|
OK, I'm just a diamond zerg so you can choose not to take my advice. I will say I've roach rushed a fair bit in my past, and there's one huge difference between this build and a normal roach rush.
The roach warren timing.
This guy throws down his warren at around 2:57, immediately after his pool finishes. He's also making a queen at this point (you can see his hatch wobbling differently so you can confirm this, no reason he wouldn't be making a queen of course).
In a standard roach rush (5RR/7RR) you throw down the warren when the queen is ~75% complete. That gives you time to get lings out and try to deny a scout. From his immediate warren, you can be fairly sure this build is coming; there's no reason to have the warren earlier otherwise.
The next important part is that you definitely could have scouted this earlier warren; it starts just as you tell your probe to leave. Had you scouted around his base for just a couple of seconds, you'd have seen how early his warren was, and there weren't any zerglings coming to stop you for a while yet.
This is where my suggestions start:
1) You definitely could and should have scouted the early warren timing. I understand you were trying to block his hatch, but that's fairly futile against pool first anyway. A speedling expand gets to around 20 supply before expanding with 4 lings, which completely stop any block attempts; at best you'd have delayed his hatch with a pylon for a couple of seconds and lost the probe. Why am I saying this? It's pretty pointless to try and delay the hatch when the pool's already finished, and you don''t lose much by staying in his base after scouting the really fast roach warren.
2) There was literally no danger at all to stop your scout until the roach warren was halfway done. This is when 2 lings pop, which couldn't have killed your probe straight away if you'd run about in his base off creep. You could probably scout his warren getting to about 75% done. At this point, even if it's a fake to trick you and he cancels it after your probe dies, he's lost so much from cancelling and lost mining time that the death of your scouting probe is justified. By staying in his base, you also see he's not building drones, further delaying his econ if he's faking the rush.
3) When the warren is 75% complete, I'd say it's a fairly safe bet he's not faking the build because the lost mining time of the drone and the lost minerals from cancel mean a standard speedling expand would always be better. If he does fake at this point, he's miles behind against standard play. From this point, we can assume he is doing the rush.
4) Let's assume that you see the warren get to 50% and your probe spontaneously combusts as his lings pop (i.e. you can't keep it alive longer than that point). If you put down a forge now, it'll be done by 4:10 and you haven't lost much. He is miles behind on econ (6 harvesters behind, in fact) already and if pressure doesn't work, he's definitely at a massive disadvantage. You can afford to delay a probe and get a forge and still be ahead. Note that this forge isn't blind, it's in response to the roach warren immediately after pool which is 100% scoutable.
5) When your not-blind-forge finishes at 4:10, 2 roaches are at 75% done and 1 is at 25% done and his army consists of 2 lings. He's not a threat until his next round of larvae. 2 roaches pop at 4:19. The next roach finishes at 4:26, and his inject finishes at this point too. He is totally committed to the rush at this point, and I'd go as far as to say at this point he's all in. The roaches arrive at 5:10 ish, which gives you a whole minute to make cannons in preparation of the rush. You can afford 2 cannons just by not making the third gate, which probably isn't going to help too much even after it finishes, especially when you can see his roaches coming. 2 cannons will go a very long way to stopping what this guy did. You can certainly justify making a delayed third cannon if you still feel pressured. At this point you know he has large amounts of lings and 3 roaches (read: not a large amount of drones back home).
Those cannons will win you the game, and even if they don't and he expands, he's massively behind. Even if he solidly makes drones for a few minutes, your econ will be better for a long, long time after holding him off, letting you expand or 1 base-allin (a 4WG response would be strong) him after a couple of minutes when your stronger econ has benefited you and when his heavy drone production hasn't had chance to benefit him yet. His army can't get bigger if he pumps drones, and yours will naturally and at a faster rate because of your econ advantage now.
If he was faking the build, the suggestions that he could double/triple expand are hard to take seriously. If you get fooled and make a forge and 2 cannons, you've spent 450 minerals to be invincible for a good long time and he's got to spend 300 + a drone of which he has very few to make a single, very delayed hatch that won't benefit him for 100 seconds. If nothing attacks you by 5:20, you can safely chrono out a shitload of probes, expand and be miles ahead. He'll have to be mass producing drones because he faked it, meaning no army, meaning a safe expo for p. Alternatively since you have a forge, get a chrono'd +1 attack upgrade and do a timing attack. He's unlikely to have gotten carapace since he'll have had to focus on drone production, so his lings will melt.
Basically, faking this build would be crazy and nearly always leave you behind unless your opponent goes nuts and over produces cannons. The build is strong, but if you drop a forge in response to his early roach warren, you can easily stop it in a similar fashion to how a 6 pool is strong until you scout it and make a complete wall.
Finally, to comment on what he said about needing 3 cannons because of a baneling followup; ignore that, he's miles behind if he has to get to that point, his econ is still trash because he can't feasibly make drones and units at the same time, and you'll have sentries with forcefields to stop him more than likely. If he's had time to get a baneling nest as well, you've had time to hold him off for a while, make a building to wall off, get more units and make more cannons if needed. You can probably even make some probes while doing all this.
Sorry for the massive wall of text  As I said you don't need to take my advice.
EDIT: removed TL;DR.
|
2.6k terran
I'm not sure if this will help too much, but I recognize the roach rush if when at 3:30 he doesn't have an expo. When I see that I send the worker into his base... right clicking the minerals to make sure it'll pass through any attacking units. If he doesn't have expo and he kills my worker I just throw down a scan... Now seeing as how you can't do that, what do you think about throwing down the forge when you scout no expo and then using the cannons to expand on top of that so that you're not behind?
|
This is a better variation of a speedling all in that I do a lot against protoss and gives me plenty of free wins (10 gas 11 pool).
Reason for it is that protoss are so used to the "fact" that a wall in with gateway, cybernetics core, and 1 zealot (and really safe, with 1 zealot 1 sentry), makes them completely safe.
So I'm scouted - repeatedly. And I do the same build against the same players two games in a row. Against different players. And still - almost no one throws down a forge then cannons. Even though they see that I am getting ling speed, and have 12 drones. So they have to know that I am going all in pretty much (what else could I do? fake all in and mass drones?).
Then I simply attack move into their base, they throw down a force field, and think that makes th em safe - but when it ends, I just power through.
The proper response to this is cannoning into a forge expand imho, or if you were smart, you threw down a proxy pylon with your scouting drone when you saw it, and then you can cannon up and go normal 4 gate. Which will crush it honestly (move into base with 6 zealots, 2 sentries, force field his ramp so his units cannot get back, and gg when you kill his drones and queen).
You are actually far ahead in economy, so yes, you can afford to throw down forge, 3 cannons, and upgrade +1 melee attack (the variation that makes it very strong against mass ling).
I mean, a zerg sacrifice an overlord to see if it's a strong 4 gate that requires cutting drones and getting spine crawlers (1 gas 20ish probles = oh yeah). That's a stronger rush actually, it's just so well known and have been practiced for months so everyone knows the theory of how to deal with it (static defense on most maps, + cutting drones and a good unit combination).
So to me, this is a strong rush, but it's easily beatable by scouting, and treating it like the rush it is: static defense being key to beating it.
Possible to beat by units? Sure ... but that's just making it harder for yourself. Why would you do it? You will be well ahead by simply getting forge, 3 cannons, and chronoboosting probes or getting tech.
You have to know that when the rush fails, he will mass drones, and expo once, and try to use his units to contain you and stop you from expanding. That means his tech will be really late. DT would be a pretty certain option if you failed to put down a proxy pylon somewhere to kill his base - and it goes well with forge expand since it's so gas heavy.
If anyone wants to practice against this and other types of early zerg all ins on EU as protoss, aebriol.380 - but I am only 2500 master.
I still don't think this is as powerfull as terran 2 rax bunker, or 4 gate (20 or so probes), as far as all ins go, but it's unusual, which makes it hard in itself.
|
On February 02 2011 17:55 travis wrote:Show nested quote +On February 02 2011 17:53 TearDrop wrote:
You can´t hold a 4gate without building 1-2 spinecrawlers yes... u certainly can. also, spinecrawlers can move, AND they are being used to protect an expansion, which makes it a ridiculous comparison anyways.
Though you can transition out of a 4 gate much more easily.
It's going to be impossible for the zerg to recover if the push does no dmg, so the cannons would be well worth it anyway.
|
A well executed 4gate, either an offensive one with temporary probe cut @22-23 or a defensive one with zealot, stalker, sentry from 1gate beats this. On close position you probably have to replace the blocking zealot with a pylon, and you should always be careful spending more then three cb on your nexus vs zerg.
|
Change the way you wall in. I change the way I walled in because of stuff like this. Line up your core and gateway at the ramp vertically with a max distance pylon (this leaves a gap for a building that is 3x3). If your scout probe sees the roach warren going down immediately after the pool, then save that gap for your next gateway. If it looks like a normal opening, you would normally have the 100 minerals for a pylon when the zealot is about to pop out. The zealot normally comes out right before lings from a standard gas/pool build gets to your base, so your ramp will still be guarded by a zealot from a normal opener (and the pylon won't be an issue with baneling busts, because you are getting sentries. This gives your wall way more HP, and 3 roaches cannot snipe a core or gateway down fast enough (especially since your core is centered as far from the ramp as possible), and it provides enough HP to get you some extra time for warpgate. Your gates warped in right as he finally got through the forcefields, and you faltered due to a risky wall. Poster a few above me is correct that this is 100% scoutable and fully telling of what is coming.
*also, quite a few people are suggesting that you do in fact get cannons, which is completely a correct decision to make in this case. Instead of adding the second gateway to wall off completely with core + initial gate, just drop a forge in its place.
Hope this helps out!
|
On February 02 2011 18:03 sgtcodfish wrote:
TL;DR: You can easily scout an immediate roach warren after pool, which is a tell for this build, and make a forge in response to this and have cannons up in time to defend. Even if it was a fake, you're ahead.
you know, instead of the 20+ zerglings he made, he could easily just pull 3 drones from gas, and add 3 more before the initial push, leaving him closer to 20 (i don't cut drones at 14 lol wtf man)
even from the replay, his zerglings were overkill.
|
I just played a few ladder games so that I could try this build out.
It worked, but it was against a much lower ranked player.
|
On February 02 2011 17:45 Obelisk7 wrote:
I watched the replay and your cybernatics core went down right before his roach warren. Also, your probe left right after he started the warren.
Cancel the core, start a second Gateway, and then get two zealots. Build a core, get two more zealots.
By the time his attack comes, the minerals you used for the third gateway with what you did normally instead is used on forces to hold the attack off. Core should finish while you chrono out zealots to deal with mass lings, and have a sentry chrono'ed with the other gate providing zealots.
From there I'm not sure how things would pan out...but if the attack is being held off then forcefield and reinforce army.
Would something like that work ?
No. Overlords can give vision, then roaches will just pick off zealots and zerg will lose nothing. Your sentries will be late.
Also it's actually ridiculous how many players comment here who haven't tested the timings. Getting an immortal out at 5:15? Really? Teching to stargate? You can't be serious. Although I understand your resentments regarding the problem of having to defend with cannons, since I've seen even HuK do this against a 7RR by Slush about 3 months ago I've come to accept that...well....it's just how it is. I don't think you can defend without forge cost efficiently (meaning without probe-pull).
Now for the theory-craft on scouting: The main point about scouting this and other early roach one-basing strats is the pool BEFORE the gas. If he doesn't get pool before the gas, then the roach warren will be severely delayed, screwing up timings. 13 pool 12 gas isn't easy to catch unless your scout arrives right before the buildings are thrown down, but if you catch him doing this, you KNOW that there's a high chance he will do something all-in-ish. At this point, I'd definitely say you need to run your probe back in and scout for the potential roach warren. If he did indeed just go for the usual speedling-expo, then his build will be quite uneconomical - worth sacking a probe for this intel, imo.
|
You're probe gets to his base and does a quick check for gas.. then stops at 2:33 and doesnt move until 3:03. His pool finishes at 2:52.. so that means that at the earliest he will have lings hatch at 3:16. He starts his roach warren at 2:57.
So you have 20 seconds after pool finishes to look for a warren. I think you just need to be more active with your probe in his base. But this is just one replay so... and you said that you knew it was coming.
|
I am not a protoss, but I noticed that my build finishes warpgate research ten seconds earlier than you did in your game. That would have been enough to warp in two sentries to force field the ramp as he walks up splitting the army, or any other combination of units. I could get three gates up by that point along with a zealot and two sentries before the tech finished. I'm not sure if it will stop this build but it definitely seems like a step in the right direction. I think i did a 12 or 13 gate, and only chronoboosted probes twice before saving energy for warpgate. I didn't chronoboost units, but one less chronoboost could be used on probes to account for that if it is important.
|
It's interesting because this is a rush in pvz that's extremely strong. It's funny, because one of my friends told me he found a new rush which was just speedlings, where he won practically all of his zvps on the ladder.
Now obviously my post sadly got kind of buried with the activity going down on the chat.
I really, dislike the idea of getting a forge. Unlike zerg or terran, a cannon is something not on our standard early tech path, and deviations can cause lots of problems, like if they then proceed to mass expand, you're left with static defense you'll hardly use, and can not move in anyway what so ever, not to mention they barely guard your natural.
So again, (page 14 if you want to check), I posted up a build I thought was viable as a counter.
2 zealot, stalker, sentry, (10 probes also, considering the advantage in economy)
vs 3 roach 12 speedlings. Considering the ff, I would imagine it's a close call, any other thoughts?
|
On February 02 2011 18:10 Gooey wrote:
Change the way you wall in. I change the way I walled in because of stuff like this. Line up your core and gateway at the ramp vertically with a max distance pylon (this leaves a gap for a building that is 3x3). If your scout probe sees the roach warren going down immediately after the pool, then save that gap for your next gateway. If it looks like a normal opening, you would normally have the 100 minerals for a pylon when the zealot is about to pop out. The zealot normally comes out right before lings from a standard gas/pool build gets to your base, so your ramp will still be guarded by a zealot from a normal opener (and the pylon won't be an issue with baneling busts, because you are getting sentries. This gives your wall way more HP, and 3 roaches cannot snipe a core or gateway down fast enough (especially since your core is centered as far from the ramp as possible), and it provides enough HP to get you some extra time for warpgate. Your gates warped in right as he finally got through the forcefields, and you faltered due to a risky wall. Poster a few above me is correct that this is 100% scoutable and fully telling of what is coming.
Can you post a screenshot of this kind of wall-in? Thanks
|
I have been nuking the shizz out of the build on Diamond players (prac-partners) and un-scouted is a insta win!!
I LOve that you found this.
|
On February 02 2011 18:08 aebriol wrote:
This is a better variation of a speedling all in that I do a lot against protoss and gives me plenty of free wins (10 gas 11 pool).
Reason for it is that protoss are so used to the "fact" that a wall in with gateway, cybernetics core, and 1 zealot (and really safe, with 1 zealot 1 sentry), makes them completely safe.
So I'm scouted - repeatedly. And I do the same build against the same players two games in a row. Against different players. And still - almost no one throws down a forge then cannons. Even though they see that I am getting ling speed, and have 12 drones. So they have to know that I am going all in pretty much (what else could I do? fake all in and mass drones?).
Then I simply attack move into their base, they throw down a force field, and think that makes th em safe - but when it ends, I just power through.
The proper response to this is cannoning into a forge expand imho, or if you were smart, you threw down a proxy pylon with your scouting drone when you saw it, and then you can cannon up and go normal 4 gate. Which will crush it honestly (move into base with 6 zealots, 2 sentries, force field his ramp so his units cannot get back, and gg when you kill his drones and queen).
You are actually far ahead in economy, so yes, you can afford to throw down forge, 3 cannons, and upgrade +1 melee attack (the variation that makes it very strong against mass ling).
I mean, a zerg sacrifice an overlord to see if it's a strong 4 gate that requires cutting drones and getting spine crawlers (1 gas 20ish probles = oh yeah). That's a stronger rush actually, it's just so well known and have been practiced for months so everyone knows the theory of how to deal with it (static defense on most maps, + cutting drones and a good unit combination).
So to me, this is a strong rush, but it's easily beatable by scouting, and treating it like the rush it is: static defense being key to beating it.
Possible to beat by units? Sure ... but that's just making it harder for yourself. Why would you do it? You will be well ahead by simply getting forge, 3 cannons, and chronoboosting probes or getting tech.
You have to know that when the rush fails, he will mass drones, and expo once, and try to use his units to contain you and stop you from expanding. That means his tech will be really late. DT would be a pretty certain option if you failed to put down a proxy pylon somewhere to kill his base - and it goes well with forge expand since it's so gas heavy.
If anyone wants to practice against this and other types of early zerg all ins on EU as protoss, aebriol.380 - but I am only 2500 master.
I still don't think this is as powerfull as terran 2 rax bunker, or 4 gate (20 or so probes), as far as all ins go, but it's unusual, which makes it hard in itself.
thanks needed that info...don't know much about Zerg as I don't play that race at all...so how many drones does Zerg tend to have to place a standard pool and extractor to get speed?
|
do what u did that game but just pylon block ur entrance. Its when the lings get in ur base when it gets bad.
|
On February 02 2011 18:35 Looky wrote:
do what u did that game but just pylon block ur entrance. Its when the lings get in ur base when it gets bad.
A pylon block actually gets sniped down pretty fast when its warping in by lings and the 3 roaches. However, a gate block is far better and serves the same purpose imho, costing only 50 minerals more. Also you can cancel it at the last moment so you dont have to blow it up to get out of the base
|
On February 02 2011 18:37 Nis wrote:Show nested quote +On February 02 2011 18:35 Looky wrote:
do what u did that game but just pylon block ur entrance. Its when the lings get in ur base when it gets bad. A pylon block actually gets sniped down pretty fast when its warping in by lings and the 3 roaches. However, a gate block is far better and serves the same purpose imho, costing only 50 minerals more. Also you can cancel it at the last moment so you dont have to blow it up to get out of the base
not when he pylon block it at his small choke where 1 ling would hit it. every protoss block with gateway and core. no way he would break a pylon with 3 roaches. plus you wouldnt want to cancel a gateway since you need to make units.
|
The warren is down before anything is out to kill the scout, so you wouldn't blindly have to cannon, just cannon in response to a rush.
|
On February 02 2011 18:28 Ingruz wrote:Show nested quote +On February 02 2011 18:10 Gooey wrote:
Change the way you wall in. I change the way I walled in because of stuff like this. Line up your core and gateway at the ramp vertically with a max distance pylon (this leaves a gap for a building that is 3x3). If your scout probe sees the roach warren going down immediately after the pool, then save that gap for your next gateway. If it looks like a normal opening, you would normally have the 100 minerals for a pylon when the zealot is about to pop out. The zealot normally comes out right before lings from a standard gas/pool build gets to your base, so your ramp will still be guarded by a zealot from a normal opener (and the pylon won't be an issue with baneling busts, because you are getting sentries. This gives your wall way more HP, and 3 roaches cannot snipe a core or gateway down fast enough (especially since your core is centered as far from the ramp as possible), and it provides enough HP to get you some extra time for warpgate. Your gates warped in right as he finally got through the forcefields, and you faltered due to a risky wall. Poster a few above me is correct that this is 100% scoutable and fully telling of what is coming.
Can you post a screenshot of this kind of wall-in? Thanks
Here ya go http://img64.imageshack.us/i/tosswallin.png/
This gives you the room to block with an additional forge/gateway against any cheese and still provides the zealot blocking gap. You will be able to afford the pylon before the zealot pops out with a normal sentry expand build. It is the safest wall imo.
|
|
|
|
|
|
EPT
