|
I know this is an old tired topic that comes up all the time and never really gets answered with any sense of solid conclusion. I've taken a slightly different tack in attacking the problem that I hope will put it to rest once and for all.
************ADDED SOMEWHERE ON PAGE 2************
Timings on 6pool through 14pool, using the various orders. Supershort results: 12pool with double extractor before overlord, 11pool with single extractor before overlord, but 13+pool should always be with 9OL. Also, never ever ever overpool. 11pool with extractor trick is superior in every possible way.
*******************************************************************
tl;dr version:
-----------------------
I used the AI to test the various builds in a very consistent manner, to try to answer the question once and for all.
The 9 overlord is the best economic build, period.
Minerals at 2:15: 370
10 overlord can be used for 10pool and overpool, but is inferior to 9OL in every other possible respect to the tune of 10 minerals, and loses the small window of advantage it gets by building the 10th drone sooner by the time 10OL finishes its 11th worker.
Minerals at 2:15: 355
A single extractor trick doing the trick before building the overlord (11 drones then overlord) gives a small time window of about a 15 mineral advantage, between workers 11 and 14, but there's a cost, of around 5-10 minerals. Best case, you *might* be able to break even when compared to 9OL, but most of the time you'll lose 5-10 minerals.
Minerals at 2:15: 364
Single extractor trick after the overlord (10 drones, overlord, trick to 11 drones) is inferior to 10-T-OL in the 11-13 worker range, and is inferior to 9OL in the 14-15 range, so I honestly see no reason to ever trick this way, UNLESS you meant to 9OL, but forgot the overlord. In that case, this is your best way to try to catch back up, because the cost is 0-5 minerals instead of 5-10.
Minerals at 2:15: 369
The double extractor trick provides such a small advantage over the single extractor trick during even a smaller window that it should probably only be used exactly to 12 pool, and even then, there is an extra cost over the single extractor trick. It probably compares to the single extractor trick much like 10OL compares to 9OL, meaning that if you're doing something exactly on drone 12, you might use it, otherwise, the single trick is probably strictly superior.
Minerals at 2:15: 353 (imprecise, but again, likely strictly inferior to single trick except for fastest 12pool)
-----------------------
BTW, are you concerned yet that the tl;dr version is that long? 
Full version
------------
In an attempt to put this question to rest once and for all, I decided to set up a computer player to perform everything at as close to the exact correct timing as possible. There are many advantages to testing this way rather than testing on your own, the primary reason being that the computer can be much closer to perfect than you.
Computer players do a pretty dang near-perfect 3/3 split (a lot better than any 3/3 split I've ever seen in a replay), and with some manual rally-point management as well as spending a *lot* of time trying to figure out exactly when the best time to trigger the extractor trick was (and how to do it to begin with in the map editor, which definitely was not easy), I was able to get truly consistent timings. What I ended up with was the as near-perfectly as possible for the first 15 drones, and certainly performing much better than any player could, especially performing the extractor trick, where the AI is able to perform a number of completely separate actions at the exact same time. (build a drone, cancel an extractor, put a worker back on minerals)
Note that there was actually some amount of variation here. This shocked me at first, because the timing numbers would come through and match EXACTLY, even for entire sequences up to 15 drones, down to the one-sixteenth of a second that it looks like the timer resolution might be. Which makes perfect sense, because the computer player is going to make the exact same decision at the exact same time *every* time.
But there turned out to be quite a lot of variation. I'd only see 2-3 patterns of timing, and once the pattern was established, it would match exactly, but there was as much as a full second's difference in later drone start timings. I puzzled over this for a really long time, and then as I was watching a sequence and saw a cocoon pop up, it hit me.
Larva movement is random, and the start position of your larva determines how far the drone has to move once it's been created. What's surprising is that this one little thing can make a rather large difference.
Since there was quite a bit of variation, I ran through a number of times, and picked out the most consistent timings for each variation. Anyway, the point here is that although this test was not absolutely consistent (even though I thought when I started that it should be), it's definitely going to be more consistent than testing by hand.
With no further ado, here are the results. sdXX is when drone number XX was started, and FDXX is when it completed building. Each point records the exact time, as well as the number of minerals that the player had at that point. sovl and FOVL represent the overlord, and EXTR and CANC are the extractor when doing an extractor trick. I also printed the number of minerals at 2:15 (135 seconds), by which time all 3 builds were complete. Time measurements are all in gametime, not real time.
9OL
sd07 0 0
sd08 15.9375 0
FD07 17 10
sd09 26.5625 5
FD08 32.9375 30
FD09 43.5625 80
sovl 46.625 0
sd10 54.375 0
FD10 71.375 105
FOVL 71.625 60
sd11 71.625 60
sd12 71.6875 10
sd13 78.125 0
FD11 88.625 70
FD12 88.6875 70
sd14 91.875 30
FD13 95.125 65
sd15 106.938 110
FD14 108.875 125
FD15 123.938 270
At 2:15, this build had 370 minerals
10OL
sd07 0 0
sd08 15.9375 0
FD07 17 10
sd09 26.5625 5
FD08 32.9375 30
sd10 38.9375 0
FD09 43.5625 30
sovl 55 10
FD10 55.9375 10
FOVL 80 120
sd11 80 120
sd12 80.0625 70
sd13 80.125 20
sd14 95.25 75
FD11 97 85
FD12 97.0625 85
FD13 97.125 85
sd15 110.813 140
FD14 112.25 150
FD15 127.813 290
At 2:15, this build had 355 minerals
There is exactly *one* point of time where 10OL is ahead of 9OL. From 55.9375 seconds until 71.375 seconds, 10OL has one more worker running, for a gain of about 15 worker seconds. But 9OL finishes drones 11 and 12 almost 10 seconds faster than 10OL, so this 15 second advantage begins to be lost at 88.625, and is completely gone somewhere around 96 seconds in. After that point, 9OL is ahead of 10OL in every way. It finishes the 15th drone 4 seconds faster, and is 15 minerals ahead at the 2:15 mark.
As a result of this, the only situations where it is not better to 9OL are when you are 10-pooling or overpooling.
SINGLE EXTRACTOR TRICK BEFORE OVERLORD (or 10-T-OL)
sd07 0 0
sd08 15.9375 0
FD07 17 10
sd09 26.5625 5
FD08 32.9375 30
sd10 38.9375 0
FD09 43.5625 30
EXTR 49.6875 45
sd11 50.3125 0
CANC 50.3125
FD10 55.9375 49
sovl 64.375 4
FD11 67.3125 29
FOVL 89.375 139
sd12 89.375 139
sd13 89.4375 89
sd14 91.5 44
FD12 106.375 159
FD13 106.438 159
sd15 106.563 114
FD14 108.5 129
FD15 123.563 249
At 2:15, this build had 364 minerals
This is a little more difficult to figure, so lets start with the easiest result: at 2:15, this build was exactly 6 minerals behind the 9OL build. That amount happens to be exactly what you lose from the extractor cancel, although technically it works out that you're 10 minerals behind, not 6, since there's nothing you can ever spend 4 minerals on.
Now the hard part: figuring out exactly where this has an advantage over 9OL. Drone 10 finishes at 55.9375, and drone 11 finishes at 67.3125. So we gain 15.43 worker seconds on drone 10, and 21.31 worker seconds on drone 11, for a total of 36.75 worker seconds. At 88.6875 seconds in, 9OL finishes worker 12 ahead of the extractor trick, gaining 17.69 worker seconds back on drone 12, and also finished worker 13 11.31 seconds ahead, which leaves a 7.75 worker advantage starting at 106 in seconds unaccounted for, which only builds on drones 14 and 15, becase 9OL finished them both about a third of a second later than the extractor trick. Call it 8.5 seconds for good measure.
So the extractor trick *should* be 8.5 worker seconds worth of mining ahead of 9OL, but ended up being at exactly the same point mineral-wise, minus what was spent on the extractor. Some of this is the loss of time for the worker building the extractor. This drone building the extractor is the one closest to the extractor, and it drops off minerals at 47 seconds, builds the extractor at 49, and touches the minerals after cancelling at 52 seconds, basically taking around 5 seconds to get from base to mineral patch, a trip that normally takes about 2 seconds.
So that accounts for 3 seconds, but we should still be 5 seconds ahead.
Well, it turns out that 5 seconds is less than the amount of time it takes for one worker to collect and return 5 minerals. So for a lot of extra complexity, we've made a very small gain in time, but not gained enough time to overcome the loss we take from cancelling. Again, there is quite a bit of variance here. I saw 2:15 marks from 364 to 374, with 364 being the most common.
There is a period here though, where this build has a 36.75 worker second advantage though, which works out to be somewhere around 25 minerals (or 15, after taking out the extractor trick penalty). If having an extra 15 minerals between the 11th and 14th drone speeds things up enough that it's worth 5-10 minerals later, then the extractor trick is worthwhile. So 12-14 pools could extractor trick rather than 9OL (but could also choose to 9OL if the 5-10 minerals aren't worth the slightly increase pool speed), everything else should 9OL.
If you perform the extractor trick well, that is.
SINGLE EXTRACTOR TRICK AFTER OVERLORD (or 10-OL-T)
sd07 0 0
sd08 15.9375 0
FD07 17 10
sd09 26.5625 5
FD08 32.9375 30
sd10 38.9375 0
FD09 43.5625 30
sovl 55 5
FD10 55.9375 10
EXTR 65.6875 45
sd11 66.1875 0
CANC 66.1875 0
FOVL 80 64
sd12 80 64
sd13 80.0625 14
FD11 83.1875 29
sd14 91.1875 34
FD12 97 89
FD13 97.0625 89
sd15 106.313 99
FD14 108.188 114
FD15 123.313 254
369 at 2:15
Over a number of runs, the final mineral result bounced pretty consistently from 369-374, which gives an overall cost of 0-5 minerals, which is better than 10-trick-OL by 5 minerals. If we compare this build to 10-T-OL, we can see the following:
Drone 11 comes out 15.875 seconds later, putting us behind by that many worker-seconds at 83.1874 seconds. We start to gain this back on drones 12 and 13, which are each faster by 9.375 seconds, and that both finish at 97 seconds. So at 104.9375 seconds, this build has caught back up to 10-T-OL, and at 106.375, we're ahead by 2.875 worker-seconds. We gain another .312 seconds on drone 14, and .25 seconds on drone 15, giving us a grand total of 3.4375 worker seconds advantage over 10-T-OL.
BUT....
We already know that both of these builds are behind, or at best equal to, 9OL, which is far easier to execute. The only reason to doing the extractor trick at all is to get an advantage at some point in the 10-15 worker range in order to get some building out faster. 10-OL-T has a smaller advantage over 9OL than 10-T-OL in the 10-13 range, and has a very negligible advantage over 10-T-OL at around the time we hit 15 supply.
So if we want to build in the 12-14 range, 10-T-OL clearly has an advantage in that time range, and if we're waiting til 15 supply (the only time 10-OL-T has an advantage over 10-T-OL), then we should just 9OL.
At the end of all that, 10-OL-T seems inferior to 10-T-OL in the 11-13 range, and inferior to 9OL in the 14-15 range, so there doesn't seem to be any valid reason to ever do this build. The only reason I can think of would be to be less behind if you meant to 9OL but forgot the overlord. In that case, tricking after the overlord will put you less behind than tricking before the overlord.
DOUBLE EXTRACTOR TRICK
The AI does not have enough flexibility to allow it to do a double extractor trick optimally. Specifically, I wasn't able to tell specific workers exactly where to build, and it makes poor decisions about when and where to take extractors and which workers to use in this case. What I can do though is extrapolate some numbers based on when the extractor trick would start, and the fact that we know it takes about 3 seconds of time away from the workers building the extractors.
We would be starting the double trick at around 55 seconds (the time when 10OL starts an overlord). This has drones 11 and 12 finishing at 72 seconds, giving us a 33 worker-second advantage for these two drones over 9OL. We lose some of the worker 11 advantage we got from 1 extractor trick by delaying though. We get 33 seconds for 11 and 12, but we were getting 21 seconds on just 11 before, so we gain 12 more worker seconds. 12 worker seconds is worth somewhere around 7-8 minerals, and it cost us 5 to do the second extractor trick. So we've basically gained 2-3 minerals worth of worker mining time over a single extractor trick, and that's only if we do things completely perfectly, which is a lot harder to do with a double cancel. In addition, drones 13-15 will be even more delayed, giving 9OL more of a chance to catch up or even pass us.
I would guess that the only time you should even consider doing a double-extractor trick is if you're going to double-trick, and build your pool on 12 before your overlord. The advantage simply isn't big enough compared to the double-extractor to make anything else worthwhile.
I did let the AI run a double-extractor trick, but it was fairly sub-optimal, wasting somewhere around 5-10 droneseconds due to poor choices when building the extractors. The tests that I did run though, showed it finishing drone 15 later (by many seconds) than the other builds, and showed it finishing with 343-353 minerals at 2:15, or 10-20 minerals behind the single trick. I would guess that an optimal build would bring that closer, but drones 13, 14, and 15 are all going to be later in a double extractor trick than a single, and I don't think the time gained from worker offsets that.
KICKS AND GIGGLES: 8 OVERLORD
Just for grins, I plugged in an 8OL build. Posting timings really isn't necessary, because this build is behind 9OL at every point, and ends up with a 25 mineral deficit. Basically, never build your overlord before 9. Ever.
|
Sadly this was already calculated very early in the beta, when slush started 9overlording and someone calculated it on spreadsheets or something ><. But nicely done ;o
|
you should put the results of the 3 tests in the tl;dr. <3
Other than that, I'm glad to see the math! great job.
|
On July 04 2010 09:43 Templar. wrote:
Sadly this was already calculated very early in the beta, when slush started 9overlording and someone calculated it on spreadsheets or something ><. But nicely done ;o
Well, the problem with calculating things on spreadsheets is that they can give inaccurate results.
For example, working from pure math, people get that doing an extractor trick can gain you minerals, when the fact of the matter is that trying to actually do it in game, there are imperfections.
For example, even the "uber-fast AI" whose reflexes are instantaneous has to wait a sixteenth of a second between making two workers "at the same" time.
|
Nice to have definitive info, all the other threads about this stuff I saw was sort of inconclusive and didn't have great control of the variables.
|
Could you please publish the map/AI, so that others could simply run each case and see for themselves?
It was on my todo list too, but glad it's already done and confirmed yet again. The advantage of doing it with AI is that there's no speculation about whether the tester does everything at maximum efficiency, and the test could be repeated many times easily, for more stable results.
|
On July 04 2010 10:14 Skrag wrote:Show nested quote +On July 04 2010 09:43 Templar. wrote:
Sadly this was already calculated very early in the beta, when slush started 9overlording and someone calculated it on spreadsheets or something ><. But nicely done ;o Well, the problem with calculating things on spreadsheets is that they can give inaccurate results. For example, working from pure math, people get that doing an extractor trick can gain you minerals, when the fact of the matter is that trying to actually do it in game, there are imperfections. For example, even the "uber-fast AI" whose reflexes are instantaneous has to wait a sixteenth of a second between making two workers "at the same" time.
im not positive if he did it on spreadsheets, but the numbers seem to be exactly the same as his anyways ;p
|
|
|
Hmmm. I'm a zerg player, but I usually just do the 10 OL. I've never done a 9 OL since broodwar. I may reconsider doing a 9 OL build now, but Im not sure, I'll probably just do the regular 10 OL. If more pros start to do this, then I will definitely do this, but a lot of research before this said that 10 OL was better. Maybe they're right, maybe you're right I don't know, but when beta comes back, I'll definitely experiment with 9 OL
|
Where exactly were the tests taken? If the spawn destination was closer or farther from the mineral patches, it would change the results.
|
On July 04 2010 10:49 EssayReader wrote:
Where exactly were the tests taken? If the spawn destination was closer or farther from the mineral patches, it would change the results.
I took the official Lost Temple map and added all the necessary stuff to get the AI to do the builds. I started out doing it at the 9 oclock position, and while I was doing that, I discovered by setting AI up at 3, 12, and 9, that there is actually a good amount of difference between the start positions. For example, there's a very slight disadvantage for being at 3 vs 9, which are 2 positions you would think are exactly equal. They're not, but it's pretty close.
I ended up running the final tests at 6 oclock after I figured out that the random larva movement was causing potentially large differences. For example, if you're at 9 and a larva moves to the far right underneath the hatchery, that has a much larger impact than if you're at 6 and the larva moves to the far right.
Obviously the exact results will vary based on the map and the individual start positions, but I haven't taken the time to test any of that yet. I might, but I wanted to get one really solid test in to see if the results were really conclusive, in a way that you couldn't suspect player error was at fault for the differences in timing.
|
On July 04 2010 10:46 Tazza wrote:
Hmmm. I'm a zerg player, but I usually just do the 10 OL. I've never done a 9 OL since broodwar. I may reconsider doing a 9 OL build now, but Im not sure, I'll probably just do the regular 10 OL. If more pros start to do this, then I will definitely do this, but a lot of research before this said that 10 OL was better. Maybe they're right, maybe you're right I don't know, but when beta comes back, I'll definitely experiment with 9 OL
If by 10OL you mean extractor trick, then there's probably not enough of a difference to make a huge difference.
If you're talking about the flat 10OL without extractor trick, you should stop doing that right now. I've never seen any pro do that, and it's provably inferior to the other builds. They're all doing 9OL or extractor tricks. (which are technically 11OL)
|
seen similar posts a lot of times, but this one was by far the most convincing. very nicely done.
|
I didn't read the entire OP but is there any downside on 9OL?
|
its up to u i guess. if u do 10 ol u should do 1 extractor trick for the scout. thats the earlier benefit. else use 9ol
|
Yeah I've always done the 9 OL unless I want to 10 pool --> 9 worker --> build drone --> extractor trick --> another drone.
|
On July 04 2010 13:22 lolaloc wrote:
I didn't read the entire OP but is there any downside on 9OL?
The only downside is that there are very small windows of time where the other builds can have a (very small) advantage.
For example, if you're going for a 12 pool, want it out as fast as possible, and don't care if it costs a little extra, you should double extractor trick. A 13 pool with the same requirements, extractor trick will get it a little faster.
If you don't build your first building til 14 or later though, you should never even consider doing anything other than 9OL.
|
I'm not even a Zerg player and I think this is really great info. Assuming your tests are accurate, this solves a lot of unanswered questions. Very comprehensive.
Now when I Z for fun (or play random) I'll know to 9ovie.
|
Like you said in the OP extractor trick is only for build timings, it is never more minerals, but sometimes if you plan to do lets say a 12 pool, 14-15 hatch at natural, using double extractor can let you delay your first overlord a little bit longer, giving you more time to get your natural down. Otherwise 9 OL is best.
|
Skrag, why do you build OL at 11?
I always do the single extractor trick AFTER I've built the overlord, and I imagine most top players do it that way.
You should try this scenario out:
sd07
sd08
FD07
sd09
FD08
sd10
FD09
sovl
FD10
EXTR
sd11 (later)
CANC
FD11
FOVL
sd12 (earlier)
sd13 (earlier)
sd14 (should be on time with 11OL since there's no larva block occuring)
Not sure if the finishing drone timings match exactly, but you get the hang of it. Single extractor trick after OL on 10.
|
161 Posts
On July 05 2010 07:20 LaLuSh wrote:
Skrag, why do you build OL at 11?
I always do the single extractor trick AFTER I've built the overlord, and I imagine most top players do it that way.
good catch, i didn't even notice that :p definitely OL on 10 and extractor trick after it.
also, the results are not entirely conclusive because people often send a drone to scout early on, and people do not have the micro of a computer. i would presume a human starts the 9OL later than the computer by a few seconds, but a human is also worse at the extractor trick.
in conclusion i think every player has to find out for themselves what works best, but the 2 methods are in the end more similar than different.
|
On July 05 2010 07:20 LaLuSh wrote:
Skrag, why do you build OL at 11?
I always do the single extractor trick AFTER I've built the overlord, and I imagine most top players do it that way.
Not sure if the finishing drone timings match exactly, but you get the hang of it. Single extractor trick after OL on 10.
I thought about trying this, but building the overlord after the 11th drone doesn't actually give you an overall economic advantage, instead only giving you a very small window of advantage from drones 11-14, and that advantage all comes from getting drones 10 and 11 faster than 9OL. If you delay the 11th drone, that advantage will shrink, likely making 10-OL-trick inferior to 10-trick-OL in every situation.
For the record though, I'm fairly sure that from the replays I've seen of top players (and I watch a LOT of replays), 10-trick-OL is the standard way of doing the trick.
Of course, the nice thing is that I already have a setup that makes it easy to try, so rather than theorizing, I can just modify the build order and see what comes out.
|
Ok, report from doing the extractor trick after the overlord rather than before:
sd07 0 0
sd08 15.9375 0
fd07 17 10
sd09 26.5625 5
FD08 32.9375 30
sd10 38.9375 0
FD09 43.5625 30
sovl 55 5
FD10 55.9375 10
EXTR 65.6875 45
sd11 66.1875 0
CANC 66.1875 0
FOVL 80 64
sd12 80 64
sd13 80.0625 14
FD11 83.1875 29
sd14 91.1875 34
FD12 97 89
FD13 97.0625 89
sd15 106.313 99
FD14 108.188 114
FD15 123.313 254
369 at 2:15
Over a number of runs, the final mineral result bounced pretty consistently from 369-374, which gives an overall cost of 0-5 minerals, which is better than 10-trick-OL by 5 minerals. If we compare this build to 10-T-OL, we can see the following:
Drone 11 comes out 15.875 seconds later, putting us behind by that many worker-seconds at 83.1874 seconds. We start to gain this back on drones 12 and 13, which are each faster by 9.375 seconds, and that both finish at 97 seconds. So at 104.9375 seconds, this build has caught back up to 10-T-OL, and at 106.375, we're ahead by 2.875 worker-seconds. We gain another .312 seconds on drone 14, and .25 seconds on drone 15, giving us a grand total of 3.4375 worker seconds advantage over 10-T-OL.
BUT....
We already know that both of these builds are behind, or at best equal to, 9OL, which is far easier to execute. The only reason to doing the extractor trick at all is to get an advantage at some point in the 10-15 worker range in order to get some building out faster. 10-OL-T has a smaller advantage over 9OL than 10-T-OL in the 10-13 range, and has a very negligible advantage over 10-T-OL at around the time we hit 15 supply.
So if we want to build in the 12-14 range, 10-T-OL clearly has an advantage in that time range, and if we're waiting til 15 supply (the only time 10-OL-T has an advantage over 10-T-OL), then we should just 9OL.
So, I realize you're a way better player than me, but this all suggests to me that you, and anybody else who is doing the trick after the overlord, should stop, and either 9OL or trick before the overlord, depending on whether you want to build a pool before 14 supply.
|
By the way, just for a frame of reference when we're talking about small numbers of worker-seconds, the conversion rate from worker-seconds to minerals ranges from .67 min/sec to .75 min/sec depending on how close the specific patch being worked is.
So 3.5 worker-seconds is a massive advantage of between 2 and 3 minerals, at best.
This is why, even though extractor tricks seem to gain an advantage in workerseconds if you just compare the drone timings, they actually end up behind in minerals over the entire time range.
Because 10 minerals (what we effectively have to make up in order to be even) is worth at least 13 workerseconds, and the drone building the extractor loses another 3 seconds to the extractor build/cancel, we'd have to be 16 workerseconds ahead at the end of the sequence just to break even, and 10-OL-Trick is only 11 seconds ahead after 15 workers, and 10-Trick-OL is only 8.5 seconds ahead of 9OL, so neither one gains enough worker-seconds to break-even on the cost of 10 minerals for canceling the extractor.
|
Also, just for shits and giggles, I threw in an 8OL build (nobody is doing this, but I was curious).
It is behind 9OL at every point, with a 20-second disadvantage on worker 9, never catches up from there, and ends up being 25 minerals short at the 2:15 mark.
|
Updated the original post with timings and conclusions on doing the extractor trick after the overlord, as well as the ever-so-unpopular 8OL.
|
I agree with LaLuSh. In the OP you compare 10ol ET to both 10 et 11ol and the 9ol. But if your comparing it with just the 10et 11ol, you will find that the overlord before 11th drone is more economic and doesn't waste larvae spawn time. It may be inferior to 9ol but it should be noted as being better than overlord on 11.
|
Thanks for running the tests Skrag. I think I'm going to 9OL in the future.
I always intuitively thought the differences between 9OL and 10-OL-T were negligible, so I didn't really care. But it's true that the execution of 10-OL-T is harder, and there's alot more elements where you can screw up. Especially with the timing of the extractor trick.
Probably settled this discussion once and for all.
|
The one thing you forgot in your tests was a scout (unless you didn't and just haven't mentioned it?), which is sort of crucial. The reason I do 10-OL-ET is because I can use the drone I use for the trick to scout (smaller maps), and that drone will be replaced within a matter of seconds. I have seen several players mention that they do the trick for the same reason.
It's all well and good to know that 9-OL is the most economic when you DON'T scout... But on smaller maps you will always scout before 15 (Steppes for example).
But it is useful information for larger maps where you may not send a scout till 14+, 9-OL being the best in these cases.
|
On July 05 2010 13:46 Swede wrote:
The one thing you forgot in your tests was a scout (unless you didn't and just haven't mentioned it?), which is sort of crucial. The reason I do 10-OL-ET is because I can use the drone I use for the trick to scout (smaller maps), and that drone will be replaced within a matter of seconds. I have seen several players mention that they do the trick for the same reason.
It's all well and good to know that 9-OL is the most economic when you DON'T scout... But on smaller maps you will always scout before 15 (Steppes for example).
But it is useful information for larger maps where you may not send a scout till 14+, 9-OL being the best in these cases.
Well, the point is that unless you're doing *something* with a worker in the small range where tricking gives you a timing advantage, then you shouldn't do it. I didn't mention scouting, but scouting is definitely "doing something", it's just good to be aware of the cost.
In this case, it is probably better to trick after the overlord rather than before, which you've said you're already doing. However, an super-early scout isn't really that necessary on smaller maps IMO. If the info you can get from a scout doesn't change what you're doing, then you've sent it too early, and a super early scout often isn't going to tell you anything actionable, so it might be worth figuring out if you can delay a little for a better economy, because tricking for a scout will increase the overall cost of the scout by quite a bit.
|
On July 05 2010 13:13 P00RKID wrote:
I agree with LaLuSh. In the OP you compare 10ol ET to both 10 et 11ol and the 9ol. But if your comparing it with just the 10et 11ol, you will find that the overlord before 11th drone is more economic and doesn't waste larvae spawn time. It may be inferior to 9ol but it should be noted as being better than overlord on 11.
I don't believe any of these builds waste any larvae. And yes, the discussion comparing the 2 extractor trick methods does say that tricking after the overlord is more economic than tricking before the overlord.
But the point is that it's still less economic than 9OL, and is inferior to tricking before the overlord during the range that the extractor trick actually gives an advantage. So if you want to act during that period of advantage (before 14 workers), then tricking before the overlord is superior, and if you don't, 9OL is superior. Tricking after the overlord is a mix between the two, but it doesn't seem to be good enough in either situation to actually be worthwhile.
|
Really nice way to do the measurements, I like how you tried to get rid of tester bias.
I guess that pros rarely fail the timing of the ET, but for me the fact that 9OL is so much more consistent, is reason enough to use it. Added bonus for being the most economic
On July 05 2010 13:46 Swede wrote:
The one thing you forgot in your tests was a scout (unless you didn't and just haven't mentioned it?), which is sort of crucial.
Then again, you also get your overlord faster with 9OL
I prefer going for an earlier pool, like 12 or 13 pool. In that case I very rarely scout before placing my pool.
From that point on, I feel your economy becomes dominated more by how hard you drone/save and how few lings you get rather than your exact opening.
|
I like to get fast ling speed so I go 10 extractor WITHOUT canceling then overlord and pull 1 drone to gas and then pool at 13 and pull 2 drones to gas when pool is almost done to make up for the 100 gas and research ling speed then pull drones back to mineral. Can anybody tell me if there's another way to get faster ling speed with a higher econ advantage build? I'd like to try. Just too lazy to test myself haha.
|
On July 05 2010 16:02 eScaper-tsunami wrote:
I like to get fast ling speed so I go 10 extractor WITHOUT canceling then overlord and pull 1 drone to gas and then pool at 13 and pull 2 drones to gas when pool is almost done to make up for the 100 gas and research ling speed then pull drones back to mineral. Can anybody tell me if there's another way to get faster ling speed with a higher econ advantage build? I'd like to try. Just too lazy to test myself haha.
Building your extractor just before your spawning pool, and putting 3 drones on gas as soon as its complete, will get you 100 gas just when the pool is finished.
|
tl;dr version:
I tested timings of 10pool up through 14pool using the various openings, and obviously, the fastest pool is going to come when you build the pool before your overlord. So, if you're going to 12 pool and you want it as fast as possible, double-extractor. If you want to pool on 11 asap, you should single extractor trick before the overlord. Duh. Genius, huh? Well, the timings below should give you an idea of how much time you're gaining and how much it'll cost.
But for 13pool and later, where you're going to have to build at least one overlord, 9OL is a pretty clear winner, meaning that the very small timing windows of advantage for extractor tricks simply aren't big enough to be worthwhile after the 13th drone.
------------------------------------------------
The "oh my god how much free time does Skrag REALLY have for fuck's sake" version:
Another update. The timings given in the OP didn't *exactly* show what was best when, so in a major fit of boredom, I manually tested a ton of different pool timings with different methods. I didn't try to automate this one, because that would have given a disadvantage to the double-extractor trick, but here's the testing methodology I used:
Playing on normal speed, which is slow enough to minimize the chance of me missing timings, and restarting every time I felt like I screwed up even the slightest amount, I tested pool timings from 10 pool all the way to 14 pool. For every test, I started a drone, and split the first 6 workers in exactly the same way, which looked identical to the computer's "perfect" split on the 6oclock position on lost temple, which is 3 on each close patch, where the extras peel off to the sides extremely nicely. I then built drones up to the specified number, rallying exactly the same way every time, cherry picking up to 10 drones, to make sure two got on the close patches, and then just leaving the rally point on the center. I tried to have the drone in position just when minerals hit 200 in all cases, and once the pool was started, continued building drones with every larva that popped.
I recorded the finish time of the key drone (drone 10 for 10pool, 11 for 11pool, etc), the start and finish times of the spawning pool, and the supply and mineral count at each point.
Results:
SPAWNING POOL TIMINGS
10Pool
Drone 10 57.25 10/10 110
72.3125 9/10 15
137.313 14/18 75
13 drones + 1 drone 10% done
Overpool (10 drones, overlord, then pool)
Drone 10 57.25 10/10 10
86.0625 9/18 5
151.0625 15/18 165
14 drones + 1 drone 50%
11Pool extractor trick before overlord
Drone 11 69.9375 11/10 129
80.125 10/10 4
145.125 14/18 154
13 drones + 1 drone 50%
11Pool 9OL
Drone 11 90.375 11/18 170
95.5 10/18 20
160.5 16/18 215
15 drones + 1 drone 75%
12Pool w/double extractor
Drone 12 85.3125 12/10 168
85.3125 11/10 8
154.375 15/18 198
14 drones + 1 drone 75%
12Pool w/double extractor OL first
Drone 12 85.375 12/10 68
102.125 11/18 3
167.125 17/18 253
16 drones + 1 drone 75%
12Pool 9OL
Drone 12 90.3125 12/18 120
101.813 11/18 10
166.813 17/18 270
16 drones + 1 drone 75%
13Pool 9OL
Drone 13 96.875 13/18 120
107.5 12/18 5
172.5 18/18 315
17 drones + 1 drone 25%
13Pool double extractor
Drone 13 117.25 13/18 23
108.688 12/18 18
173.688 18/18 273
16 drones + 1 drone 75% + 1 drone 25%
13Pool single extractor before overlord
Drone 13 107.188 13/18 4
107.25 12/18 4
172.25 18/18 284
17 drones + 1 drone 25%
13Pool single extractor after overlord
Drone 13 98.5625 13/18 134
109.063 12/18 19
174.063 18/18 299
17 drones + 1 drone 25%
14Pool 9OL
Drone 14 110.625 14/18 180
112.813 13/18 10
177.813 18/18 360
17 drones + 1 drone 50%
14Pool double extractor trick
115.063 13/18 13
Drone 14 117.125 13/18 28
180.063 18/18 353
17 drones + 1 drone 50%
14Pool single extractor before overlord
Drone 14 110.125 14/18 179
112.625 13/18 4
177.625 18/18 354
17 drones + 1 drone 50%
14Pool single extractor after overlord
Drone 14 110.063 14/18 179
113.438 13/18 4
178.438 18/18 359
17 drones + 1 drone 50%
Obviously, builds that got the requisite number of drones first, before an overlord (single extractor trick for 11 pool, and double for 12 pool), get the pool up quite a bit faster. Some comparisons:
10 POOL
Never never never ever ever overpool. Extractor trick 11-pool gets the pool out 6 seconds *faster*, and wastes almost exactly the same amount of larvae.
11 POOL
Unsurprisingly, a single extractor trick gets the pool out 15 seconds faster than 9OL to 11 drones. (which is actually kinda hard to remember to do right, and probably not smart, because 9OL has you starting drones 11 and 12 at the exact same time normally, when the overlord pops) If you're 11 pooling, speed is probably all that matters, and the cost doesn't appear to be too severe.
12 POOL
12 pooling with a double extractor trick beats getting to 12 drones via 9OL by just over 12 seconds. But it does waste larvae, and therefore your drone count will be lower, and your overall minerals lower by a fair amount. If getting a queen or zerglings 12 seconds faster is important, and worth the cost in delayed economy, double-extractor then pool then overlord is the way to go here.
I also tested a 12pool build I've seen TheLittleOne do a fair amount, which is double-extractor, overlord, then pool. I'm not sure why he does this, because this build is inferior to 12pooling via 9OL in every way. The pool comes about the same time (half a second slower actually in my tests, but as I already mentioned, just the way the drones move around the hatchery can make at least that much of a difference), but is about 20 minerals short. Zero speed benefit, for a decent cost.
13 POOL
I tested a *lot* of ways with a 13 pool, because this is the area where it's somewhat unclear which is best, because this is the time range where extractor tricks are starting to lose their advantage over 9OL.
9OL was clearly the winner economically, ending up with the best drone and mineral count, as well as being faster than everything except a single extractor trick before overlord, which came in only *slightly* faster (half a second), and had a clear cost associated with it.
14 POOL
Again, the single extractor trick before overlord came in *slightly* faster than 9OL (a fraction of a second again), but with a cost of 10-15 minerals, more than I personally would be willing to pay for a fraction of a second.
So again, 9OL is a pretty clear winner.
Just for reference sake, I measured 6pool through 9pool using the same methods, except that I only cared about the pool finish time. If you're pooling this early, you're doing it to rush, and you already know you're various levels of screwed economically, so the finish time is the only thing that matters.
6POOL: 106.75
7POOL: 114.125
8POOL: 122.438
9POOL: 130.188
|
Always done 9 OL because I heard it was good lol, but thanks for proving it
|
Thanks for doing such comprehensive testing on this. I just got the beta two days ago and was completely lost what opening was the best. 9OL from now on for me
|
This is impressive. Skrag I don't see where you account for larvae production and availability over time. Would there be a quick way to run these tests again and find when the larvae is produced and exhausted in the sdXX FDXX OLxx method? Of course we can infer the larvae used from your data, but it wouldn't account for the larvae production or availability at any given times.
If you want to distribute the work load you could share the map with TL too as someone else suggested earlier.
|
So,
From all the builds i like:
12Pool w/double extractor
13Pool 9OL
14Pool 9OL
I think 13 pool might catch up economically but i am not sure about 12Pool
Can you please test this builds for the same time and compare? or maybe until you build your expansion. like 12Pool Queen Hatch
let`s say we want as strong build so after pool you will star queen and build drones
I wonder ho much does 12Pool costs you maybe we can always do 12Pool and if we scout no rush push drones to keep up or lings to kill him
|
On July 14 2010 03:58 projectdp wrote:
This is impressive. Skrag I don't see where you account for larvae production and availability over time. Would there be a quick way to run these tests again and find when the larvae is produced and exhausted in the sdXX FDXX OLxx method? Of course we can infer the larvae used from your data, but it wouldn't account for the larvae production or availability at any given times.
If you want to distribute the work load you could share the map with TL too as someone else suggested earlier.
For the pool timings, where I 13pooled all the various ways for example, no build was ever left with any available larvae at the recorded end time, although many of them would have had one pop out shortly after the pool finished. 9OL never wastes any larvae, no matter where you build the pool (unless for some reason you wanted to 9OL 9Pool, but that would be silly), but any build that puts pool before overlord is going to waste some.
None of the builds ever wasted larvae after the overlord finished, so it probably wouldn't be that big of a deal to go through 10-12pools and see what larvae get wasted where.
If you're looking at the original list, the only build that wasted any larvae was 10OL, and you could just look at the drone start times to see when larvae pop, since after the first few, every larva was used as soon as it popped out. Of course, I didn't build a pool at any point in the original list either, since the whole point was to try to figure out what periods of time the various starts had advantages over each other.
As for distributing the map, it's just lost temple with like 2 triggers to print out text messages when things finish.
|
Here's the original thread for those who are interested:
http://www.teamliquid.net/forum/viewmessage.php?topic_id=118754
You can show logically that 9ol is the best without needing too many calculations.
On April 08 2010 22:04 Klive5ive wrote:
Ok let's sort this out then.
Here's my measurements from replays, feel free to check this:
With 9 drones it takes 14 tics to mine the extra 100 minerals needed to get an overlord on 9.
Drones mine at 5 minerals every 7 tics
Extractor trick costs effectively 10 minerals (it's actually 6 but you can't spend the extra 4 unless you cancel another building).
10o
This build wastes larvae time, setting all your future drones back. The result is less minerals in the long run.
However at the 10 larvae stage this build has the most minerals. Therefore it will get a 10 pool down faster.
If you plan a fast pool or something crazy like a 12hatch this would be the correct build.
10o+e Vs 9o
Neither of these builds waste larvae time.
9o gets it's 10th drone 100 minerals later. Effectively that's about 14 tics. 1 drone would mine 10 minerals in that time.
The 10o+e loses an effective 10 minerals on the extrick so it's immediately equal. Factor in the lost mining time on the extrick drone and the later 12th drone and you can see this build is worse.
Conclusion:
9o is the strongest economic opening for Zerg. This should always be done before the popular 15pool 16hatch.
|
Other people concluded in that thread that 10OL is still good for overpool.
I just want to make it abundantly clear that while overpool might have been a viable option in BW, it is NEVER a good idea in this game, because extractor tricking for an 11 pool gets the pool *faster* than overpool, is more economic, and I'm fairly sure it wastes fewer larvae too.
Which, of course, means that 10OL is never a good option under any circumstance. If you happen to forget to build your OL on 9, go ahead and build it on 10 and then extractor trick. That will put you as close to 9OL as you can be.
|
Thanks for all this. I assumed that 10 OL/ET was done to optimize the larva usage (when I do 9 OL I find I have 3 larvae for about a second or so, but it seems that's because I'm not quite as quick as the computer AI). Regardless, great guideline for the various build orders based on when the spawning pool comes out, I've been looking for this kind of breakdown!
|
I have tested my builds on the slowest speed (so i can manualy do everything almost as fast as the computer can). I vastly perfer the 10-T-OL build, the reason is that I like to scout with my 11th or my 13th drone, depending on if its vr terran 11 drone, or protoss/zerg 13 drone.
Terran is an 11 drone, so I can do more damage to the terran eco by forcing him to pull 2 more scvs early to stop my drone from killing the scv building the rax, and I feel that 2 scvs pulled off minerals is more worth my one drone being a little earlier. As well as having a quick gas steal (before he can build 2 refineries) after harrassing the rax is over.
Zerg and protoss dont change all that much till the middle game usualy, so i can send a slightly latter scout, but still early enough to catch any silly 6/8/11 pools, as well as getting a scout to protoss quick enough to tell me if I can safely FE at 16 or so (1 gateway in far positions is clear indication that a FE is good before pool, 1 gateway in close positions and you can do it very quickly after the pool, otherwise you need to stay on one base for a while).
So all in all given that a 10-T-OL might be 1 mineral worse on avr if you dont scout early (and my testing showed it EXACTLY the same not 1 mineral worse then a 9 OL). I will still choose the 10-T-OL VASTLY over a 9 OL if your doing a early scout.
Clearly this isnt as good an idea on something like scrap station or LT if you scout him in the close air distance location. The early drone scout isnt needed as much (esp against zerg or protoss), and the 9OL is probably better. Although an early drone scout can still harrass vr terran which can be good enough to still send it.
|
9 OL seems like you have less drones stock up larva early on, and 10 lord with extractor trick uses the larva the most efficiently. the mineral difference between 9 and it are negligible so you shouldn't use 9 then.
Did op even take larva into consideration?
|
With 9ol you get the 12th and 13th drones faster than the 10ol et, and 9ol doesn't have any lost larva like I think your suggesting CHarlieMurphy.
|
This study needs to take into account the worker scout.
I would imagine that having an early scout makes a much larger economic impact on 9 OL as opposed to 10 OL trick as you lose 1 of 10 drones with 9 OL and you lose 1 of 11 drones with 10 OL.
|
On July 21 2010 13:43 obsid wrote:
So all in all given that a 10-T-OL might be 1 mineral worse on avr if you dont scout early (and my testing showed it EXACTLY the same not 1 mineral worse then a 9 OL). I will still choose the 10-T-OL VASTLY over a 9 OL if your doing a early scout.
1 mineral isn't 1 mineral, it's 5. You can't spend 4 minerals at any point in the game. Also, how many times did you test on slow? Because I tested far many more times than I'd care to count, both letting the AI do the work, and doing it myself on normal speed (slow was too slow since I was letting them run longer, and I could easily do everything I needed to, including a perfect 3/3 split, on normal), and 10-T-OL was very consistently 10 minerals behind.
But even if it's just 1 mineral, behind is behind is behind, so unless you're doing something very specific with that 11th drone, then why pay a price to do a more risky build (misclicks can cost you a LOT here)?
As for scouting, if you insist on scouting at 11, and that exact timing is useful and meaningful, then sure, extractor trick. But 9OL will get the 13th drone almost 10 seconds before 10-T-OL, so if you still choose it vastly for that reason, you're doing something wrong IMO.
|
On July 21 2010 14:06 CharlieMurphy wrote:
9 OL seems like you have less drones stock up larva early on, and 10 lord with extractor trick uses the larva the most efficiently. the mineral difference between 9 and it are negligible so you shouldn't use 9 then.
Did op even take larva into consideration?
A single extractor trick doesn't waste any larvae spawn time. 9OL wastes less than a second of larvae spawn time, while you're waiting for minerals after building your overlord. Every other build I've tried, including the double-extractor trick overlord that TLO claims to do for larvae management, wastes quite a bit more than that.
Is less than a second worth of larvae worth 10 minerals, which represents about 15 seconds worth of worker mining time?
|
So... I have a fairly unique BO and am wondering how the best way to go about testing the economic impact would be.
I do:
Extracto-drone 11
Spawning pool
OL
extracto-drone 11
12 & 13 when OL pops
14 when larva pops
Queen/ling (lings do my scouting around the 3:30 mark)
Ling
Extractor
Drone
OL
Spawn larva when queen pops
2nd queen when OL pops
Power drone to 26
Yes, a lot of drawn out steps, but very cheese resistant, allows for some early aggression, and getting spawn larva so early let's me catch up in drone numbers to more standard builds.
I know I have lost minerals doing this build, but how many?
|
this is very helpful, thank you for the information.
I'm going to start using 9ol, and I know I'm going to mess it up and go for 10 ol-extractor trick and then try to switch back and mess myself up big time with my opener >.>
But that will only last for the first few days, or if I'm really drunk.
Thanks for the research, much appreciated.
Repped
|
I have to agree that your testing proves conclusively that 9ol is superior to 10ol purely in terms of minerals mined for the rest of the game. It doesn't however take into account when you get those minerals and the fact that at certain points in the early stages zerg really can't spend all their minerals even if they have them.
For example, from my own testing doing double ET, 12 ol, 12 pool you will be able to do a 12 pool without wasting larvae, which to my knowledge is the fastest possible pool without permanent economic damage(basically, as opposed to other races zerg only loses cumulative income if they are larvae blocked, cutting drones will not permanently delay other drones unless it wastes larvae).
This means that your queen spawns roughly 10-15 seconds earlier than other "early" pool timings that do not waste larvae(13-14). This can help tremendously against certain ultra fast rushes, such as 6 rax proxy reaper or 6 pool. Not that these builds are hard to stop regardless, but it makes stopping them a lot less of a hassle. You can from there either get a really quick ling force, or power drones like any of the "economic" builds thanks to that early larvae pop.
So while your testing is very much appreciated, it doesn't paint the entire picture as well as you make it sound like(although for most builds, it's very informative). Especially since you say you can't get the AI to properly double extractor trick.
Edit: Posted before fine tuning my post due to stress, so some things may have changed if you read this post just as I originally posted it. Sorry about that.
|
I checked BO's on last days of phase2 from several best known zergs and all seemed to be doing 9OL now. I did think until then that gastrick (after OL ofc) was slightly ahead, but have started switching to 9OL now. Last few days I tried to remember this, but often noticing my brains autopiloted myself to always 10/10 before OL  Oh well, training training required.
|
On July 22 2010 21:37 Alsn wrote:
I have to agree that your testing proves conclusively that 9ol is superior to 10ol purely in terms of minerals mined for the rest of the game. It doesn't however take into account when you get those minerals and the fact that at certain points in the early stages zerg really can't spend all their minerals even if they have them.
For example, from my own testing doing double ET, 12 ol, 12 pool you will be able to do a 12 pool without wasting larvae, which to my knowledge is the fastest possible pool without permanent economic damage(basically, as opposed to other races zerg only loses cumulative income if they are larvae blocked, cutting drones will not permanently delay other drones unless it wastes larvae).
Check out the spawning pool timings on the second page. I tested various ways of pooling at different supply amounts (6 pool all the way through 14 pool), all manually because I couldn't get the AI to double-extractor efficiently, but it was all done on normal speed instead of faster so that there were fewer opportunities to screw up, and doing each build numerous times.
I also did separate testing on how much larvae time was wasted by each build. As I already mentioned, 9OL wastes a very small amount of time, where single extractor tricks don't waste any. But a double extractor trick into overlord then pool does waste more larvae time than 9OL by a few seconds.
In addition, 12 pooling with 9OL gets the pool slightly *faster* than double-trick-OL (and by slightly, I mean *very* slightly, because it's only by about 3/4 of a second on average), and is better economically.
So if you're 12 pooling but building an overlord first, you should *never* double extractor trick.
The only reason you should ever double extractor trick IMO is if you're going to build the pool first then the overlord. Otherwise, you're sacrificing minerals for absolutely zero benefit.
|
On July 22 2010 21:45 Ouga wrote:I checked BO's on last days of phase2 from several best known zergs and all seemed to be doing 9OL now. I did think until then that gastrick (after OL ofc) was slightly ahead, but have started switching to 9OL now. Last few days I tried to remember this, but often noticing my brains autopiloted myself to always 10/10 before OL Oh well, training training required.
Heh, yeah, I still find myself forgetting to OL at 9 sometimes. The good news is that if you build the overlord on 10, and then extractor trick, you only end up very slightly behind 9OL, so its a good fallback if you forget the OL.
|
You should always do 9OL unless your BO requires you to do extractor tricks (very fast exp/pool)
|
Hi, you got the replays/maps? That would be cool to see firsthand.
|
So I've always built my OL when I'm at 10worker count, done the ET, and when the OL pops I make my 2 drones right away, then another when the larva is out.
So if I'm reading this correctly, the superior build to this is actually to get to 9 supply, save to 100minerals for the OL,(while still at 9 supply), THEN get my 10th drone?is there time to do the ET before th OL pops, still giving you 11/10? then the OL pops, and you'll have 100 minerals ready to instantly make another 2 drones?
so...forgive my for being baked and unable to comprehend, and dumb this shit down for me
I go
9 OL
drone for 10
ET
OL pops
drone,drone
then shortly after there's a third larva for another drone, but the third larva is quicker than my 10 OL, ET, double drone when OL pops, and another drone when the 3rd larva appears?
I have this correct?
9OL is superior build?
sorry i'm retarded right now, I'm just making sure I'm understanding this
|
nm I just now saw the post with the link and quote at the top of page three
DERP
thanks ^^ lol
|
"Computer players do a pretty dang near-perfect 3/3 split (a lot better than any 3/3 split I've ever seen in a replay)"
Have you ever looked at really early APM for the computer? I've seen them hit 400 before, it's insane.
|
vvvvvvvvvvvvvv This is why I don't come here vvvvvvvvvvvvvvv
Honestly. Thanks for informing us that the trick is a failure. But it didn't take reading all that.
On July 04 2010 09:37 Skrag wrote:I know this is an old tired topic that comes up all the time and never really gets answered with any sense of solid conclusion. I've taken a slightly different tack in attacking the problem that I hope will put it to rest once and for all. ************ADDED SOMEWHERE ON PAGE 2************ Timings on 6pool through 14pool, using the various orders. Supershort results: 12pool with double extractor before overlord, 11pool with single extractor before overlord, but 13+pool should always be with 9OL. Also, never ever ever overpool. 11pool with extractor trick is superior in every possible way. ******************************************************************* tl;dr version: ----------------------- I used the AI to test the various builds in a very consistent manner, to try to answer the question once and for all. The 9 overlord is the best economic build, period. Minerals at 2:15: 370 10 overlord can be used for 10pool and overpool, but is inferior to 9OL in every other possible respect to the tune of 10 minerals, and loses the small window of advantage it gets by building the 10th drone sooner by the time 10OL finishes its 11th worker. Minerals at 2:15: 355 A single extractor trick doing the trick before building the overlord (11 drones then overlord) gives a small time window of about a 15 mineral advantage, between workers 11 and 14, but there's a cost, of around 5-10 minerals. Best case, you *might* be able to break even when compared to 9OL, but most of the time you'll lose 5-10 minerals. Minerals at 2:15: 364 Single extractor trick after the overlord (10 drones, overlord, trick to 11 drones) is inferior to 10-T-OL in the 11-13 worker range, and is inferior to 9OL in the 14-15 range, so I honestly see no reason to ever trick this way, UNLESS you meant to 9OL, but forgot the overlord. In that case, this is your best way to try to catch back up, because the cost is 0-5 minerals instead of 5-10. Minerals at 2:15: 369 The double extractor trick provides such a small advantage over the single extractor trick during even a smaller window that it should probably only be used exactly to 12 pool, and even then, there is an extra cost over the single extractor trick. It probably compares to the single extractor trick much like 10OL compares to 9OL, meaning that if you're doing something exactly on drone 12, you might use it, otherwise, the single trick is probably strictly superior. Minerals at 2:15: 353 (imprecise, but again, likely strictly inferior to single trick except for fastest 12pool) ----------------------- BTW, are you concerned yet that the tl;dr version is that long? Full version ------------ In an attempt to put this question to rest once and for all, I decided to set up a computer player to perform everything at as close to the exact correct timing as possible. There are many advantages to testing this way rather than testing on your own, the primary reason being that the computer can be much closer to perfect than you. Computer players do a pretty dang near-perfect 3/3 split (a lot better than any 3/3 split I've ever seen in a replay), and with some manual rally-point management as well as spending a *lot* of time trying to figure out exactly when the best time to trigger the extractor trick was (and how to do it to begin with in the map editor, which definitely was not easy), I was able to get truly consistent timings. What I ended up with was the as near-perfectly as possible for the first 15 drones, and certainly performing much better than any player could, especially performing the extractor trick, where the AI is able to perform a number of completely separate actions at the exact same time. (build a drone, cancel an extractor, put a worker back on minerals) Note that there was actually some amount of variation here. This shocked me at first, because the timing numbers would come through and match EXACTLY, even for entire sequences up to 15 drones, down to the one-sixteenth of a second that it looks like the timer resolution might be. Which makes perfect sense, because the computer player is going to make the exact same decision at the exact same time *every* time. But there turned out to be quite a lot of variation. I'd only see 2-3 patterns of timing, and once the pattern was established, it would match exactly, but there was as much as a full second's difference in later drone start timings. I puzzled over this for a really long time, and then as I was watching a sequence and saw a cocoon pop up, it hit me. Larva movement is random, and the start position of your larva determines how far the drone has to move once it's been created. What's surprising is that this one little thing can make a rather large difference. Since there was quite a bit of variation, I ran through a number of times, and picked out the most consistent timings for each variation. Anyway, the point here is that although this test was not absolutely consistent (even though I thought when I started that it should be), it's definitely going to be more consistent than testing by hand. With no further ado, here are the results. sdXX is when drone number XX was started, and FDXX is when it completed building. Each point records the exact time, as well as the number of minerals that the player had at that point. sovl and FOVL represent the overlord, and EXTR and CANC are the extractor when doing an extractor trick. I also printed the number of minerals at 2:15 (135 seconds), by which time all 3 builds were complete. Time measurements are all in gametime, not real time. 9OL sd07 0 0 sd08 15.9375 0 FD07 17 10 sd09 26.5625 5 FD08 32.9375 30 FD09 43.5625 80 sovl 46.625 0 sd10 54.375 0 FD10 71.375 105 FOVL 71.625 60 sd11 71.625 60 sd12 71.6875 10 sd13 78.125 0 FD11 88.625 70 FD12 88.6875 70 sd14 91.875 30 FD13 95.125 65 sd15 106.938 110 FD14 108.875 125 FD15 123.938 270 At 2:15, this build had 370 minerals 10OL sd07 0 0 sd08 15.9375 0 FD07 17 10 sd09 26.5625 5 FD08 32.9375 30 sd10 38.9375 0 FD09 43.5625 30 sovl 55 10 FD10 55.9375 10 FOVL 80 120 sd11 80 120 sd12 80.0625 70 sd13 80.125 20 sd14 95.25 75 FD11 97 85 FD12 97.0625 85 FD13 97.125 85 sd15 110.813 140 FD14 112.25 150 FD15 127.813 290 At 2:15, this build had 355 minerals There is exactly *one* point of time where 10OL is ahead of 9OL. From 55.9375 seconds until 71.375 seconds, 10OL has one more worker running, for a gain of about 15 worker seconds. But 9OL finishes drones 11 and 12 almost 10 seconds faster than 10OL, so this 15 second advantage begins to be lost at 88.625, and is completely gone somewhere around 96 seconds in. After that point, 9OL is ahead of 10OL in every way. It finishes the 15th drone 4 seconds faster, and is 15 minerals ahead at the 2:15 mark. As a result of this, the only situations where it is not better to 9OL are when you are 10-pooling or overpooling. SINGLE EXTRACTOR TRICK BEFORE OVERLORD (or 10-T-OL) sd07 0 0 sd08 15.9375 0 FD07 17 10 sd09 26.5625 5 FD08 32.9375 30 sd10 38.9375 0 FD09 43.5625 30 EXTR 49.6875 45 sd11 50.3125 0 CANC 50.3125 FD10 55.9375 49 sovl 64.375 4 FD11 67.3125 29 FOVL 89.375 139 sd12 89.375 139 sd13 89.4375 89 sd14 91.5 44 FD12 106.375 159 FD13 106.438 159 sd15 106.563 114 FD14 108.5 129 FD15 123.563 249 At 2:15, this build had 364 minerals This is a little more difficult to figure, so lets start with the easiest result: at 2:15, this build was exactly 6 minerals behind the 9OL build. That amount happens to be exactly what you lose from the extractor cancel, although technically it works out that you're 10 minerals behind, not 6, since there's nothing you can ever spend 4 minerals on. Now the hard part: figuring out exactly where this has an advantage over 9OL. Drone 10 finishes at 55.9375, and drone 11 finishes at 67.3125. So we gain 15.43 worker seconds on drone 10, and 21.31 worker seconds on drone 11, for a total of 36.75 worker seconds. At 88.6875 seconds in, 9OL finishes worker 12 ahead of the extractor trick, gaining 17.69 worker seconds back on drone 12, and also finished worker 13 11.31 seconds ahead, which leaves a 7.75 worker advantage starting at 106 in seconds unaccounted for, which only builds on drones 14 and 15, becase 9OL finished them both about a third of a second later than the extractor trick. Call it 8.5 seconds for good measure. So the extractor trick *should* be 8.5 worker seconds worth of mining ahead of 9OL, but ended up being at exactly the same point mineral-wise, minus what was spent on the extractor. Some of this is the loss of time for the worker building the extractor. This drone building the extractor is the one closest to the extractor, and it drops off minerals at 47 seconds, builds the extractor at 49, and touches the minerals after cancelling at 52 seconds, basically taking around 5 seconds to get from base to mineral patch, a trip that normally takes about 2 seconds. So that accounts for 3 seconds, but we should still be 5 seconds ahead. Well, it turns out that 5 seconds is less than the amount of time it takes for one worker to collect and return 5 minerals. So for a lot of extra complexity, we've made a very small gain in time, but not gained enough time to overcome the loss we take from cancelling. Again, there is quite a bit of variance here. I saw 2:15 marks from 364 to 374, with 364 being the most common. There is a period here though, where this build has a 36.75 worker second advantage though, which works out to be somewhere around 25 minerals (or 15, after taking out the extractor trick penalty). If having an extra 15 minerals between the 11th and 14th drone speeds things up enough that it's worth 5-10 minerals later, then the extractor trick is worthwhile. So 12-14 pools could extractor trick rather than 9OL (but could also choose to 9OL if the 5-10 minerals aren't worth the slightly increase pool speed), everything else should 9OL. If you perform the extractor trick well, that is. SINGLE EXTRACTOR TRICK AFTER OVERLORD (or 10-OL-T) sd07 0 0 sd08 15.9375 0 FD07 17 10 sd09 26.5625 5 FD08 32.9375 30 sd10 38.9375 0 FD09 43.5625 30 sovl 55 5 FD10 55.9375 10 EXTR 65.6875 45 sd11 66.1875 0 CANC 66.1875 0 FOVL 80 64 sd12 80 64 sd13 80.0625 14 FD11 83.1875 29 sd14 91.1875 34 FD12 97 89 FD13 97.0625 89 sd15 106.313 99 FD14 108.188 114 FD15 123.313 254 369 at 2:15 Over a number of runs, the final mineral result bounced pretty consistently from 369-374, which gives an overall cost of 0-5 minerals, which is better than 10-trick-OL by 5 minerals. If we compare this build to 10-T-OL, we can see the following: Drone 11 comes out 15.875 seconds later, putting us behind by that many worker-seconds at 83.1874 seconds. We start to gain this back on drones 12 and 13, which are each faster by 9.375 seconds, and that both finish at 97 seconds. So at 104.9375 seconds, this build has caught back up to 10-T-OL, and at 106.375, we're ahead by 2.875 worker-seconds. We gain another .312 seconds on drone 14, and .25 seconds on drone 15, giving us a grand total of 3.4375 worker seconds advantage over 10-T-OL. BUT.... We already know that both of these builds are behind, or at best equal to, 9OL, which is far easier to execute. The only reason to doing the extractor trick at all is to get an advantage at some point in the 10-15 worker range in order to get some building out faster. 10-OL-T has a smaller advantage over 9OL than 10-T-OL in the 10-13 range, and has a very negligible advantage over 10-T-OL at around the time we hit 15 supply. So if we want to build in the 12-14 range, 10-T-OL clearly has an advantage in that time range, and if we're waiting til 15 supply (the only time 10-OL-T has an advantage over 10-T-OL), then we should just 9OL. At the end of all that, 10-OL-T seems inferior to 10-T-OL in the 11-13 range, and inferior to 9OL in the 14-15 range, so there doesn't seem to be any valid reason to ever do this build. The only reason I can think of would be to be less behind if you meant to 9OL but forgot the overlord. In that case, tricking after the overlord will put you less behind than tricking before the overlord. DOUBLE EXTRACTOR TRICK The AI does not have enough flexibility to allow it to do a double extractor trick optimally. Specifically, I wasn't able to tell specific workers exactly where to build, and it makes poor decisions about when and where to take extractors and which workers to use in this case. What I can do though is extrapolate some numbers based on when the extractor trick would start, and the fact that we know it takes about 3 seconds of time away from the workers building the extractors. We would be starting the double trick at around 55 seconds (the time when 10OL starts an overlord). This has drones 11 and 12 finishing at 72 seconds, giving us a 33 worker-second advantage for these two drones over 9OL. We lose some of the worker 11 advantage we got from 1 extractor trick by delaying though. We get 33 seconds for 11 and 12, but we were getting 21 seconds on just 11 before, so we gain 12 more worker seconds. 12 worker seconds is worth somewhere around 7-8 minerals, and it cost us 5 to do the second extractor trick. So we've basically gained 2-3 minerals worth of worker mining time over a single extractor trick, and that's only if we do things completely perfectly, which is a lot harder to do with a double cancel. In addition, drones 13-15 will be even more delayed, giving 9OL more of a chance to catch up or even pass us. I would guess that the only time you should even consider doing a double-extractor trick is if you're going to double-trick, and build your pool on 12 before your overlord. The advantage simply isn't big enough compared to the double-extractor to make anything else worthwhile. I did let the AI run a double-extractor trick, but it was fairly sub-optimal, wasting somewhere around 5-10 droneseconds due to poor choices when building the extractors. The tests that I did run though, showed it finishing drone 15 later (by many seconds) than the other builds, and showed it finishing with 343-353 minerals at 2:15, or 10-20 minerals behind the single trick. I would guess that an optimal build would bring that closer, but drones 13, 14, and 15 are all going to be later in a double extractor trick than a single, and I don't think the time gained from worker offsets that. KICKS AND GIGGLES: 8 OVERLORD Just for grins, I plugged in an 8OL build. Posting timings really isn't necessary, because this build is behind 9OL at every point, and ends up with a 25 mineral deficit. Basically, never build your overlord before 9. Ever.
|
What about overpool:
+ Show Spoiler +10 Overlord,
10 Pool,
13 Queen
or trick overpool:
+ Show Spoiler +10 Trick
11 Overlord
11 Pool
14 Queen
..with a fast Queen and (optionally) a pair of Lings build with the Queen?
I feel that early Queen is boosting your economy quite a bit, especially if you're on one base or expanding at 20 or so. Really doesn't feel far behind compared to 9 OL in range 15-25. Anyone willing to verify that?
|
On October 29 2010 19:29 slith wrote:What about overpool: + Show Spoiler +10 Overlord,
10 Pool,
13 Queen or trick overpool: + Show Spoiler +10 Trick
11 Overlord
11 Pool
14 Queen ..with a fast Queen and (optionally) a pair of Lings build with the Queen? I feel that early Queen is boosting your economy quite a bit, especially if you're on one base or expanding at 20 or so. Really doesn't feel far behind compared to 9 OL in range 15-25. Anyone willing to verify that?
I've done quite a bit of testing on this, and it turns out the early queen doesn't actually give your economy a boost, because you've sacrificed so much for the early queen. And 11overpool is *always* better than 10overpool btw. It's ahead at every point.
What the early queen does do is allow you to start to catch back up after the first larvae spawn though, so you don't end up very far behind overall. You are still pretty significantly behind up to around 30 supply though.
|
Awesome results. Thank you very much. I've been 10 overlording for a while now, and occassional extractor trick thingy, but now i know that 9 overlord is best i will do it from now on. Cheers.
|
Great research, but leaves me with a question very pertinent to me atm.
I have a 13 pool build I am using, so according to you 10-T-OL 11OL is the way to go to get my pool slightly faster without much eco hit.
But, how does 10-T-OL 11 OL compare to 10-T-OL-TT 11 OL (trick into lord, then double trick waiting for lord to pop)?
|
On blisterings sands YABOT
i attempted 9OL 13 drones
at 3:15 i got 985 mins
10OL 3:15 i got 1015 mins
14 drones
9OL 3:25 1065 (3:15 965mins)
10OL 3:25 1100 (3:15 1000 mins
I did this quite a few times and i have the replay, ill test abit more and with more drones etc but i did not test the extractor trick, and in theory 9OL and 10-OL-T have the exact same larvae useage (no larvae spawn time wasted) but the 10-OL-T uses 2 of them much earlier than 9OL with only a 6 mineral difference. Definately interesting whether or not i did it wrong but id like some feedback on this, thanks.
|
On January 12 2012 11:29 whitePuma wrote:
On blisterings sands YABOT
i attempted 9OL 13 drones
at 3:15 i got 985 mins
10OL 3:15 i got 1015 mins
14 drones
9OL 3:25 1065 (3:15 965mins)
10OL 3:25 1100 (3:15 1000 mins
I did this quite a few times and i have the replay, ill test abit more and with more drones etc but i did not test the extractor trick, and in theory 9OL and 10-OL-T have the exact same larvae useage (no larvae spawn time wasted) but the 10-OL-T uses 2 of them much earlier than 9OL with only a 6 mineral difference. Definately interesting whether or not i did it wrong but id like some feedback on this, thanks.
I'd love to see TLer's saying : "Dude this topic is OLD why bring it back" ? When you think about it, if that was the same guy that made another thread about it, it'd be "Dude this topic has already been discussed multiple times, use the search function !"
On topic : I don't think we need to think too much about this build order. It's roughly the same really.
Plus, it's hard to be precise because of the spawn larvaes being far more or less from the minerals...
Did you try @ 20 drones to see if there's still a 30/35 minerals difference ? Otherwise this would bring back to bw with the ovi 8 / 9 stuff :p
|
I'm a big fan of the 9 overlord myself, but I do see certain pros (Suppy and Life come to mind) that just love to do extractor trick. So at the end of the day, it really doesn't seem to matter too much considering pros are doing what is seemingly a less effective opening.
|
|
|
|
|
|
EPT
