives but I don't think they'd finish in time for ultra den to be built and ultras to spawn). And by facing a lair tech army, you're going to be more cost efficient probably, just by the nature of Zerg lair tech units? If it seems like I'm asking an impossible task, it's because in my opinion it is impossible (without significant computing power to get results in decent speed or something else like a genetic algorithm), or at least very difficult. | ||
|
Flamingo777
United States1190 Posts
On a more serious note, how do the bases for zerg differ from that of the other two '1 worker building at a time' bases such as Terran's Command Center and the nexus of Protoss? | ||
|
Rannasha
Netherlands2398 Posts
In reality, if you lose 1 worker, you will be one worker behind on your build until the point where you reach your intended number of workers. So the number of 'cycles' lost is much larger than indicated. However, as a counterbalancing effect: If you kill workers while the opponent has more than 16 workers per mineral line, the lost mining will be less than if there are 16 or less workers per mineral line. | ||
|
Nekovivie
United Kingdom2599 Posts
The Equation: 50x+(0.5x)(x+1)(11.33333)((((-1)^(b))-1)/2))^2)+(1/8)(7+(-1)^x+8x+2x^2)(11.33333)(((-1)^(b)+1)/2) where x=WORKERS and b=BASES Mother of god. | ||
|
strongandbig
United States4858 Posts
That said, I think this could be really useful in two contexts and I really commend OP on the work. (1) in combination with an evolutionary algorithm to create build orders (remember the seven roach rush?), this could be very useful. You could give the evolutionary algorithm a parameter to maximize other than "I want more roaches". AFAIK so far most of the successful builds from evolutionary algorithms have been allins, because it's so hard to parameterize anything else to optimize. (2) more importantly, for casters. This would be a really interesting thing for casters to show during games where some sort of worker harass takes place. It would give a sort of running update on how strong the economic harass was from either player. | ||
|
Rabiator
Germany3948 Posts
On August 30 2012 10:59 Maxd11 wrote: How do you modify the equation for each race to factor in chrono boost/the facts that zerg can (theoretically) replenish all their workers at the same time using larvae? For Zerg there could be an additional factor in it because you force them to spend larvae on Drones and not units, thus limiting their offensive potential. Killing enough Queens to have less than one per Hatchery would also deal a lot of economic damage. On the other hand I dont think it is a good thing to "figure out SC2 mathematically", because that makes it less interesting as a game if people start looking at pure numbers only. Some benefits cant be counted like pulling the army of your opponent out of position. If a math solution is available people will start asking for that ... just like the idiots who ask in a brand new MMO "which is the best [=most powerful] class?" and totally ignore the fact that there are situational decisions which tip the scale to one side or the other. Relying on pure math is never a good thing! | ||
Teoita
Italy12246 Posts
| ||
|
brachester
Australia1786 Posts
it doesn't look that complicated if you work a lot with programs such as maples or sites like wolframalpha | ||
|
Broodwurst
Germany1586 Posts
| ||
|
DontNerfInfestors
Spain280 Posts
| ||
|
Rannasha
Netherlands2398 Posts
On August 30 2012 10:59 Maxd11 wrote: How do you modify the equation for each race to factor in chrono boost/the facts that zerg can (theoretically) replenish all their workers at the same time using larvae? For Chrono, just divide the 11.33333 values (2 occurrences) in the equation by 1.5 (assuming constant Chrono on probes). Zerg is more complicated, since Larvae are a resource and their value depends on the build the Z is doing and the current game-state. | ||
| ||
