• Log InLog In
• Register

[D] harvester optimization - Page 5

Post a Reply Prev 1 2 3 4 5 6 Next All

MindRush Romania916 Posts April 21 2010 01:27 GMT #81 tested and true ! no actual timings and such, didn't time my builds i played against the AI on Desert Oasis ..... i played terran vs protoss i kicked the AI's ass and i left him with just 1 assimilator then i made 1 cc at a high yield and one at another , one space farther the number of SCVs were 24 at one, 24 at the other, the one fartherst finished faster the ai didn't mine from any high yield, and i saw that on the replay i don't post the replay, just try it yourselves

Aether Canada123 Posts April 21 2010 01:47 GMT #82 OK, I did a more thorough test on guysers. As I expected there is a difference but it's incredibly small. A guyser takes around 23 minutes to exhaust with normal placement and moving your HQ +1 only adds about 9 seconds to that 23 minutes. If we break it down into seconds, this is 2500/1380 vs 2500/1390. 1.811GPS or 108.7 GPM vs 1.798GPS or 107.9 GPM. As I expected, the difference is less than 1 units of gas per minute. You lose approximately 0.8 GPM, and you would ONLY lose that on the high-yield guysers. This equates to only about 14 Gas over the course of over 23 minutes. There could be a very slight difference depending on if the guysers are diagonal and in the centre or not, but I would still estimate an incredibly small difference. Possibly .9 as opposed to .8 The reason for this is as I stated earlier, with normal placement workers actually have to wait a split second before entering the guyser, so the added distance cuts into the waiting time before it cuts into the mining time, meaning that most of the extra distance is absorbed by the fact that probes don't have to wait to enter the guyser.

GreggSauce United States566 Posts April 21 2010 01:52 GMT #83 I think the real issue with this is its generally impractical with how the game is currently, could be useful in the future. From my understanding of reading all of this, and you will lose minerals until you hit 18* because of the distance right? before it starts increase because you can have 3 maxed out instead of 2 right? Most of the high resource games generally revolve around harrassing entire mineral lines and or killing off bases, making sure you have minimum 18 per minerals seems to be like another layer of complexity that might not even benefit you it could be very useful as terran or zerg that are taking high yield resources, but I don't know, someone may find this useful, I just don't see it.

GreggSauce United States566 Posts April 21 2010 01:56 GMT #84 also you can use a cheat program to speed up the single player game you seem to be a pretty bright person already so it shouldn't be too hard for you to figure out thats how i tested stuff like this in wc3

Aether Canada123 Posts April 21 2010 02:27 GMT #85 That's true, I did mention that myself. This is why I realize it has limited uses. I completely lied about this point that I'm editing out. It does require complete saturation to work. If you have 1 less than full saturation they work out to be pretty much even. ie: 17 vs 17. is equal and 18 vs 18 gives a bonus to the misplaced CC After some more tests this might actually be MOST beneficial for Terran because of what you're saying here. I did another test building 2 command centres, building probes from each one with a rally point at a single mineral node. When both patches have 20 workers, I queue up an additional 4 on 1 and 1 on the other. when the single SCV is done, I lift the CC off and replace it 1 space farther. The other CC continues building SCVs. Despite the fact that I fumbled badly trying to land my CC and kept it floating for a good 15-20 seconds, this mineral patch STILL caught up and finished first by about 15 more seconds. So considering the point about harassment, if a Terran used just a few workers at their high yield and spammed all of their MULEs there, while replacing a more well protected expo, the player will see a benefit without the risks associated with other races. As far as speeding the game up, meh. I just set them to mine, then minimize it and do something else. Only takes a couple of minutes to build the first few resource buildings and set up HQs in the right place. I think at this point I've done a pretty exhaustive battery of tests, I don't know how much more information there is to be acquired. I guess it's up to players who are better than me now to decide whether this will be beneficial in any given situation and whether it's worth using as a result. I think the practicality of it will depend largely on the map you're playing. On a map like metalopolis for example I think it would be highly impractical since the high yields are in a really difficult position to defend, but if you do this on say an island expo with some turrets or a high yield that's more easily defended then it could be viable. Something for people to play around with and decide if they can make use of it. edit P.S. I'm really interested to hear what any of the top level guys around here think about this. I think they're the ones who have the potential to make use of something as small as this. Nazgul, Idra, day[9], etc. (I only know a few from the replays I've watched, heh)

Moletrap United States1297 Posts April 21 2010 07:21 GMT #86 Mad fucking props, Aether. Just to be clear, this works with probes and drones as well?

Aether Canada123 Posts April 21 2010 07:22 GMT #87 yeah, all workers mine at the same speed.

imPERSONater United States1324 Posts April 21 2010 07:31 GMT #88 haha, it's amazing what we bored SC players will unearth! Discoveries like this are why I am so excited for beta to end and get the real game out the masses. The first year of a game has massive development in strategies and I love to see these come to light so early. Great work Aether and for the rest of out sake, I hope you find some more free time to goof around soon!

s031720 Sweden383 Posts April 21 2010 09:15 GMT #89 As rufio said. In addition, early game; Lets say your are making a FE. Your main will be saturated, so you can move that one 1 step away from the mineral lines and build a few more SCVs there to saturate it agian, and have your mules at the expo. All you ever need as terran is one base close to the minerals for your mules to work at and the rest one step away. Since the positive effect of "One-Step-Away-Mining" only kicks in when the mineralline is fully saturated, and when not using mules; the mules should always go to the newest expansion. (not to mention that mules also always should go to the mineralline with most unmined minerals left which in this case coincide). It would be interestin to see a spreadsheet of how much the income would benefit from having 2 ordinary minerallines and 1high yield mining with the "One-step-away-mining" idea, compared to the same but mining ordinarliy.

HoracE United Kingdom180 Posts April 21 2010 10:02 GMT #90 nothing short of genius.. very impressed. It is one of those simple scientific ideas that you think gosh should of thought of that but didn't. Love the logic and the way you have backed up your argument... you may well see this being used by the pros in tournaments which is definately kudos for your work. Congrats.

shoop United Kingdom228 Posts April 21 2010 10:22 GMT #91 Excellent post. Judging by the reactions, it could double as an intelligence test.

MiraMax Germany532 Posts April 21 2010 10:25 GMT #92 I'd like to join the praise. Even if it does not provide a huge advantage, it can be done with little effort, so it might just become as standard as worker splitting. I agree that it might be especially valuable for expanding terrans. I already foresee the new Flash of SC2 maynarding his scvs right before aethering his CC. Hell, it even sound right!!

sob3k United States7572 Posts April 21 2010 11:22 GMT #93 Can we get a picture showing this new positioning vs old positioning?

Aether Canada123 Posts April 21 2010 12:07 GMT #94 sorry clicked wrong button.

Liquid`Nazgul 22427 Posts April 21 2010 12:10 GMT #95 The problem here is that when you place your nexus one spot away anticipating saturation you are ill positioning yourself for worker harass. During the game he might kill off your workers and in plenty of games your count of 18 goes to a count of 12. Despite that there are certain scenarios where it could be worth it. Turtling Terrans, or games where you need to pull out a move to come back into it so you take a bit of risk trying to close the gap. Games where your main is mining out so you will know for sure you have 18 spare probes to mine to perfection on the new expo. Etc. Really good stuff though I had been playing around with a lot of this stuff already such as for example where in SC:BW you needed to send your first 8 probes to 8 minerals, in SC2 if you get the opportunity you need to send a second probe to the closer minerals so two of them can return minerals from that one more quickly.

Aether Canada123 Posts April 21 2010 12:19 GMT #96 right, but what do you think about Terrans using it in a main after they've expanded. ie you're in mid game, your expo has basically become your main because the main is mined out, and most of the harassment is actually happening at your high yield. So you lift off and replace the command center in the expo and spam mules at the high yield where the CC is in it's proper position. Then they have to either harass both mineral lines or choose between the high yield and the expo mining at slightly higher capacity. At this point you should be able to hold off harass at one of the two mineral lines so you're pretty much guaranteed a bonus beyond what the MULEs normally supply.

Liquid`Nazgul 22427 Posts April 21 2010 12:20 GMT #97 Yep I definitely think there are situations in practice where it will help you in the game. It's going to be super tricky to get a feel for these situations. In short: I love it. It's the sort of thing SC2 needed.

Aether Canada123 Posts April 21 2010 12:26 GMT #98 Yeah, like I said initially, I really think the uses are highly limited. I also think there are a small handful of situations where knowing to do it can give you a bit of an edge though, and I will be excited to see if someone is able to use it effectively some time in the future.

mathemagician1986 Germany549 Posts April 21 2010 14:29 GMT #99 Could someone please test how long it takes to mine out a gold mineral base with 16 (and mabye also with 17) workers with the standard CC position? In my opinion, that's the real question here. Are the 2 extra SCVs worth the investment in terms of mining efficiency or not? If a gold expo with 16 workers gets mined out just as fast as with 18 (due to workers changing patches regularly), the minerals/worker/time ratio is actually better than with the new method.

Kezzer United States1268 Posts April 21 2010 14:35 GMT #100 Am I missing something? I see in the OP the numbers to back up his arguement, bonus income. But what I don't get is that once you are fully saturated *normally* with about 20-24 workers, the mineral nodes are constantly being mined from. And that rate is constant... no amount of space from the CC/nexus/hatch will change that.

Prev 1 2 3 4 5 6 Next All

Please log in or register to reply.