EPTBuild Order Calculator for Brood War? - Page 2
| Forum Index > BW General |
|
brooklyn kurtz
United States1 Post
| ||
|
rushz0rz
Canada5300 Posts
On December 28 2013 19:43 Birdie wrote: If I can, I'll try get it to work on Mac too, although I don't have one to test it on. I wouldn't bother. BW doesn't even work on OSX and most people who have OSX and play BW play it on Wine (can install your app with it probably anyway) or boot camp. | ||
|
J.Dong
United States102 Posts
is an extension of the build order search which has been in UAlbertaBot for years, but now works with all 3 races. You can put in a build order goal and it will try to give you time-optimal build orders to achieve that goal. It's open source on github.From what the author says, it's extremely accurate for everything excluding drone transfers. Unfortunately it is poorly documented and there are no pre-compiled binaries that I see. Edit: I found some stuff that may work. BOSS binaries don't seem to be on the list though... | ||
|
[[Starlight]]
United States1578 Posts
On January 01 2015 03:01 f10eqq wrote: Here's the old program you're thinking about (latest version is for 1.161). It's not. Cool program, but the one I'm thinking of was cross-platform and ran inside a browser. | ||
|
TelecoM
United States10670 Posts
On January 01 2015 06:16 J.Dong wrote: Here's another program at the cutting edge of Broodwar research: + Show Spoiler +. BOSS is It's open source on github. From what the author says, it's extremely accurate for everything excluding drone transfers. Unfortunately it is poorly documented and there are no pre-compiled binaries that I see. Edit: I found some stuff that may work. BOSS binaries don't seem to be on the list though... NICE, that is high tech stuff ! | ||
|
[[Starlight]]
United States1578 Posts
On January 01 2015 05:05 rushz0rz wrote: I wouldn't bother. BW doesn't even work on OSX It does work under OS X, though technically OS X is running the PowerPC code (emulation?). IIRC, it works all the way up through Snow Leopard (10.6), because Lion (10.7) and later don't support PowerPC code. Lion got released in mid-2011. One of my comps is a Mac with Tiger (OS X 10.4) that runs BW just fine. Also looks like Birdie flew the coop on this project anyhow. Can't blame him, it would've been a time sink. | ||
|
Birdie
New Zealand4438 Posts
| ||
|
lantern77
Canada60 Posts
I know this is an old thread and bumping it is kinda odd. But I'm totally down to build this in Web. Assuming that data from liquipedia (or some source) is reliable. It gets over the platform/compatibility issue. I have four months free time, so why not. The question is what features would be best implemented on top of this and what is the most reliable way to get all the data on build times? | ||
|
Jealous
10129 Posts
On April 13 2016 08:47 lantern77 wrote: Hey guys, I know this is an old thread and bumping it is kinda odd. But I'm totally down to build this in Web. Assuming that data from liquipedia (or some source) is reliable. It gets over the platform/compatibility issue. I have four months free time, so why not. The question is what features would be best implemented on top of this and what is the most reliable way to get all the data on build times? I would say the features necessary is build selection before you start a game. I will assume that you need a launcher or plug-in to do so. Then, you select builds based on Race -> Build. For example, Zerg -> 3H Muta -> Timing. If you are late on the timing, then alert goes up, or timer switches to negative. Oh, yea, timer is important for this imo. As for the second question, I feel there are three ways to approach this. The easiest, IMO, would be having a strong gamer do 3H muta timing vs. an empty slot and get the time from the rep. The second easiest is programming an AI to do the same, which I think LetaBot might be able to help you with. The third is to watch ZvT replays and see what timing the players hit and take the lowest time available. | ||
|
lantern77
Canada60 Posts
On April 13 2016 09:01 Jealous wrote: I would say the features necessary is build selection before you start a game. I will assume that you need a launcher or plug-in to do so. Then, you select builds based on Race -> Build. For example, Zerg -> 3H Muta -> Timing. If you are late on the timing, then alert goes up, or timer switches to negative. Oh, yea, timer is important for this imo. As for the second question, I feel there are three ways to approach this. The easiest, IMO, would be having a strong gamer do 3H muta timing vs. an empty slot and get the time from the rep. The second easiest is programming an AI to do the same, which I think LetaBot might be able to help you with. The third is to watch ZvT replays and see what timing the players hit and take the lowest time available. Oh I definitely like the idea for adding a timing overlay. Like a few minutes after my post someone asked this question http://www.teamliquid.net/forum/brood-war/507531-build-order-overlay Combine the two the overlay and the web build order timeline app and that would be Great. Possibly the web version can export a format that the overlay can take for users to get their timings. As to my second question I should have rephrased my question, "Where can I get the data for building/researching units/upgrades". Like how long it takes to build a single scv. I know liquipedia has that data, but is it reliable. With this data users can create their own build order timelines(Web app). | ||
|
molotow[eef]
Germany81 Posts
1. Feed an amount of replays into the program (all played on a chosen map) 2. Be able to set filters á la: "Max amount of hydras at time 7:00" + "Hydra Speed at time 7:00" or "earliest Dts" Get a list of replays which fullfill your filters + Informations á la: Amounts of Nexy, Hatcheries etc. / Amount of Workers. This would be freaking cool! | ||
|
pebble444
Italy2497 Posts
| ||
|
Endymion
United States3701 Posts
On January 01 2015 06:16 J.Dong wrote: Here's another program at the cutting edge of Broodwar research: + Show Spoiler +. BOSS is It's open source on github. From what the author says, it's extremely accurate for everything excluding drone transfers. Unfortunately it is poorly documented and there are no pre-compiled binaries that I see. Edit: I found some stuff that may work. BOSS binaries don't seem to be on the list though... i dont believe this accounts for worker transfers and drone travel time to construction sites' hit on mineral/gas income unfortunately, but if you use just to compare the validity of two BOs the differences should cancel out in all but the tightest timings... i also found one of the architect's papers on the subject which also doesnt mention travel time as a restriction, which is strange given how much it plays into combat calculations: paper | ||
|
kogeT
Poland2037 Posts
For me basic assumptions were: - You put the exact requirements on what you want to have (e.g 12 mutas + 30 drones vs 12 mutas ASAP would probably give you 12 mutas earlier but with less drones) - Gathering rate for gas mineral (per minerals setup / avg per map etc.) - Timeof going to all important locations (e.g natural, 2nd natural) and also avg. distance (or just time spent) to build specific buildings. This is for workers that have to be in a position - All variables on costs, time of building units/buildings etc. Now from there it's quite simple as it's just a set of optimization equations, step by step. By principle you always start with straight up macro (number of workers * avg gather rate for this amount of workers to get exact amount of mineral/gas that matches the sum of requirered tech/units), and afterwards you add up different constraints like (I need 1 turret max 7 min mark etc.) | ||
|
lantern77
Canada60 Posts
On April 13 2016 19:33 kogeT wrote: Funny I never seen that thread and also considered creating an engine for that. For me basic assumptions were: - You put the exact requirements on what you want to have (e.g 12 mutas + 30 drones vs 12 mutas ASAP would probably give you 12 mutas earlier but with less drones) - Gathering rate for gas mineral (per minerals setup / avg per map etc.) - Timeof going to all important locations (e.g natural, 2nd natural) and also avg. distance (or just time spent) to build specific buildings. This is for workers that have to be in a position - All variables on costs, time of building units/buildings etc. Now from there it's quite simple as it's just a set of optimization equations, step by step. By principle you always start with straight up macro (number of workers * avg gather rate for this amount of workers to get exact amount of mineral/gas that matches the sum of requirered tech/units), and afterwards you add up different constraints like (I need 1 turret max 7 min mark etc.) I agree to that was my thoughts, however at the time I didn't consider base layouts and worker movement. As indicated by Endymion. Plus this can get very theoretical, as molotow indicated. Mur, I hope I can get something started tonight. | ||
|
Endymion
United States3701 Posts
On April 13 2016 19:33 kogeT wrote: Funny I never seen that thread and also considered creating an engine for that. For me basic assumptions were: - You put the exact requirements on what you want to have (e.g 12 mutas + 30 drones vs 12 mutas ASAP would probably give you 12 mutas earlier but with less drones) - Gathering rate for gas mineral (per minerals setup / avg per map etc.) - Timeof going to all important locations (e.g natural, 2nd natural) and also avg. distance (or just time spent) to build specific buildings. This is for workers that have to be in a position - All variables on costs, time of building units/buildings etc. Now from there it's quite simple as it's just a set of optimization equations, step by step. By principle you always start with straight up macro (number of workers * avg gather rate for this amount of workers to get exact amount of mineral/gas that matches the sum of requirered tech/units), and afterwards you add up different constraints like (I need 1 turret max 7 min mark etc.) there are a lot of small considerations that are extremely hard to take into account, but add up to a material amount... like travel time to construction site.. it might not seem difficult, but there is a lot that goes into deciding where to place a building and when to send the drone depending on what's happening in the game.. for a super simplified optimizer you could just give it coordinates depending on your starting position, kinda like how people generally place their buildings in the same exact place depending on the build, but for a webapp it's difficult to simulate that. | ||
|
ArmadA[NaS]
United States346 Posts
| ||
|
ProMeTheus112
France2027 Posts
| ||
|
[[Starlight]]
United States1578 Posts
On April 13 2016 17:22 molotow[eef] wrote: In my opinion a build order calculator would always be too theoretical and abstract. It is nearly impossible to adjust it to the different maps (different flow of income, different amount of min patches) and different baselayouts. There was nothing theoretical or abstract about the BW Build Order Calculator I used back in the day. It worked, and worked well. It wasn't smart enough for you to just put in 'fastest BO to get 12 hydras' or whatever. You had to input various BOs, and through trial and error figure out what was fastest. But that was actually an advantage, because you got a feel for/better understanding of exactly WHY a particular BO was faster or better. Kinda like driving a manual vs an automatic. | ||
|
[[Starlight]]
United States1578 Posts
If so, perhaps y'all could coordinate/work together, thus making less work for each individual, and a greater chance of success via your pooled knowledge? Just a suggestion. | ||
| ||
