EPTSC2 Math! How Many Workers for Constant Prodcution - Page 3
| Forum Index > SC2 General |
|
Kaal
Djibouti2539 Posts
| ||
|
Moonling
United States987 Posts
 but anyways thanks for this i know it took a lot of time TL Appreciates | ||
|
Godstorm
Romania845 Posts
| ||
|
Qzy
Denmark1121 Posts
On October 11 2010 13:27 s2pid_loser wrote: tl;dr could u possibly bold ur results so i can jus see what ur findings were i dont doubt ur proof, im curious to see ur result thanks User was warned for this post I know he said it a bit rude, but this is what you do in the scientific world. Too much information about a proof can overwhelm you, therefore a conclusion need to be taken out of the calculations. Looks nice tho... Can it tell you how many workers you need on the new expo, before it can maintain itself? | ||
|
kidcrash89
198 Posts
On October 11 2010 16:15 Jaeger wrote: Getting #DIV/0! in your spreadsheet (for protoss & terran) if any of the units I'm creating don't cost gas. In fact it seems completely bonkers? this data: + Show Spoiler + Number of buildings producing 1st unit = 3 Mineral cost of 1st unit = 100 Gas cost of 1st unit = 1 Time to produce 1st unit = 28 Supply cost of 1st unit = 2 Number of buildings producing 2nd unit = 6 Mineral cost of 2nd unit = 50 Gas cost of 2nd unit = 150 Time to produce 2nd unit = 45 Supply cost of 2nd unit = 2 Number of buildings producing 3rd unit = 3 Mineral cost of 3rd unit = 100 Gas cost of 3rd unit = 1 Time to produce 3rd unit = 28 Supply cost of 3rd unit = 2 yields: Workers to put on minerals = 55 Workers to put on gas = 1334 Total workers = 1389 Retyped the gas code. It should give the following now: + Show Spoiler + Number of buildings producing 1st unit = 3 Mineral cost of 1st unit = 100 Gas cost of 1st unit = 1 Time to produce 1st unit = 28 Supply cost of 1st unit = 2 Number of buildings producing 2nd unit = 6 Mineral cost of 2nd unit = 50 Gas cost of 2nd unit = 150 Time to produce 2nd unit = 45 Supply cost of 2nd unit = 2 Number of buildings producing 3rd unit = 3 Mineral cost of 3rd unit = 100 Gas cost of 3rd unit = 1 Time to produce 3rd unit = 28 Supply cost of 3rd unit = 2 Workers on minerals = 55 Workers on gas = 30 Total workers = 85 Let me know if the div by zero thing still occurs, I don't think it will now. | ||
|
bud_b
Turkey4 Posts
This is a really useful tool! Thanks a lot! :D To all those who say that this is pretty much useless data, try to think of it in another way: We can make use of this information (and tool) to calculate X (the number of production facilities) instead of G (the number of gatherers). We can assume that G is a given function of number of bases (I mean assume full saturation). That would answer questions like: "How many barracks can a fully saturated single base support?" "If i have a fully saturated zerg base and i am somehow unable to expand, do i need a second hatchery to be able to spend all my minerals while making zerglings and banelings?" | ||
|
cryu
37 Posts
| ||
|
PsykoMantis
United States203 Posts
I want this info more than anything to see if it is better on 2 base for a certain build to have 3 geysers fully saturated, or 4 geysers with 2 probes on each geyser. edit: just read the link to the tl post for measuring mining speed and that answers my question. edit: now that I thought about it some more, I definitely think I was just over thinking things. + Show Spoiler + One last thing that I was thinking would be really interesting is to see the effect of upgrades on production, I think you could represent it using the following: If you consider the result that is calculated for the production for a given #probes/bases as a steady state, you could graph production (y-axis) vs time (x-axis) and it would give you a straight line. However, if you were to throw in an upgrade into the graph at some time t0, you would see a dip in production (obviously) on the graph, but what I am more interested in is how long it takes for production to recover to its steady state value. For example this would answer a question such as: if i am making constant marine production, i know in T amount of time, I will have X marines. However, if I were to put in an upgrade that cost 100 minerals somewhere in the time period T, do I just end up with 2 less marines at time T or do I have even less marines? I am probably just over thinking the situation on this one, any thoughts? | ||
|
Takkara
United States2503 Posts
You don't want to add macro gaps, harassment, drone caps, etc to the spreadsheet because it's not a simulator. It's a model. It gives you the optimal minimum number of drones harvesting at max capacity you need to handle your unit production. Of course harassment lowers your abilities, of course drone mashing lowers your yield, and supply facilities, unit producing structures, lag, etc, etc, etc. But you know that if you don't AT LEAST have what this model says, then you can't support what you're trying to produce. A very useful and educational tool. The web app particularly is amazingly well designed. | ||
|
kidcrash89
198 Posts
On October 11 2010 22:27 PsykoMantis wrote: Wow this is awesome, I was wondering about one addition to this though (I haven't looked at the excel sheet yet, just the web app) but being able to input how many probes are gathering gas from each geyser. I know it seems like really I am just splitting hairs, but I sort of want to know the exact amount of probes required, for the time being I am just using some fraction of geysers used to find any non-multiple of 3 probe amount on gas.. so for example 5/6 probes on 2 geysers would be 1.667 geysers? I want this info more than anything to see if it is better on 2 base for a certain build to have 3 geysers fully saturated, or 4 geysers with 2 probes on each geyser. edit: just read the link to the tl post for measuring mining speed and that answers my question. edit: now that I thought about it some more, I definitely think I was just over thinking things. + Show Spoiler + One last thing that I was thinking would be really interesting is to see the effect of upgrades on production, I think you could represent it using the following: If you consider the result that is calculated for the production for a given #probes/bases as a steady state, you could graph production (y-axis) vs time (x-axis) and it would give you a straight line. However, if you were to throw in an upgrade into the graph at some time t0, you would see a dip in production (obviously) on the graph, but what I am more interested in is how long it takes for production to recover to its steady state value. For example this would answer a question such as: if i am making constant marine production, i know in T amount of time, I will have X marines. However, if I were to put in an upgrade that cost 100 minerals somewhere in the time period T, do I just end up with 2 less marines at time T or do I have even less marines? I am probably just over thinking the situation on this one, any thoughts? Between the web app and the spreadsheet I think you should be able to get a good picture of generally how many probes you'll need. Realistically mechanical errors, like losing .25 sec for not having made something really throw the data off a lot more than you might think. The best way to use this is to use the estimate and then see how it works in application. Ultimately, you'll need to tweak it. These will send you in the right direction though. | ||
|
kioskmongo
Sweden33 Posts
| ||
|
CursOr
United States6335 Posts
Imagine cutting off 2 bases with the exact saturation to crank an exact composition at maximum efficiency (perfect amount of production buildings)... it would be damn near impossible to stop. Anyone who is interested in ridiculously strong 1/2 base pushes should try probe/scv cutting. | ||
|
Zocat
Germany2229 Posts
On October 11 2010 12:51 kidcrash89 wrote: *** Mineral patches never become over-saturated. We can have an infinite number of workers on one mineral patch and each will mine at the same rate as one worker would from its own distinct patch. Imho this assumptions kills your entire model. 8+ and especially 16+ workers on 1 base (assuming 8 mineral patches) will heavily influence your results. Apart from that though your work looks great. | ||
|
Merikh
United States918 Posts
| ||
|
kidcrash89
198 Posts
On October 12 2010 02:44 Merikh wrote: So I was reading the liquidpedia zerg tips section. Has the op took in consideration about close and far mineral patches? Apparently to saturate a field you only need 2 workers on close mineral patches and 3 on the farther back mineral patches so I'm curious. I averaged the results from the far/close mineral fields based on Kulas top-left spawn, which give something like 1 close field and 7 far fields. The gas geysers provided different mining rates as well, something particularly noticeable on Kulas. You could try to make a perfect model for every map and starting position, but it didn't seem practical to me. You could simply substitute in different values of the average time it takes to mine the fields for any specific map and starting zone. I just chose Kulas arbitrarily. You could account for oversaturation too, but that adds a lot of trouble in the math that I'm not as familiar with since the effects of saturation aren't prominent until about 16+ drones or so (and then you need to factor in how many drones are at ease base, etc). Rounding the drone count (since you can't have 13.5 drones) and considering realistic assumptions, like the fact you can't do 3 things at once and you are likely to miss timings, even if only by a hair of a second, seem to counterbalance the fact that oversaturation isn't considered. No amount of math can account for mechanical errors. I can try to add something after the calculations to consider saturation. | ||
|
Pokornyx
United States19 Posts
| ||
|
Rokk
United States425 Posts
On October 12 2010 02:37 Zocat wrote: Imho this assumptions kills your entire model. 8+ and especially 16+ workers on 1 base (assuming 8 mineral patches) will heavily influence your results. Apart from that though your work looks great. Mining is a near constant increase until you get to over 2 workers per patch. After that I think it increases about half as much per worker until 3 per patch. The model still works fine as long as you understand the limitations of it. | ||
|
GeorgeForeman
United States1746 Posts
| ||
|
onmach
United States1241 Posts
I've been making mistakes like trying to do a hydra break with 3 gas and 2 queens, when I could have gotten away with one queen and 4 gas, and gotten significantly more hydras out in the process. Also my muta build had too many drones on minerals, more than I need to make a few zerglings and crawlers. I can also see that it matches up with destiny's zergling bust on protoss fast expands he was doing last night. Roughly 20-25 drones on minerals, speed, +1, two queens, constant production, enough to beat multiple sentries, cannons, etc. with good control. | ||
|
AdelSC123
France362 Posts
 | ||
| ||
