EPTOn November 14 2004 17:51 BigBalls wrote:
i still think you should upload that rep
i still think you should upload that rep
Oh indeed. TL needs more consistant rep uploading =O
Maximum # of Worker per se - Page 5
| Forum Index > Brood War Strategy |
|
Commander{+}
United States2878 Posts
Oh indeed. TL needs more consistant rep uploading =O | ||
|
LOcDowN
United States1015 Posts
On November 14 2004 16:55 Firazpiral wrote: you need 5 drones per patch, 2 probes per patch, and 1 scv per patch Please read the my opening topic damn it. | ||
|
LOcDowN
United States1015 Posts
On November 14 2004 13:25 LordOfDabu wrote: So I did a little test on Lost Temple. I basically edited the map and added the following two triggers: Always: Create x Protoss Probe Player 1 Accumulates 12000 [that's 8 x 1500, the amount at one starting location] ]ore: End scenario in victory I then tried various values of x and sent all probes to mine. When the scenario ended I took note of how long it took to mine out the location. I first did 20-25 and then 30 and 35. All of these tests were done at the 9 spot. Hopefully my chart will come out okay. Number of Probes / Time taken to mine / Minerals/min / Cost of probes / Money gained 20 14:05 852 1000 11000 21 13:20 900 1050 10950 22 12:52 933 1100 10900 23 11:54 1008 1150 10850 24 11:13 1070 1200 10800 25 11:17 1064 1250 10750 30 11:00 1090 1500 10500 35 11:02 1088 1750 10250 Some things of note is that it sometimes took longer with more probes: I assume this is due to the randomness and/or movement of the probe AI (such as when a probe arrived at a location already being mined). It looks like the desired number of probes is 24, which is surprisingly exactly 3x the number of mineral patches. I'll probably rerun the test again later to verify that the times I get are close, as well as with scvs and drones. Thank you for sharing this with us. I appreciate the datas and your effort for taking the initiative. <3 | ||
|
88)WhyYouKickMyDog
United States608 Posts
On November 14 2004 10:33 LOcDowN wrote: Don't tell me there is no optimal value, as it is mentioned many times in this thread. Several peoples here also posted the saturation limited. From all these datas the reader should have a sense of the range in term of maximum efficiency & understanding the saturation limit. Ofcourse there is an inefficient factor as the number of SCV increase. However, I do think the inefficient factor can be decreased and this is the point of this thread - to find the optimal value for mineral efficiency. You do realize by understanding one variable, in this case our optimal value, we can compare it to other higher or lower set of SCV ranging from 15 (as you stated) to 40 (as GroT stated). One variable can trigger other questions to be ask. It can also set a standard value to measure for something else. nope, i was right. the saturation limit will change by like +-3 SCV's each time someone does testing, because theres a random element. also, the number they come up with will very rarely be useless, unless you're playing BGH and you have a REALLY long time for your SCV to bring in its 50 minerals to make it worth it. Say you had 23 SCV's and this magic number you guys come up with is 24. If the patches have say ~400 minerals left each, and you built a 24th, it would only hurt you. So, everything I said it right, and if somehow im wrong, I was just trying to help you out anyways. Don't need to be an asshole about it (or at least it seems like you were when reading your response). | ||
|
88)WhyYouKickMyDog
United States608 Posts
| ||
|
Abahgus
United States323 Posts
| ||
|
mr.FiSt
Czech Republic97 Posts
 | ||
Liquid`Drone
Norway28876 Posts
I noticed a slight, but quite small difference between 2.5 probes per patch and 3 probes per patch and 3 and 4 probes per patch gave me the *exact* same results. although 4 might give you sliiiiiiiiiightly more in the long run / depending on map, it's DEFINITELY not worth spending money on. apart from the fact that you want to expand, and you want to maynard peons. the term maynard didn't originally mean transferring peons btw, it meant overproducing peons in one base so you could transfer them and immediately have a new expo running. overproducing. in pvp you don't (depending on map) want more than 2.5 peons per patch because you usually don't want to expand very fast. in pvt you can make far more (at first, but eventually you stop producing them. ) zerg is obviously completely different from terran and toss, personally I very rarely have more than 1.5 peons per patch with zerg, usually between 1 and 1.5. | ||
Twisted
Netherlands13554 Posts
Just take another expand if you want more income -.- | ||
|
Maksim
United States22 Posts
I remember this from a study posted on Brattsunami's site a couple years ago | ||
|
ssregitoss
Turkey241 Posts
| ||
|
ssregitoss
Turkey241 Posts
Well, consider this: Player A loses 9 probes to a reaver drop (which is a reasonable number. You can almost always get 2 scarabs into the probes, particularly if you suicide the reaver). Heres what Player A REALLY lost, assuming that they IMMEDIATELY build their probes back up no matter what, possibly ignoring their defenses (which they generally cannot do): 50n + [12 * (n + (n-1) + ... +1)], which when simplified ammounts to: 50n + 6n(n+1) Don't worry about the math. I assure you its correct. n stands for the number of probes lost, and the first term is the cost of replacement, and the second term is the mining you lost out on. Obviously in the very very long term (mined out minerals), the second term becomes irrelevant, but since we ALL know 1000 minerals now is FAR better than 1000 minerals 25 minutes from now, the second term IS very important, especially since you usually only kill probes on somewhat fresh mineral batches. Just as a reference, I thought I'd just calculate a few sample loss numbers. Losing one probe is 62 minerals; 62 minerals per probe Losing five probes is 480 minerals;95 minerals per probe Losing 9 probes is 990 minerals; 110 minerals per probe Losing 10 probes is 1160 minerals; 116 minerals per probe | ||
|
LordOfDabu
United States394 Posts
| ||
|
BigBalls
United States5354 Posts
dabu got it before me | ||
|
Physician
United States4146 Posts
~ because efficiency will ultimately depend on ur survival ~ perfect economy can mean death. The balance of economy/unit production can only be achieved with a lot of experience and very thorough scouting. Terran 3 workers/patch >> bump to 3.5-4/patch when planning expo. Toss 2.5 workers/patch >> bump to 3/patch when planning expo. Zerg 1.5 workers/patch >> bump to 2.5/patch when planning expo. Terran and Toss are similar, making non stop workers one by one is usually done, worker production is only halted when mass units is a priority. Zerg is perhaps the most critical race in economy/unit production balance ~ a little mistake here usually costs u the game ~ specially given the fact that more units are needed early on that other races. Zerg needs less workers per patch always than the other 2 races. Hope it was of help. ~ Physician | ||
|
Beast_Bg
Bulgaria1623 Posts
 | ||
|
Jesus Christ
United States35 Posts
| ||
|
Allko
China297 Posts
| ||
|
BigBalls
United States5354 Posts
perfect example. he made 110 probes, i made 58. UNNECESSARY | ||
|
yeehaw
San Marino888 Posts
On November 14 2004 13:25 LordOfDabu wrote: So I did a little test on Lost Temple. I basically edited the map and added the following two triggers: Always: Create x Protoss Probe Player 1 Accumulates 12000 [that's 8 x 1500, the amount at one starting location] ]ore: End scenario in victory I then tried various values of x and sent all probes to mine. When the scenario ended I took note of how long it took to mine out the location. I first did 20-25 and then 30 and 35. All of these tests were done at the 9 spot. Hopefully my chart will come out okay. Number of Probes / Time taken to mine / Minerals/min / Cost of probes / Money gained 20 14:05 852 1000 11000 21 13:20 900 1050 10950 22 12:52 933 1100 10900 23 11:54 1008 1150 10850 24 11:13 1070 1200 10800 25 11:17 1064 1250 10750 30 11:00 1090 1500 10500 35 11:02 1088 1750 10250 Some things of note is that it sometimes took longer with more probes: I assume this is due to the randomness and/or movement of the probe AI (such as when a probe arrived at a location already being mined). It looks like the desired number of probes is 24, which is surprisingly exactly 3x the number of mineral patches. I'll probably rerun the test again later to verify that the times I get are close, as well as with scvs and drones. Just a point, I think mining out the location is inaccurate. Several times I have mineral patches mined out, while there are still some with almost 100 minerals. This is probably due to the mining algorithm. I suggest you try editing mineral values or do the test on BGH and see if there is a difference. | ||
| ||
