|
On December 11 2010 13:40 jdseemoreglass wrote:
Once the opportunity for flaming disappears, Skrag leaves as quickly as he came...
No apology for turning my thread into a childish mudslinging battle for hours... He attacks me and calls me a jackass twice, a moron twice, a blind-single-minded fanatic, and so on, and he won't even bother to end with a civil comment after I give him what he wants.
Trying to provide for the TL community is really not worth the trouble it entails... which explains a lot.
Yeah, I prefer to try and avoid all of the name calling as much as possible. I'm not sure what order things happened, but you did make an unwarranted comment on the other thread too 
He was probably more annoyed, as was I, that you kind of gave up on any testing so easy. It just appeared that you were trying to champion the 11P/18H as something it might not be no matter what. The reason I really stopped arguing before was because it was obvious that it would take a while to get more reliable data, and I didn't want to stir the pot far before I could really give any evidence. After our first discussion like 5 days ago I started trying to think of ways to work with the huge discrepancies in results from replay to replay. It seemed like an obvious problem that was making it hard to really be sure about any conclusions based on our replays.
|
I have to say I really like this build in team games. 1v1... I'm not really sold because creep spread is so important and the opposition can force you to get zerglings before you put down the hatch on 18 making you way behind in the long run. I could see it being effective in PvZ... but for the other two matchups I don't think its ideal.
BTW 2100 1v1, 1700 2v2 team.
|
On December 11 2010 13:52 grannock wrote:
I have to say I really like this build in team games. 1v1... I'm not really sold because creep spread is so important and the opposition can force you to get zerglings before you put down the hatch on 18 making you way behind in the long run. I could see it being effective in PvZ... but for the other two matchups I don't think its ideal.
BTW 2100 1v1, 1700 2v2 team.
If you have a replay of this problem people are claiming to have, of having to make zerglings before hatch, I would love to see them. I have heard this several times but haven't been able to actually witness what people mean, or how it would differ from a hatch-first play.
Hatch first will certainly allow you to get an earlier spine, but I don't see why it is necessary to sacrifice so many drones because of the short time between spine placement...
I heard some people say that zealot aggression would hurt this build. I actually encountered a game on the ladder where someone got quite a few early zealots and put 4gate pressure early on. Here is a rep for those interested. I suppose they are referring to earlier pressure, but I've yet to encounter this on the ladder.
![[image loading]](http://www.gamereplays.org/community/uploads/repimgs/repimg-33-172974.jpg)
|
On December 11 2010 14:03 jdseemoreglass wrote:Show nested quote +On December 11 2010 13:52 grannock wrote:
I have to say I really like this build in team games. 1v1... I'm not really sold because creep spread is so important and the opposition can force you to get zerglings before you put down the hatch on 18 making you way behind in the long run. I could see it being effective in PvZ... but for the other two matchups I don't think its ideal.
BTW 2100 1v1, 1700 2v2 team. If you have a replay of this problem people are claiming to have, of having to make zerglings before hatch, I would love to see them. I have heard this several times but haven't been able to actually witness what people mean, or how it would differ from a hatch-first play. Hatch first will certainly allow you to get an earlier spine, but I don't see why it is necessary to sacrifice so many drones because of the short time between spine placement... I heard some people say that zealot aggression would hurt this build. I actually encountered a game on the ladder where someone got quite a few early zealots and put 4gate pressure early on. Here is a rep for those interested. I suppose they are referring to earlier pressure, but I've yet to encounter this on the ladder.
Its early zealot aggression or a pylon at your natural + 1 zealot and probe. You would need at least 6 zerglings to fight it off and by the time your natural is up your extremely vulnerable to the 6 zealot 2 stalker 1 sentry push. The real problem with this build as you have posted it, is you are saving your 3rd overlord for after the hatchery. Against an aggressive opponent this will not be an option, which delays the hatch more, which then skews all of your data results.
Honestly, I have really never tried this in 1v1 because you cant scout on 9, and then you are very vulnerable to early pressure because your in the dark, and just praying that you can get a hatch off at 18. I dont see it working well. I have 800+ games played at the upper level of the diamond bracket and I see too many holes for this in 1v1. The >slight< advantage this has over a 14 pool is negligible because you are slightly stronger with 14pool prior to the first larva inject, which is when your fending off early probe/scv harass to put down your hatch.
Hatch delayed, no spines, you wont stand a chance against a an early non-allin from a terran or protoss, let alone an allin.
Like I said, I'm loving this build in my team games. It enables you to defend any cheese, or drone up hard if they are teching. Also, hatch first in 2v2 is very risky, so this strat allows for one to get a nasty econ without rolling the dice on if your opponent is rushing.
The issue with 1v1, is if you are good at hatch first, or vs a protoss, the 14 pool, you are not rolling the dice. You know its going to work, and its slightly better than this build you have posted. Sure for lower calliber players that cant stop the 10 pool, or the 3 zealot/5 marine rush, try this build out and you might like it for 1v1. For the players on the upper end of the ladder, there is no point because these early rushes lead to an easy victory after they are squashed.
|
Does any one know of any VODs of 14/15 hatch/pool vs 2 Rax FE on youtube? I want to throw the OP linked games and one of those on the doubler (http://www.youtubedoubler.com/) to compare stuff.
EDIT: I just trawled through the post GSL2 stuff on Husky's channel and no dice on an aggressive 2 Rax FE that was executed well. It doesn't help that he looks at completely random stuff either.
|
On December 11 2010 14:21 Antisocialmunky wrote:
Does any one know of any VODs of 14/15 hatch/pool vs 2 Rax FE on youtube? I want to throw the OP linked games and one of those on the doubler (http://www.youtubedoubler.com/) to compare stuff.
I've seen plenty of Hatch-first vs. 2rax in the GSL... I doubt the reps on youtube will defend better than NesTea or IdrA, etc.
But yeah, I would love to see them either way. I hear people saying that a spine helps the defense but I've never seen a spine finish in time so I can't really assess this...
|
Hmm for defense vs 2rax a fast spine is very good.
Ret and IdrA have been working on a build were you get your 2nd overlord late in order to get a fast spine off your FE. If your doing the 11pool 18hatch you should be able to hold off 2rax easily enough, but if you have problems you also make a spine at ur main while ur FE is building and then move it over, assuming the walk distance isn't too big (could be tough on steppes).
I'm a little confused by the lack of replay reliability. I'm a 2,300 pt ranked 1v1 random diamond on NA and I'm considering trying each of these builds with perfect execution and see how they stack up. The main contenders are 14/15 hatch/pool vs 11/18 pool/hatch right? It seems easy enough to compare this across several games as an impartial 3rd party. Would this be useful?
|
On December 11 2010 14:40 Alternity wrote:
Hmm for defense vs 2rax a fast spine is very good.
Ret and IdrA have been working on a build were you get your 2nd overlord late in order to get a fast spine off your FE. If your doing the 11pool 18hatch you should be able to hold off 2rax easily enough, but if you have problems you also make a spine at ur main while ur FE is building and then move it over, assuming the walk distance isn't too big (could be tough on steppes).
I'm a little confused by the lack of replay reliability. I'm a 2,300 pt ranked 1v1 random diamond on NA and I'm considering trying each of these builds with perfect execution and see how they stack up. The main contenders are 14/15 hatch/pool vs 11/18 pool/hatch right? It seems easy enough to compare this across several games as an impartial 3rd party. Would this be useful?
At this point, any replays from diamond players would be useful... I seem to be the only one posting them. It would be even more useful if someone had a practice partner who is competent at 2rax pushes so we could do some comparisons.
|
Was it Skrag who mentioned the AI had 4,000 APM while testing that build? For a layman, I hope you don't have to have insane APM to get consistently better results with the 13p/15h over the 11p/18h...I hope. Do we have replays of actual human players doing these builds and the differential still being about 150 minerals and more larvae?
|
On December 11 2010 14:40 Alternity wrote:
Hmm for defense vs 2rax a fast spine is very good.
Ret and IdrA have been working on a build were you get your 2nd overlord late in order to get a fast spine off your FE. If your doing the 11pool 18hatch you should be able to hold off 2rax easily enough, but if you have problems you also make a spine at ur main while ur FE is building and then move it over, assuming the walk distance isn't too big (could be tough on steppes).
I'm a little confused by the lack of replay reliability. I'm a 2,300 pt ranked 1v1 random diamond on NA and I'm considering trying each of these builds with perfect execution and see how they stack up. The main contenders are 14/15 hatch/pool vs 11/18 pool/hatch right? It seems easy enough to compare this across several games as an impartial 3rd party. Would this be useful?
I'm not sure because I haven't done enough testing with human vs. AI to know how much human error comes into play when you have a person executing really well. If it is a significant effect, really that just means you need more trials to get reliable data. However I tend to think, given a person who can actually do perfect execution every time, the reliability should be same. The other thing to note is that even with AI/perfect execution there are differences which make it hard to compare some builds. These differences, surprisingly enough arise from the larva random walk. The drones are produced at different locations, which affects the state that they reach minerals at. If they reach a mineral patch at the wrong time they will find it occupied and waste time moving to other minerals. The only way to really get reliable results is perfect execution and have many trials to average. I am however naturally suspicious that a person will play exactly the same every time. Will they grab a drone for the expansion at the same point every time? Will they place the spawning pool at the same time? That's why I prefer AI. I don't have to wonder about those little things being an influence on the results.
Anyways, after that long winded response, if you want to try and show things with replays played by people, the more replays the better I suppose.
|
On December 11 2010 14:44 jdseemoreglass wrote:Show nested quote +On December 11 2010 14:40 Alternity wrote:
Hmm for defense vs 2rax a fast spine is very good.
Ret and IdrA have been working on a build were you get your 2nd overlord late in order to get a fast spine off your FE. If your doing the 11pool 18hatch you should be able to hold off 2rax easily enough, but if you have problems you also make a spine at ur main while ur FE is building and then move it over, assuming the walk distance isn't too big (could be tough on steppes).
I'm a little confused by the lack of replay reliability. I'm a 2,300 pt ranked 1v1 random diamond on NA and I'm considering trying each of these builds with perfect execution and see how they stack up. The main contenders are 14/15 hatch/pool vs 11/18 pool/hatch right? It seems easy enough to compare this across several games as an impartial 3rd party. Would this be useful? At this point, any replays from diamond players would be useful... I seem to be the only one posting them. It would be even more useful if someone had a practice partner who is competent at 2rax pushes so we could do some comparisons.
I think this is a great idea, no matter what the build. Getting together and testing a build against different common builds is a great idea.
|
On December 11 2010 15:13 Enyalus wrote:
Was it Skrag who mentioned the AI had 4,000 APM while testing that build? For a layman, I hope you don't have to have insane APM to get consistently better results with the 13p/15h over the 11p/18h...I hope. Do we have replays of actual human players doing these builds and the differential still being about 150 minerals and more larvae?
That early in the game considering there is no fighting involved does saying 4,000 APM even matter (meaning that this entire strategy can be accomplished in relatively small number of actions), doesn't that just suggest that the actions taken by the computer are almost instantaneous?
|
On December 11 2010 15:53 eth3n wrote:Show nested quote +On December 11 2010 15:13 Enyalus wrote:
Was it Skrag who mentioned the AI had 4,000 APM while testing that build? For a layman, I hope you don't have to have insane APM to get consistently better results with the 13p/15h over the 11p/18h...I hope. Do we have replays of actual human players doing these builds and the differential still being about 150 minerals and more larvae? That early in the game considering there is no fighting involved does saying 4,000 APM even matter (meaning that this entire strategy can be accomplished in relatively small number of actions), doesn't that just suggest that the actions taken by the computer are almost instantaneous?
If you were just holding a button down you wouldn't hit 4000 apm.
|
On December 11 2010 15:13 Enyalus wrote:
Was it Skrag who mentioned the AI had 4,000 APM while testing that build? For a layman, I hope you don't have to have insane APM to get consistently better results with the 13p/15h over the 11p/18h...I hope. Do we have replays of actual human players doing these builds and the differential still being about 150 minerals and more larvae?
Why are people so suspicious of AI? It's not cheating or anything. It's just reliable.
Anyways, I was just watching a replay of the AI and in this particular test it ended the replay with an average APM of 12. Do you think you can manage 12 APM?
|
On December 11 2010 15:56 grannock wrote:
If you were just holding a button down you wouldn't hit 4000 apm.
Ya but thats my point, that number is meaningless if being used to suggest this build is beyond human APM abilities.
|
Hmm, well it seems worth trying. I have a friend who is also 2k+ diamond and is quite good at 2rax pressure. I'll play vs him with the 14hatch/15pool vs 11pool/18 hatch and see if my ability to defend is affected.
The purpose of these builds is to find the best econ opening that allows you to defend correct? At mid diamond level of play most players will check your pool timings, so the power of a 11pool shouldn't be under estimated.
I'm not sure if this has been discussed in detail but the "most optimal" build is going to be very map dependant. Something like the 11pool is great for steppes, but 14hatch might be better on maps like shakuras. IdrA and Ret's fast spine crawler on i believe 16 before 2nd overlord helps to defend vs 2rax aggression in that scenerio. Were as steppes probably couldn't finish spine before the push comes.
|
Can you post the replays when you get them? I would really love to get some actual game replays for builds other than the 11 Pool for my thread. I haven't gotten a single one yet.
|
On December 11 2010 15:57 jacobman wrote:Show nested quote +On December 11 2010 15:13 Enyalus wrote:
Was it Skrag who mentioned the AI had 4,000 APM while testing that build? For a layman, I hope you don't have to have insane APM to get consistently better results with the 13p/15h over the 11p/18h...I hope. Do we have replays of actual human players doing these builds and the differential still being about 150 minerals and more larvae? Why are people so suspicious of AI? It's not cheating or anything. It's just reliable. Anyways, I was just watching a replay of the AI and in this particular test it ended the replay with an average APM of 12. Do you think you can manage 12 APM?
I can definitely do that. It's just I had read in this thread and previous ones discussing the 11p/18h the need for rallying individual drones to certain mineral patches and splitting them correctly and doing all this (seemingly) complex stuff...and for a player who is just not that good (myself!) - I just want the best economic build that is also safe. Without having to execute tons of mechanical actions that will probably cause me to forget the actual steps of the build I'm executing. And so I wasn't positive this was possible with normal players and not perfect AI. If I'm wrong - awesome. I've got a new favorite build.
No disrespect or suspicion intended.
|
On December 11 2010 17:16 Enyalus wrote:Show nested quote +On December 11 2010 15:57 jacobman wrote:On December 11 2010 15:13 Enyalus wrote:
Was it Skrag who mentioned the AI had 4,000 APM while testing that build? For a layman, I hope you don't have to have insane APM to get consistently better results with the 13p/15h over the 11p/18h...I hope. Do we have replays of actual human players doing these builds and the differential still being about 150 minerals and more larvae? Why are people so suspicious of AI? It's not cheating or anything. It's just reliable. Anyways, I was just watching a replay of the AI and in this particular test it ended the replay with an average APM of 12. Do you think you can manage 12 APM? I can definitely do that. It's just I had read in this thread and previous ones discussing the 11p/18h the need for rallying individual drones to certain mineral patches and splitting them correctly and doing all this (seemingly) complex stuff...and for a player who is just not that good (myself!) - I just want the best economic build that is also safe. Without having to execute tons of mechanical actions that will probably cause me to forget the actual steps of the build I'm executing. And so I wasn't positive this was possible with normal players and not perfect AI. If I'm wrong - awesome. I've got a new favorite build. No disrespect or suspicion intended.
It's okay. Admittedly the AI does a really good job playing the builds because it does execute many actions right when the opportunity is available. So even though it has really low APM, it is really precise with its actions. However I do believe it is possible to replicate the results if you really put some effort into it. Also my AI doesn't do any of the fancy drone splitting or sending drones to particular mineral patches in the beginning.
|
On December 11 2010 16:42 Alternity wrote:
Hmm, well it seems worth trying. I have a friend who is also 2k+ diamond and is quite good at 2rax pressure. I'll play vs him with the 14hatch/15pool vs 11pool/18 hatch and see if my ability to defend is affected.
I would like to do this myself but alas I dont know any terrans. This might be obvious, but your "ability to defend" is obviously higher with the earlier pool, what you want to see is if tthat is too high of a price to pay compared to spine and drone micro. If you mind, try the following:
- contain. can terran force you to cancel the later hatch and bunker up outside your main while expanding / teching freely?
- assuming you build a bunch of units to hold off the rush, and that terran plays smartly (i.e. doesnt suicide everything but backs out once he sees your army, can you keep up economically in the midgame - not dying to a stim timing push / siege tanks slow push etc?
|
|
|
|
|
|
EPT
