|
On May 18 2010 05:00 Azuremen wrote:Show nested quote +On May 18 2010 04:54 Wr3k wrote:On May 18 2010 04:51 Azuremen wrote:On May 18 2010 04:38 Project.SmyD wrote:
im assuming you don't know that you can chrono boost a warpgate while is cooling down for faster cooldown I believe anyone looking at this would be fully aware of that you can Chronoboost a cooldown on a Warpgate. See below... On May 18 2010 04:41 randomnine wrote:
Wait. If it's running the warpgate cooldown and building a unit as a gateway at the same time, won't chronoboosting speed up both those timers?
If so, chronoboosting this would be insanely awesome. I'm going to go test it. This would be nearly OP I feel, as you can also Chrono Warpgate change over as well, so literally half the math involved, plus ignoring the 5 second Warp in phase as I don't think you can cut the Warp in time (please correct me if I am mistaken, I usually Chronoboost my gates after Warp in finishes) On May 18 2010 04:42 Wr3k wrote:
You are wrong, switching the gateway back and forth results in the exact same production rate as leaving it a warpgate. The 30% reduction in build time is not factored into your calculations. I thought about this very early in the beta and tested it. Provided warp gates have not been stealth changed it is still the same. There is absolutely no benefit to swapping your gates. You are mistaken. As has been mentioned several times in this thread, Cooldown is unit Production time minus 10 seconds. See my math above. And before you attempt to tell us we are wrong again, Liquipedia entry on Warpgates. Dare I say liquipedia could be wrong? Until someone tests it and provides proof I am going to go with what I observed through my own testing months ago. You do realize liquipedia is a WIKI right? I'll screenshot the game for you when I get home in 4 hours. Or you can do it yourself right now and look at the cooldown timers displayed in the overlays at the Warpgates. With a Zealot, it is 30%, but with a Stalker/Sentry its around 22%, a HT/DT is around 18%. This could explain your testing results.
If I had access to SC2 at I would. Could someone please post the cooldown times for each unit on the warp gate? Because all empirical testing I have done and seen done indicates that it is 30%, which if it is true, makes this thread pointless.
|
I just tested this with Stalker->Zealot->Stalker and got a difference of 5 seconds total.
Replay.
Whatever the cooldown reductions, it does work. Whether 5 seconds is enough to make it worthwhile is another question.
[edit] So you dont have to pay attention:
First I test Warpgate->Gateway->Warpgate. The first Stalker begins warping in at 6:11, and the final Stalker finishes at 7:07, for a total of 56 seconds.
Then I test Warpgate->Warpgate->Warpgate. The first Stalker begins warping in at 7:44, and the final Stalker finishes at 8:45, for a total of 61 seconds.
|
On May 18 2010 05:15 kzn wrote:I just tested this with Stalker->Zealot->Stalker and got a difference of 5 seconds total. Replay. Whatever the cooldown reductions, it does work. Whether 5 seconds is enough to make it worthwhile is another question. [edit] So you dont have to pay attention: First I test Warpgate->Gateway->Warpgate. The first Stalker begins warping in at 6:11, and the final Stalker finishes at 7:07, for a total of 56 seconds. Then I test Warpgate->Warpgate->Warpgate. The first Stalker begins warping in at 7:44, and the final Stalker finishes at 8:45, for a total of 61 seconds.
Did you take into account the 5 seconds it takes to convert to warp gate each cycle?
It should be Warp Stalker> Convert to gateway> Build stalker> Convert to warp gate
Compared to the time it takes to warp 2 stalkers.
|
You should try with the higher tier units as well. As said previously in the thread, the difference is bigger if you warp in HTs and build a zealot, for example.
I'd try myself but B.net is down in Europe atm.
-Edit- @Wr3k: Obviously, he just did as the OP said that he should do and took the time when it ended and subtracted it with the time it started. Everything is accounted for within.
|
Thats what I did. I only built one Gateway, researched warpgate, and build Stalker->Zealot->Stalker with each methodology.
I'm fairly certain there is no reason to do this if you're just trying to spam out identical units (I could be wrong) but that it works only on the basis of different build times (and therefore cooldowns) for each unit.
[edit] Its probably a second less than 5, since I forgot what to hit to turn Warpgates back into Gateways and had to look.
|
On May 18 2010 04:31 Tozar wrote:
I can see this really making a difference in the early game when I am being pressured by marauders. Warp in a stalker, switch to gateway, make a zealot, switch to warpgate, warp in a stalker...
Agreed, very interesting, and the only practical application I can see for this. It would be most useful in the early game when 150 minerals is more than just a drop in the bucket, and possibly mid-late game if your macro slips during a battle.. but it's pretty easy to keep your macro up during a fight as long as you have enough pylons and production facilities. I will definitely give this a shot next time I'm dealing with early aggression!
|
On May 18 2010 05:22 kzn wrote:
Thats what I did. I only built one Gateway, researched warpgate, and build Stalker->Zealot->Stalker with each methodology.
I'm fairly certain there is no reason to do this if you're just trying to spam out identical units (I could be wrong) but that it works only on the basis of different build times (and therefore cooldowns) for each unit.
[edit] Its probably a second less than 5, since I forgot what to hit to turn Warpgates back into Gateways and had to look.
It makes no sense to test it by making 3 units like that, you are essentially including 1.5 "cycles" you need to be very specific about whether or not you are including the conversion time, and test it in one FULL cycle (one unit from gate one from warp), otherwise you are gaining or losing 5 seconds depending on how you time it.
To test it properly you need to time:
Convert to warpgate > warp stalker > convert to gateway > build stalker
and compare it to:
Warp stalker > Warp Stalker
Look, when my desktop computer is available I will do a comprehensive analysis of this.
|
On May 18 2010 05:32 Wr3k wrote:Show nested quote +On May 18 2010 05:22 kzn wrote:
Thats what I did. I only built one Gateway, researched warpgate, and build Stalker->Zealot->Stalker with each methodology.
I'm fairly certain there is no reason to do this if you're just trying to spam out identical units (I could be wrong) but that it works only on the basis of different build times (and therefore cooldowns) for each unit.
[edit] Its probably a second less than 5, since I forgot what to hit to turn Warpgates back into Gateways and had to look. It makes nosense to test it in a cycle of 3 units like that, you need to be very specific about whether or not you are including the conversion time, and test it in cycles of 2 or more, otherwise you are gaining or losing 5 seconds depending on how you time it. To test it properly you need to time: Convert to warpgate > warp stalker > convert to gateway > build stalker and compare it to: Warp stalker > Warp Stalker
I don't see what you're trying to get at. You can start converting the warpgate back to a gateway while you're warping in the stalker. Also, to me, it makes more testing a cycle of 3 instead of two.
|
Bnet shut down as I was testing, but I can confirm that Liquipedia is wrong about warpgates. Stalkers certainly warp in faster than Liquipedia thinks they should.
You can't easily get the benefit of chronoboost on both timers. When building a stalker and a zealot [using constant chronoboost], the normal stalker warp-in cooldown finishes before you can build the zealot and transform the gateway back to a warpgate. I'll test with templars after the system comes back up, since their basic build time is longer at 55s.
|
8751 Posts
Unfortunately, paying for units in advance (regular Gateway) kinda sucks. When the Warp cooldown finishes, you pay for a new unit, then 3 seconds later you pay for another new unit. It's generally better for builds to pay for things more gradually. Getting slightly faster build times but having to pay for units two at a time is kinda iffy. I'm not sure if any builds can actually use this to be more effective. I guess it's worth looking into but the time sink isn't the only trade-off going on here. I guess some panic situations might use it in early game. For late game, it's definitely better to spend a few hundred extra minerals on Gateways and spend more time looking after your army and economy.
|
On May 18 2010 05:38 Aldehyde wrote:Show nested quote +On May 18 2010 05:32 Wr3k wrote:On May 18 2010 05:22 kzn wrote:
Thats what I did. I only built one Gateway, researched warpgate, and build Stalker->Zealot->Stalker with each methodology.
I'm fairly certain there is no reason to do this if you're just trying to spam out identical units (I could be wrong) but that it works only on the basis of different build times (and therefore cooldowns) for each unit.
[edit] Its probably a second less than 5, since I forgot what to hit to turn Warpgates back into Gateways and had to look. It makes nosense to test it in a cycle of 3 units like that, you need to be very specific about whether or not you are including the conversion time, and test it in cycles of 2 or more, otherwise you are gaining or losing 5 seconds depending on how you time it. To test it properly you need to time: Convert to warpgate > warp stalker > convert to gateway > build stalker and compare it to: Warp stalker > Warp Stalker I don't see what you're trying to get at. You can start converting the warpgate back to a gateway while you're warping in the stalker. Also, to me, it makes more testing a cycle of 3 instead of two.
If something has two phases, why would you ever measure one cycle as 3 phases? It makes each alternating cycle different, warp>gate>warp vs gate>warp>gate.
|
I'm so confused 
|
Changing back and forth when building units with the same cooldown is effectively a 23 second buildtime for every second unit built, no matter the actual unit buildtime.
Using it to mass zealots is obviously useless, since they have a 23 second warp in anyway.
Sentry/Stalker production is effectively increased to 2 per 55 seconds, instead of 2x32=64 seconds.
HT/DT production is effectively increased to 2 per 68 seconds, instead of 2x45=90 seconds.
It is of course most effective if you're warping in high cooldown units and training low cooldown/buildtime units from the gateways. For example: warping in a HT, switching and training a zealot, then switching back will take a total of 46 seconds, while the cooldown of the HT is normally 45 seconds, so you effectively get an extra zealot for 1 second of buildtime.
Here's the gateway and warpgate build data copied directly from the data files from the latest patch:
+ Show Spoiler [Warpgate] + <CAbilWarpTrain id="WarpGateTrain"> <EditorCategories value="Race  rotoss,AbilityorEffectType:Structures"/> <InfoArray index="Train1" Category="Army" Time="5" Unit="Zealot"> <Resource index="Minerals" value="100"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="23"/> <Button DefaultButtonFace="Zealot" State="Restricted"/> </InfoArray> <InfoArray index="Train2" Category="Army" Time="5" Unit="Stalker"> <Resource index="Minerals" value="125"/> <Resource index="Vespene" value="50"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="32"/> <Button DefaultButtonFace="Stalker" State="Restricted" Requirements="HaveCyberneticsCore"/> </InfoArray> <InfoArray index="Train4" Category="Army" Time="5" Unit="HighTemplar"> <Resource index="Minerals" value="50"/> <Resource index="Vespene" value="150"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="45"/> <Button DefaultButtonFace="HighTemplar" State="Restricted" Requirements="HaveTemplarArchives"/> </InfoArray> <InfoArray index="Train5" Category="Army" Time="5" Unit="DarkTemplar"> <Resource index="Minerals" value="125"/> <Resource index="Vespene" value="125"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="45"/> <Button DefaultButtonFace="DarkTemplar" State="Restricted" Requirements="HaveDarkShrine"/> </InfoArray> <InfoArray index="Train6" Category="Army" Time="5" Unit="Disruptor"> <Resource index="Minerals" value="50"/> <Resource index="Vespene" value="100"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="32"/> <Button DefaultButtonFace="Disruptor" State="Restricted" Requirements="HaveCyberneticsCore"/>
+ Show Spoiler [Gateway] + <CAbilTrain id="GatewayTrain"> <EditorCategories value="Race rotoss,AbilityorEffectType:Structures"/> <InfoArray index="Train1" Time="33"> <Resource index="Minerals" value="100"/> <Button DefaultButtonFace="Zealot" State="Restricted"/> <Unit value="Zealot"/> </InfoArray> <InfoArray index="Train2" Time="42"> <Resource index="Minerals" value="125"/> <Resource index="Vespene" value="50"/> <Button DefaultButtonFace="Stalker" State="Restricted" Requirements="HaveCyberneticsCore"/> <Unit value="Stalker"/> </InfoArray> <InfoArray index="Train4" Time="55"> <Resource index="Minerals" value="50"/> <Resource index="Vespene" value="150"/> <Button DefaultButtonFace="HighTemplar" State="Restricted" Requirements="HaveTemplarArchives"/> <Unit value="HighTemplar"/> </InfoArray> <InfoArray index="Train5" Time="55"> <Resource index="Minerals" value="125"/> <Resource index="Vespene" value="125"/> <Button DefaultButtonFace="DarkTemplar" State="Restricted" Requirements="HaveDarkShrine"/> <Unit value="DarkTemplar"/> </InfoArray> <InfoArray index="Train6" Time="42"> <Resource index="Minerals" value="50"/> <Resource index="Vespene" value="100"/> <Button DefaultButtonFace="Disruptor" State="Restricted" Requirements="HaveCyberneticsCore"/> <Unit value="Disruptor"/>
The Sentry is reffered as Disruptor in the data files.
As you can see all units have their cooldowns set to be 10 seconds less than gateway training time. Having it as a percentage would be more intuitive, but it's just a flat reduction.
|
On May 18 2010 05:54 lololol wrote:Changing back and forth when building units with the same cooldown is effectively a 23 second buildtime for every second unit built, no matter the actual unit buildtime. Using it to mass zealots is obviously useless, since they have a 23 second warp in anyway. Sentry/Stalker production is effectively increased to 2 per 55 seconds, instead of 2x32=64 seconds. HT/DT production is effectively increased to 2 per 68 seconds, instead of 2x45=90 seconds. It is of course most effective if you're warping in high cooldown units and training low cooldown/buildtime units from the gateways. For example: warping in a HT, switching and training a zealot, then switching back will take a total of 46 seconds, while the cooldown of the HT is normally 45 seconds, so you effectively get an extra zealot for 1 second of buildtime. Here's the gateway and warpgate build data copied directly from the data files from the latest patch: + Show Spoiler [Warpgate] + <CAbilWarpTrain id="WarpGateTrain"> <EditorCategories value="Race rotoss,AbilityorEffectType:Structures"/> <InfoArray index="Train1" Category="Army" Time="5" Unit="Zealot"> <Resource index="Minerals" value="100"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="23"/> <Button DefaultButtonFace="Zealot" State="Restricted"/> </InfoArray> <InfoArray index="Train2" Category="Army" Time="5" Unit="Stalker"> <Resource index="Minerals" value="125"/> <Resource index="Vespene" value="50"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="32"/> <Button DefaultButtonFace="Stalker" State="Restricted" Requirements="HaveCyberneticsCore"/> </InfoArray> <InfoArray index="Train4" Category="Army" Time="5" Unit="HighTemplar"> <Resource index="Minerals" value="50"/> <Resource index="Vespene" value="150"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="45"/> <Button DefaultButtonFace="HighTemplar" State="Restricted" Requirements="HaveTemplarArchives"/> </InfoArray> <InfoArray index="Train5" Category="Army" Time="5" Unit="DarkTemplar"> <Resource index="Minerals" value="125"/> <Resource index="Vespene" value="125"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="45"/> <Button DefaultButtonFace="DarkTemplar" State="Restricted" Requirements="HaveDarkShrine"/> </InfoArray> <InfoArray index="Train6" Category="Army" Time="5" Unit="Disruptor"> <Resource index="Minerals" value="50"/> <Resource index="Vespene" value="100"/> <Cooldown Link="WarpGateTrain" Location="Unit" TimeUse="32"/> <Button DefaultButtonFace="Disruptor" State="Restricted" Requirements="HaveCyberneticsCore"/> <CAbilTrain id="GatewayTrain"> <EditorCategories value="Race rotoss,AbilityorEffectType:Structures"/> <InfoArray index="Train1" Time="33"> <Resource index="Minerals" value="100"/> <Button DefaultButtonFace="Zealot" State="Restricted"/> <Unit value="Zealot"/> </InfoArray> <InfoArray index="Train2" Time="42"> <Resource index="Minerals" value="125"/> <Resource index="Vespene" value="50"/> <Button DefaultButtonFace="Stalker" State="Restricted" Requirements="HaveCyberneticsCore"/> <Unit value="Stalker"/> </InfoArray> <InfoArray index="Train4" Time="55"> <Resource index="Minerals" value="50"/> <Resource index="Vespene" value="150"/> <Button DefaultButtonFace="HighTemplar" State="Restricted" Requirements="HaveTemplarArchives"/> <Unit value="HighTemplar"/> </InfoArray> <InfoArray index="Train5" Time="55"> <Resource index="Minerals" value="125"/> <Resource index="Vespene" value="125"/> <Button DefaultButtonFace="DarkTemplar" State="Restricted" Requirements="HaveDarkShrine"/> <Unit value="DarkTemplar"/> </InfoArray> <InfoArray index="Train6" Time="42"> <Resource index="Minerals" value="50"/> <Resource index="Vespene" value="100"/> <Button DefaultButtonFace="Disruptor" State="Restricted" Requirements="HaveCyberneticsCore"/> <Unit value="Disruptor"/> The Sentry is reffered as Disruptor in the data files. As you can see all units have their cooldowns set to be 10 seconds less than gateway training time. Having it as a percentage would be more intuitive, but it's just a flat reduction.
Cool beans, so you can actually see some significant advantages with this if you are making templar. I'm glad someone with a copy of SC2 available was able to get some real numbers here, because this thread was getting wayyyy derailed. I guess I was mistaken about the 30% its only 30% for zealots and on the tooltip, so you will infact see slight production increase for stalkers/sentry and even more for templar. I'm surprised I didn't notice this when I tested it before, must've only tested with zealots.
Assuming you do this perfectly you save:
-4.5 game seconds per stalker or sentry
-11 game seconds per templar (WOAH)
|
Tested it and yes it works. But it felt like the time gained from making just stalkers was very small (who builds ht's anyway ). Even though switching back and forth between gateway and warpgate does increase the production rate, the 2nd unit you build always pops earlier if you just warp them in which kind of ruins the chance of it saving you in most heated moments. Still cool though, hope blizzard keeps it this way just for the crazyness.
|
|
|
On May 18 2010 06:38 airen wrote:Tested it and yes it works. But it felt like the time gained from making just stalkers was very small (who builds ht's anyway ). Even though switching back and forth between gateway and warpgate does increase the production rate, the 2nd unit you build always pops earlier if you just warp them in which kind of ruins the chance of it saving you in most heated moments. Still cool though, hope blizzard keeps it this way just for the crazyness.
Yeah, 4.5 game seconds isn't much. You probably lose half of that with the slight delays between conversions unless you are constantly babysitting your gates, but the templar production reduction is pretty huge. Might be very useful if you have saved up gas and your templar tech finishes.
|
I've done the math for this so many times >.>
The way it works out, both methods (Pure Warpgate vs ReGating, which I like to call it) get their second unit out at around the same time. HOWEVER, ReGating means you can get a THIRD unit MUCH earlier, while the pure Warpgates will have a huuuuuuge wait before you can get the third unit.
Here are a couple of links:
http://www.teamliquid.net/forum/viewmessage.php?topic_id=117241
http://www.teamliquid.net/forum/viewmessage.php?topic_id=113568¤tpage=15#287
In the end, it's generally worth switching between them. It's a huge APM sink that grows by tremendous amounts as you get more Warpgates, and it speeds up your production by quite a bit.The ONLY moment that just using Warpgates will get units out faster is when you create only 2 units, but when are you going to make JUST 2 units? >.>
Note that if you're not using a Warpgate build, it would be negligible in the first place. But the more Warpgates you have, the quicker it all adds up (Imagine pumping out 3 units from 5 Warpgates, assuming you have infinite APM [just for the purposes of this example]; You have your units minutes before you would have them otherwise)
EDIT: Here's the most interesting part (I think) from the posts I linked:
On March 26 2010 09:12 Zeke50100 wrote:
A small list, if you guys are wondering:
Without the Trick (Standard Warpgate)
Templar + Zealot = 50 seconds to finish Zealot, 73 seconds for third unit
Stalker/Sentry + Zealot = 37 seconds to finish Zealot, 60 seconds for third unit
Templar + Stalker/Sentry = 50 seconds to finish St/Se, 92 seconds for third unit
With the Trick (Reverting WarpGate - ReGate?)
Templar + Zealot = 46 seconds to finish Zealot, 51 seconds for third unit
Stalker/Sentry + Zealot = 46 seconds to finish Zealot, 51 seconds for third unit
Templar + Stalker/Sentry = 55 seconds to finish St/Se, 60 seconds for third unit
Note that what the third unit is doesn't matter, because it will always take 5 seconds to warp in. Of course, it does matter if you want to continue the math, but you get similar results.
|
i tried that in the beginning of beta and considered it to be not worth it. Maybe they changed times? Well, maybe it's good for templars.
|
So the question is, how good of a player do you have to be to really take advantage of this. It would seem if you don't time it right the gain won't be that significant.
|
|
|
|
|
|
EPT
