With zerg its actually possible to calculate and estimate i think. The calculation depends on how many spare larva you have stockpiled and how many empty mineral patches are left..
Lets say we have infinity larva and infinity mineral patches but limited resources. We assume our resources are always at zero, because we always spend all of it. Now if a drone gets killed, that means 50 minerals lost for the drone obviously, but lets put that aside for a minute. The mining time that this drone lost, is equivalent to 40 minerals per minute. 40 Minerals is 80% of another possible drone, as a drone costs 50. That means that we not only lose those extra 40 minerals per minute for the original killed drone, but also 0.8 times this value for the ghost drone that you couldve built if the first drone wouldnt have died, in all subsequent minutes after the first, on top of those 40. For that ghost drone, we also need to calculate the additional lost minerals that we wouldve gotten if the ghost-drone wouldve been built and mined minerals, which in return couldve built another drone, the ghost-ghost drone so to say. This obviously leads to an infinity amount of income lost.
If we limit the amount of larva and mineral patches, we can calculate how much it actually cost us. If we have 3 stockpiled larva (after using all our money that we get without the drone that gets killed), this means we lose 50 for the original drone, 40 for another minute, 40 + 0.8x40 for the next minute and 40 + 0.8x40 + 0.8x0.8x40 for the third minute. This amounts to 259 extra minerals per minute until all mineral patches are full, at that time the player that lost the drone catches up with the one that didnt lose the drone. Now you sum up how many minutes passed until the mineral patches got fully saturated and you have a number. Basically the formula is: 50+ (40*larva + pow(0.8,(larva-1)x40 + pow(0.8,(larva-2)x40 ... until larva reaches 0 or all mineral patches are saturated) As you can do this with sigma, you can calculate the amount of resources lost as long as you know how many spare larvae you accumulated due to the drone dying and how many minutes are left until all mineral patches are fully saturated. If someone could figure out the formula for extra larvae that pile up due to a drone dying, we could actually completely answer the question.
This got a little longer than i had anticipated, i hope there arent any glaring mistakes left in the calculations due to me being too tired to look over it again. I could also be completely wrong and everything i wrote down is complete bullshit, so take it with a grain of salt.
On June 02 2011 04:16 VirtualAlex wrote: Here is the strategy I am employing. This is 2v2, although I want to attempt this in 1v1.
In this replay me (Zergxes) and my partner Quad-Harass the terran and obviously in this replay it's a huge success. We kill 7 total workers and delay terrans barracks buy well over 90 seconds. It's really funny.
I want to figure out what the minimum amount of damage I need to inflict is, to make this type of gambit worth it. Because I am losing minerals sending workers away for so long as well.
While you could figure out the mineral gain/loss, the true advantage you gain from doing harass like that is psychological.
And while that harass is nice and you did a lot of damage, you also had 700+ minerals banked at the end of it. So I'd say it was more detrimental to you in that regard. If your opponents weren't terrible and sending two workers at a time to deal with four (what?), or the protoss pulling all his probes to deal with it, you would have been very far behind.
This is extremely hard to compute. However I can see its usefulness, ie: knowing if a harass was cost-effective, or sucessful. For example, I killed 2 Probes with a Reaper. Did the Reaper payed it cost? Was it worth? Most of the time you cant tell, sometimes it gets tricky and you just "feel" behind/ahead but that's not really what's happening.
Assuming you were planning to make workers constantly until you reach, say, 70 workers, you can't replace the lost worker until you've created 70 workers (read: have 69 workers), and can build a worker for the 71st time, giving you 70 workers. The mining time lost is the time between when the worker was lost and when you build that last worker.
On June 02 2011 05:53 MShaw006 wrote: Assuming you were planning to make workers constantly until you reach, say, 70 workers, you can't replace the lost worker until you've created 70 workers (read: have 69 workers), and can build a worker for the 71st time, giving you 70 workers. The mining time lost is the time between when the worker was lost and when you build that last worker.
That depends on how many mining bases you have. You will replace the lost worker upon getting max saturation at all of your bases.
On June 02 2011 03:51 VirtualAlex wrote: I appreciate your collective inputs. It seems that losing early workers is quite detrimental. I came up with a good test, according to my assumptions of course.
Simply get into a test map as terran, and make non-stopped mining workers until 25, and record my total collect resources at the moment the 25th worker is constructed.
Then run the same test, except get one of my 11th scv killed. Work up to 25 workers and record my resources. This should accurately show the difference between the total mined and include the 50 extra minerals spent on the fresh SCV. I am at work now, but I will try to run this test when I get home. Unless someone else is willing.
This test is really not necessary. Lost minerals is: 50 minerals + 40 minerals/minute * 3.97 minutes (the time to build the last 14 probes).
I really thought my earlier response would answer the question you were asking - was there something else you were looking for here?
Protoss can make the loss much smaller by chronoboosting out (about) 3 workers per worker lost, so the number of workers catches up again.
Zerg can try to make up for the difference by making extra drones and making the army units somewhat later, so if you don't keep continous pressure against a zerg, the harass will lose effectiveness.
if you lose 1 worker before you reach - let's say - 10 supply as terran the worker you lose makes you get an extra worker at 9/10 when your opponent has 10/10 - we assume your opponent is also terran and hasn't lost any workers while your opponent makes a supply depot, you make your worker, so all you lose is the 50 minerals and build time since you start your depot later. A worker is not that much, I saw alot of pro games where a guy would lose the scouting worker without it affecting his game too much. The only real fact is when you lose your SCV to a protoss and he is blocking with a zealot, you are pretty much forced to defend vs 4gate.
Thanks alot for your input. Of course psychological damage is a huge factor. I have about 10 replays of this type of harassment working out amazingly well in which the terran player losts 4+ workers. One time we actually killed an entire supply depot. Another time, after he build his rax, he lifted it off and moved it instead of just making a marine.
But it's not possible to calculate that so I just wanted to get some hard numbers. It looks like killing 2 workers definitely makes it worth it. So I will keep developing this strategy.
EPT
