|
ZvX scouting is an insanely hard problem, and is something that has engrossed me for a long time. There are so many things that seem to have no solution - you must guess - but in a lot of these cases there is no need to guess (there are some, definitely, but not as many as 99.999% of zergs currently think I'm sure).
ZvT scouting
This MU is not as seedy w.r.t. scouting as most people believe. It can still get pretty gross, but there are ways around some of the guessing people think is necessitated by wall-ins.
2:35-2:38 is 12 rax finish time. This means you can deduce the exact supplies of the 2 rax (11/11, 12/14, 12/12, etc), and also deduce gas first, which is very important because gas first hellions can start very early (4:21 for 11rax12gas instead of 4:38 for 12rax13gas) and the gas first banshee pops at 6:00, which is 33 seconds earlier than 12 rax 13 gas banshee.
1) General non close-air positions with 0 gas:
You begin with your drone scout. If you see the rax finishing at 2:35-2:38 like normal , and no gas, ATM you know he's at least 12 rax 0 gas. You have to figure out if he's proxy rax. Go to his mineral line and count the SCVs. At the timing just preceding OC morph, if he has 13 then you know he isn't proxy (the 14th pops the moment OC starts basically). If he has 12 and an scv doesn't come and scout you at appropriate time, he's proxy. I believe it is beyond debate that this is critical due to the diversity of proxy locations, and you will outright die if he pools marines and scv pulls at the right time and you've droned till your 24/25 overlord like you should against 1 rax FE. The only problem is you have to pull your drone from mining at around 10 supply on mid sized maps otherwise he can wall you in with a 15 depot if he's 0 gas 12 rax or 12 rax 13 gas (you need about 6 ingame seconds to count his scvs at the 2:30 mark, and it's quite easy to mess up with SCVs bouncing off full patches and so on).
2) Standard gas, 12 rax 13 gas:
You can figure if he's gas first or 12 rax 13 gas in the same fashion. Gas should finish a few seconds before rax, so if rax finish 2:35-2:38 and u see gas already done (and has 14 scv at OC morph or 13 just before) then by necessity he is 12 13. If you're hatch first and not 14g14p then you will leave your drone in his natural and around 4:00 you will check to see if he's reactor hellion. This is important because the first 2 hellions finish at 5:08 off 12 rax 13 gas 2 marine reactor hellion (production starts at 4:38, although it can be delayed to 4:42 with the slightest inefficiencies), and your first two lings off 1515 and even 1514 (yes this is faster lings than 1414) get there too late to put 2 spines in time. Hellions will be in your natural before the spines finish if you rely on this.
Once reaper and reactor hellion are eliminated from his range, you must be aware of 12 rax 13 gas banshee finishing 6:33, and if he's 1 port then cloak finishes at 7:15. If you're one of those zergs who does quick 3 hatch without lair then against 12 rax 13 gas with no reaper or reactor hellion (so 1-1-1 str8 into cloak banshee is a large part of his range), you'll need to start the evo at 6:10 so your spore finishes at 7:15 for the 12 gas 13 rax cloak banshee that will be on your base at this moment. I'm not sure whether it's good to patrol lings around his main and cancel the evo at 6:45 when he shouldn't be 1 port cloak (I already know he's not proxy because I have the APM to search the whole map) except still make queens as 2 port is in his range. I will have to experiment with this for a long time.
3) Gas first:
For gas first, you just have to be aware of how much faster his tech will be out. If his rax finishes at 2:53 and already has gas then enemy can 11 gas 12 rax into 4:21 reactor hellion (17 seconds earlier than normal). I'm not sure of the optimal defence against this as you have to make 2 spines so early if you're doing ling spine queen as a hold. Gas first banshee comes about 30 seconds earlier, 11 gas 12 rax banshee finishes 6:00. So if you've eliminated gas first reactor hellion and 1111/1112/1212 reaper with the drone you've left in his natural, then you need extra queen production before 6:00 and detection soon after. It might be simpler just to steal.
4) 2 rax:
The 12 rax 2:35-2:38 timing is important because if terran is 2 rax and the first rax finishes non-trivially before 2:35 you know you need to invest much more in defenc(s?)e because he's more all in (11/11, 11/12, whatever).
This timing allows you to tell what type of 2 rax it is, or whether it's 1 rax (12 rax 0 gas), in close air meta/shattered without even checking the front - which can actually afford you some pretty awesome close air OL scouting patterns, consdering some good terrans rely on the OL to go to the wall position after bouncing off the CC. Which leads me to:
5) General close-air position 0 gas:
a) 2:38 OC morph with 14 scv (or 13 before morph) is 12 rax 0 gas, 1 rax into CC or 1 rax into mass rax.
b) 2:38 OC morph with 13 scv (or 12 before morph) is 12 rax 14 rax economic 2 rax.
c) No OC at 2:38 with 15 SCV at 2:38 (or 14 with him scouting) is CC first.
d) No OC but 11-12 SCV at 2:38 usual SCV finish timing hes 11/11 or 12/12 or 11/12. If he has 13 SCVs just proceeding 2:38 then he's 12/13, expect OC morph at around 2:42-2:45 to validate this. 11/12 has the OC morph at 2:48 with 2 marines out of each rax and 13 or 14 scvs in total (i forget the exact number, this number includes the 2 OCs proxying the rax), so this is what you're looking for as your overlord leaves in this situation to cement the read in your mind (even though you already knew before hand).
6) Bottom ramp wall-in:
If your opponent is doing the bottom ramp wall in and all you see is 2:38 rax finish, you need to look for 2 pieces of information. Don't get concerned about 2:40 rax finish, some ramps are quite far away. The first marine pops at 3:05 and the next at 3:30, and so on. If he reveals a small batch of marines early on, COUNT THEM, and see if it's the correct number for 1 rax. If he just shows 1 rax worth of marines, this of course doesn't mean he's not 2 rax or 4 rax or something stupid, but it does help you narrow his build range and you are playing towards a more defined and accurate distribution which only serves to improve your winrate assuming you react appropriately to his range. But really you're looking for 2 rax worth of marines. And this is for all you guys who have been criticising the legitimacy of these timings.
Idra's stream about 2 months ago. He sees 3:37 marine running down enemy's ramp. He didn't know this timing (I must presume), so he says it's probably 2 rax (but he did express that he wasn't 100% sure). This is another reason why these timings are vital. Random stuff like this that can screw you up 1 in every few thousand games even if you're incredible.
The second piece of information is whether he cuts marines at 3:05. If so, given my experience, 12 rax 16 command dominates his build range. Which leads me to the following discussion:
7) A problem I haven't figured out yet:
The puzzle I'm now trying to get past right now is what to do in the face of a terran who ceases marine production at 3:05-3:08 (so he makes 1 marine). I have learned that 4-6 rax can follow off this but it is rare. My goal in the coming weeks is to find out when 17 marine starts after 16 command is laid down. (So how long it takes 15 scv and whatever portion of the mule to mine 450 mineral with average scv stacking). Not a 100% reliable tell, but again helps in my goal of narrowing the range i assign to my opponent, and it's good that super high level tricks like this (making 17 marine at exacting time when he's doing 4-6 rax) isn't part of the meta game, and probably will never be because terran has to assume zerg is the biggest sicko nerd ever.
Conclusion:
I'm actually so excited about what I will learn in the next few months about zvt scouting!!! So much to learn! Especially w.r.t. builds that follow 12 rax 16 command. Only timing I know is the earliest blue flame possible off 12 rax 16cc into quick double gas is done at 7:55 (which happens to also be the time t hat the 5th and 6th hellion is finished), so I always make sure to have at least 2 spines done in my nat if he shows CC, only marine in natural, and especially if no gases are taken and his marines only have 45 hp (albeit I dunno the terran range that well when he doesn't show techlab on wall in rax, so i take this last tell very lightly. This just stems from my inexperience).
ZvP scouting
Despite ZvP having more timings to learn, timings aren't as important as his range is betrayed by the composition he shows in comparison to nexus and usually this with general gas timings and chrono allocation is enough. However, timings can be used to objectify scouting. Something like "Oh, I feel like he should have extra 3 sentry warp in 20 second ago and not now given he 2 gas normal time" is not an accurate way of going about things, and it requires conscious thought. Timings eliminate conscious thought. 5:09 nexus. BAM. You KNOW he is 1 gate expansion with only 32 or 33 supply in probes and no 34 pylon. There is 0 thought. You KNOW his range is dominated by this 1 build (You've scouted 14 gas 19 gas). There is no thinking and no second guessing. An individual who sticks to experience and intuition will be like "hmm.. 1 gate expo or 1 gate twilight into DT", because nexus timing for both these builds is very similar and both nexus timing is earlier than 3 gate expansion. BAD, even if 1 gate twilight nexus DT is much more uncommon vs 1 gate nexus. And please don't be like "but if 5:20 nexus u coudl get tricked!", no you simply use your brain and know that either build is possible, only 5:09 necessitates 1 gate nexus, 5:20 does not necessitate anything just leaves his distribution with more elements.
First, the bare fundamentals:
1) Cyber core:
- You first look at his cyber finish timing. This in turn leads to the deduction of gateway timing. If the core finishes at 3:42-3:50, then he did a 13 gate with 1 chrono preceding it. 1 chrono 12 gate cyber finish timing is slightly early as the gateway finishes at 2:42 instead of 2:53. Why is this important? Easy. If 14 or 15 gate, don't you think you should know? What if he slips in gas first, are you just going to NOT notice? What if he does a 10 pylon 10 gate, notice! Timing is the efficient way of noticing, comparing to drone count is bad because every map has different drone run distance so you have to learn relative timings for what 8-10 maps, learning 1 timing is better. This timing was most important in previous patch, because he could have wg research done at 4:55 if hes doing the 10pylon 10gate 50-gas 18 probe zealot 4 gate (a script that has everything perfect and there is no 9 scout and builds everything close to nexus could actually complete research at 4:45!) or 5:15 for 20 probe 1 gas 4 gate but with the wg nerf it's not as crucial. At present it's most important in order to see if he's gas first. So the core finishes significantly after 3:50, what do you do?
2) Checking gas:
Check his first gas. Click on it. At 4:00, it should have 2355 if he went 9 pylon 1 chrono 13 gate 14 gas. But what if your drone isn't there at 4:00? Gas geysers mine at 116/minute (1-2 different for more efficient geysers, such as the far one on metalopolis by the way!). So if you see this at a different time just click and do the maths. Don't worry, just before you actually get in this situation do (2355) - (116/2) = ? and this is your 4:30 amount. Boom! You've solved a big part of gas first. Yes, I realise it is very rare among good protoss. It is still good to have. Why? Some players like Huk will just not get gate for ages until he knows you're pool first and not hatch first. Others, MC doing DT for example, will gas first your ass. You need to tell the difference as you have to play completeeeelllllly different vs both options.
After this there's not much more to do with the drone. You of course do basic thing like check 2nd gas timing, and chrono save. If you want to save drone just leave via mineral walking before zealot+0:35 or zealot+0:25 with 1 chrono on gate after zealot (this includes a 2s safety window. If he's stalker you have an additional 5 seconds).
3) Nexus timing:
After this point, you're mainly relying on your lings outside his base. For this portion of scouting, you have
- 5:31 nexus is really the earliest 3 gate expo. You have 33 in probes, no 34 pylon.
- 5:49 nexus is what you saw quite often in the last patch, where you have 34 supply. I really don't know what causes the 18 second disparity. I forget if 5:49 nexus has the 34 pylon before it.
- 6:09 is the latest 3 gate nexus you're going to see unless you keep all inning him or something. It's 34 pylon finished, 37 in probes, and 3 sentries made out of the first gateway.
All these compose the realistic 14 gas 19 gas timings 3 gate expansion nexus timing range. Before and after are possible, but quite rare, and you will know anyway.
I'm not saying that 6:09 nexus is 14gas 19gas 37 supply 3 gate expansion. No. I'm saying this information is needed as you will have to cross reference future data about warpins and unit composition back to his nexus timing to know what buildings he has in his main and thus how many drones you are allowed to make etc.
4) Special situation 1: Stargate and blink cheese.
After 6:10, if there's no nexus and he's 2 gas, you become very suspicious. After this timing, the following are common gas cheeses that you will face off 14,19:
- 6:45 blink research finish, full chrono. If you don't drone scout, or you leave his base at zealot+0:25 because you don't want to die to a chronod sentry, he can have blink research finished at 6:36 because you not being there affords him an auto twilight. This does not change anything as he's not going to have 4 stalkers before 6:50 anyway. It doesn't actually increase the power of the rush in any way so you the earlier twilight doesn't change the amount of units you need.
- 7:10 4th phoenix leaves stargate, full chrono
Unfortunately I know not the timing for 1 base VR plays, such as 2 gate 1 void ray expansion (i.e. when the voidray pops) or 1 void 3 phoenix or 2+ voidray 3-5 gate cheese. But we'll see that it doesn't matter.
If he is 14gas 18/19 gas, drone see saved chronos, and no chrono on core, and no nexus: he's one of these two builds a huge amount of the time disregarding warpprism and other uncommon cheese (He could be another variation of stargate that has nexus such as 2 gate voidray but my suggested response doesn't, and CAN'T, change, due to the necessitation of units due to blink remaining in his range). I'm not sure the optimal response against the above information. I'm feeling mass ling, as this is good vs both elements of his range and is necessatated pre-6:45-6:55 anyway. I'm also not sure on whether I should get a spore in each base or 4 queens when I know he's either blink or stargate. I'm not sold on mass droning after 6:58 when you "know" he's stargate doing a stargate all-in or semi all-in, as you could be playing someone who just waits for 8 stalkers before moving out, or you may have missed a proxy pylon.
He's not DTs. DTs use chronos on nexus and then core. Further, the vast majority of DT builds involve nexus because it adds trick value, you have excess minerals, and you're still in the game if your DTs do damage but don't cripple.
5) Special situation 2: DTs
These usually involve nexus as you can afford it. Nexus can be very early or very late, and it all depends on whether P is 1 gate twilight nexus into 3 gate DT or 3 gate twilight nexus into DT. Odd nexus timing is not a necessary condition for DT though. The tell of a DT rush is the lack of 3rd-5th sentry.
a) A 14 gas 19 gas can fit 2 sentries and still have gas for 2 DT warp at 7:10. If your lings and initial drone see 2 chrono on core and hes doing usual 3 gate expo you would expect 3rd-5th sentry warp at 6:20 assuming he's not freaked about anything. So what you do is you keep checking for this and if not you put evo at 6:25. This means you have spore finished in both bases at 7:30, and you will never die to the DTs in this scenario.
b) A 14 gas 19 gas without sentry OR stalker can have 3 DTs warp in at 6:45. You'll know, however, as he may try to fill in with a stalker made past 4:45, which is way too late. Further, this scenario will never happen if you drone scout because he can't auto twilight due to nothing denying it and your lings will check everywhere for proxy.
c) A 14 gas 17 gas can have 1 sentry 1 stalker and 2 dts warped in at 6:55.
You've cleared everywhere of proxy pylons so you a lot of the time because he has to run the DTs from his base to your base. (A worker takes 0:53 to traverse xel naga to give you some perspective). Sure he can hide sentries, but you have to respect the information.
Stopping DTs with 100% success is pretty simple, actually. If he's 1419 and he makes the initial 2 sentries, and he doesn't warp sentries in front of you at 6:20 off 2 chronos (adjust by subtracting or adding 0:10 according to chronos on core) then make an evo 5 seconds after. If he's c), 1417 and shows a stalker and a sentry make an evo at about 6:10 without waiting for the 6:20 (off 2 chrono) warp in, and simply cancel if he warps in sentries at 6:20 (off 2 chrono, adjust accordingly).
If he's b), then you're in this situation because you didn't drone scout. I don't know how to stop this one without drone scouting, as he can autotwilight and 2 DT at 6:45 with zealot followed by stalker. I'll have to ask a better, non drone scouting player about this.
If he hides sentries, you have to bear the burden of lost minerals. I don't know any lair timings for this because I'm a later lair kinda guy.
6) Unsolved problems:
How to distinguish 6:09 nexus from 4 gate with 33 in probe (cancel strat) to 3 gate with 37 in probe (non-greedy 3 gate expansion). I'm thinking if there's no building wall I should be biased towards cancel. [Plz don't get all nazi about 6:09. I know 6:09 doesn't necessitate anything. But I do know that a 6:09 nexus has a range composed of these 2 exact builds which is why I'm using it]
I still haven't solved 3 gate nexus into 5 gate cancel or 1 gate expansion into 4 gate cancel, but I'm sure certain timings and composition on the building wallin at the natural should cause me to alter the probabilities I assign to the elements of his distribution, I'm just not sure how and to what extent.
A complete forge FE deconstruction will come in a few months.
So much to learn!
Arbitrage.505
|
I take it there's like a 0% chance of these accuracies translating to the lower leagues?
Either way, excellently written, and boatloads of information, thank you.
|
Everyone can learn openings, even silver players 
Thanks for the well-written post and information
|
On July 07 2011 14:19 Asday wrote:
I take it there's like a 0% chance of these accuracies translating to the lower leagues?
Either way, excellently written, and boatloads of information, thank you.
Knowing these timings will not help you if your opponent is not executing their build "properly."
Good read overall. I agree with most of your zvp analysis. Checking how much gas has been mined if you scout protoss last is smart. I'll incorporate that into my PvT more often if I scout late and cannot determine if they went gas first.
Thanks for the post.
|
On July 07 2011 14:53 JonnyLaw wrote:Show nested quote +On July 07 2011 14:19 Asday wrote:
I take it there's like a 0% chance of these accuracies translating to the lower leagues?
Either way, excellently written, and boatloads of information, thank you. Knowing these timings will not help you if your opponent is not executing their build "properly." Good read overall. I agree with most of your zvp analysis. Checking how much gas has been mined if you scout protoss last is smart. I'll incorporate that into my PvT more often if I scout late and cannot determine if they went gas first. Thanks for the post.
If you are playing fine yourself it still helps, because it means that he really cannot have X when he went for Y (for example a protoss can't have a lot of stalkers early, when you see a lot of zealots)
Imo you should always assume that your opponent is exectuing everything perfectly. (not missing workers, not getting supply blocked, his build lines up perfectly, his micro will be perfect, his FFs and blink will be perfect...)
One thing I want to add: Even when you know that stuff, PvZ will be a coinflip most of the time, because any good protoss will not show more than 1-2sentries and 1zealot, which could be nearly any build. sacrificing an overlord and scouting the front at the same time is necessary to get a slight glimps of his sentry/stalker count (he has to send at least 2sentries/1stalker for the OL and then you will see sentries/stalkers at the front to kill your zerglings.
|
+ Show Spoiler +On July 07 2011 15:15 Big J wrote:Show nested quote +On July 07 2011 14:53 JonnyLaw wrote:On July 07 2011 14:19 Asday wrote:
I take it there's like a 0% chance of these accuracies translating to the lower leagues?
Either way, excellently written, and boatloads of information, thank you. Knowing these timings will not help you if your opponent is not executing their build "properly." Good read overall. I agree with most of your zvp analysis. Checking how much gas has been mined if you scout protoss last is smart. I'll incorporate that into my PvT more often if I scout late and cannot determine if they went gas first. Thanks for the post. If you are playing fine yourself it still helps, because it means that he really cannot have X when he went for Y (for example a protoss can't have a lot of stalkers early, when you see a lot of zealots) Imo you should always assume that your opponent is exectuing everything perfectly. (not missing workers, not getting supply blocked, his build lines up perfectly, his micro will be perfect, his FFs and blink will be perfect...) One thing I want to add: Even when you know that stuff, PvZ will be a coinflip most of the time, because any good protoss will not show more than 1-2sentries and 1zealot, which could be nearly any build. sacrificing an overlord and scouting the front at the same time is necessary to get a slight glimps of his sentry/stalker count (he has to send at least 2sentries/1stalker for the OL and then you will see sentries/stalkers at the front to kill your zerglings.
I agree that not underestimating your opponent is great. It's easy to lose games by assuming your opponent is bad when you see that their timings are off. Really, they're just cheesing you.
But, exact timings and gas count revealing builds in lower leagues will not give you much information. Seriously. It could be gas first, no warp gate double stargate or 5 gate sentry...you cannot know.
|
One thing I want to add: Even when you know that stuff, PvZ will be a coinflip most of the time, because any good protoss will not show more than 1-2sentries and 1zealot, which could be nearly any build. sacrificing an overlord and scouting the front at the same time is necessary to get a slight glimps of his sentry/stalker count (he has to send at least 2sentries/1stalker for the OL and then you will see sentries/stalkers at the front to kill your zerglings.
Many times you don't have to sack OL. Just make an evo then cancel when he shows the sentries if that's the part of his range you're concerned with. And to say it "could be any build" is correct but inaccurate. Gate timing, nexus timing chrono allocation, gas timing, natural wall composition and timing, will all allow you to construct an extremely accurate range with very few elements, and in many instances you don't need to see these units to play catch all without being behind to the most economic element of his range.
|
Fantastic post. I'm only a Bronze level Terran, so a) I know what to benchmark myself at; b) I know what to look for in my matchups now; and c) I know how to hide units and buildings to confuse scouts :p. Thanks for an extremely well-written post!!
|
i've been thinking about doing this myself, GJ OP!
The biggest thing other than nexus timings is counting sentries. If he has a zealot a stalker and a sentry when he should have 3+ sentries, you know something is up, make 2 spores.
Cancel 4gate is pretty easy to distinguish. The nexus goes down later than 6:20 (6:40 in my experience) and he has few sentries. What you do is have a ling scouting _everywhere_ for proxy pylons and start pumping lings past 24 drones while putting up a roach warren.
|
Oh my god. Giggle.
I am so pumped about this thread.
I'm gonna go over my replays, see where these timings are accurate and start memorizing the timings. Excellent work OP.
|
i don't even play zerg but just wanted to say SICK guide!!
|
Wow. Very well done OP
|
I am blown away. I've always told myself, I play pretty well but the one thing I don't know are timings. I read this and my problem is solved. Not only that, a whole new way of looking at the game is revealed to me. I've always played with intuition and a feeling of "flow". This precision is wonderful.
|
The 4 gate from nexus cancel is easily scouted by seeing a nexus going down very late, past 6 or 6:30. If the P decides to 1 gate expand cancel THEN put down the gates, the push will come much, much later. The 5 gate off of a 3gate expand cancel shouldn't be an issue - the timing for that is easily scoutable with a lower sentry count and a very late nexus (in which case you suspect something is up anyway).
|
On July 07 2011 20:05 panzzzzz wrote:
The 4 gate from nexus cancel is easily scouted by seeing a nexus going down very late, past 6 or 6:30.
So how do you distinguish 13 gate 33 probe 4 gate 6:10 nexus from 13 gate 37 probe 3 gate 6:10 nexus?
The 4 gate cancel nexus goes down at 6:30 if the P is doing it wrong. Watch MC vs July. 6:10. Same as 37 supply 3 gate nexus.
If the P decides to 1 gate expand cancel THEN put down the gates, the push will come much, much later.
The 2nd-4th gate goes down before the cancel.
The 5 gate off of a 3gate expand cancel shouldn't be an issue - the timing for that is easily scoutable with a lower sentry count and a very late nexus (in which case you suspect something is up anyway).
No warpin at 6:20 off 2 chronos on the core can mean DTs at 7:10 with 2 sentries and 14g19g. Are you going to make pure units against DTs and he won't even be behind when it fails? What if he warps into the main, and you've stopped drones at like 32? Besides, I dispute your claim that you can't warp in 150 minerals worth of sentries at 6:20 with this build. Do you know this as a fact?
|
I think differentiating between dt and other tech allins is pretty important. Mass ling is good against blink (obviously) and stargate, because even if he follows up with an expand you can easily crush his low-unit, low-sentry army. DTs however can transition into a 1 base archon+zealot allin and I think you need roaches to defend that, or a bunch of spine crawlers. either case you need to be droning early.
If you cannot tell which one he's going for, the safest reponse is probably to save some larvae and put down a few spine crawlers.
|
Even though you've put very much thought into this...the exact timings are a bit ridiculous. I mean 5:09 Nexus for one gate expo. Really? I've seen MC do one gate expo at 5:14 and I've seen HwangSin one gate expo into stargate (!) at 4:55.
Breaking it down to "seconds" seems to overdo it.
Therefore you focus too much on very specific builds. For example, what I've been playing around with recently is, chrono-boost the gate for sentries, then go nexus off one gate and then go either stargate or TC. Delays DTs by about one minute (they are out at 8:30) but makes the build look exactly like a one gate expo into 5 gate aggression, blink or w/e. You've said "if toss doesn't build 3rd-5th sentry". The DT build I use gets 4 sentries.
Your write-up is an excellent guide to use certain timing-signals to structure your own strategies, but trying to "reverse engineer" a very specific protoss build isn't really possible when there are so many deviations floating around...
|
HwangSin one gate expo at 4:55.
1 gate cyber with 2 sentries and 32 in probes is the 5:09 timing. Did hwangsin have 2 sentries? If P has 1 zealot 2 sentries and nexus at 5:10 you know he's 1 gate expansion. In response to your comment that MC 1 gate expod at 5:14 as an argument for why my nexus timings are "ridiculous", I refer you to my added addendum to the 5:10 nexus discussion that a 5:20 or so nexus does not mean anything specific, his range is more diverse. A Z who sees a 5:10 nexus and knows this timing is undeniably informationally ahead of a Z who uses generalized timings and not specific timings, as this Z knows that P can only have 1 gateway and not a twilight. This cannot be debated. The knowledge that his range consists of ONLY 1 gate 1 core is far superior to the zerg who thinks that his range STILL consists of 1 gateway cyber twilight before nexus. If I was to only learn general timings I would not know that he does not have a twilight yet, et cetera. So, 5:10 is not ridiculous and it results in the ability to possess a more accurate range against this build.
All these timings lead to the zerg possessing a far more accurate distribution of builds that the Zerg is thinking up in his head during the game. Every player assigns a range. A Z who knows what 5:10 means has a better range than a Z who does not.This leads to superior decision making, be it on the spot decision making or rehearsed decision making in response to certain pieces of information..
Same with all the other timings.
Therefore you focus too much on very specific builds. For example, what I like to do is, chrono-boost the gate for sentries, then go nexus off one gate and then go either stargate or TC. Delays DTs by about one minute (they are out at 8:30) but makes the build look exactly like a one gate expo into 5 gate aggression. You've said "if toss doesn't build 3rd-5th sentry". The DT build I use gets 4 sentries.
It's common sense that you keep scouting with lings. "no 3rd-5th sentry" is supposed to be interpreted within the context of 6-7 minutes. I expect Z to be toggling back to lings every 6 or so ingame seconds to check for expansion gas timings, wall in timings and composition and additional units. If you only have 4 sentries I will know later DTs remain a huge part of your range and I will always have spores no matter what. My discussion of DTs was w.r.t the very tight timing of quick DTs where you need an evo at about 6:30, which you NEED specific timings for so you don't have to guess.
If P plays Z and is doing the nexus 2 sentry DT expand, and this Z is consciously using these timings, P will catch Z offguard 0% of the time! That's a pretty good success rate against this very common DT build I'd say!
Q.E.D.
|
You misinterpreted my statement - of course, if zerg sees a 5:09 nexus he knows that it's one gate. But that's also the case for a 5:08 nexus, a 5:10 nexus, a 5:11 nexus....you get my point?
Even small stuff like 9 scout to 13 scout, 13 gate to 14 gate, 17 core to 18 core, 19 gas to 20 gas (one additional chrono-boost on probes before assimilator) can mix up timings. I wasn't critizising your interpretation of the EXACT timings at all, I was critizing your approach of USING these exact timings without putting enough emphasis on the possible "time-stretch" of the respective build the toss player could be using.
It's not the 5:09 timing that players should remember. Players should learn the CUT OFF-point for a one gate expo compared to a 3 gate. Meaning, the pattern of thought shouldn't be "I'll check at exactly 5:09 for a nexus" but "I'll check for a nexus and if he hasn't built one until 5:30 he's probably not one gate expo-ing".
Later on, everything becomes much more vague. I mean, you said you'd expect DTs if you only see 4 sentries. Now what if I just had another bunch of sentries in the back and you just didn't see them?
EDIT: probably my first post came off too harsh; you did a good job deconstructing possible protoss techs, my main point was, that you could've simplified it by focussing on the timing"range" instead of certain points in time.
|
Later on, everything becomes much more vague. I mean, you said you'd expect DTs if you only see 4 sentries. Now what if I just had another bunch of sentries in the back and you just didn't see them?
This bit is in my OP. If you hide sentries then I've wasted money on spores. As I said my goal in the game is to construct a range, and if you only show 4 sentries a large part of your range is DTs. I have to respect the range I'm assigning to you, where DTs form a large part and make spores.
Even small stuff like 9 scout to 13 scout, 13 gate to 14 gate, 17 core to 18 core, 19 gas to 20 gas (one additional chrono-boost on probes before assimilator) can mix up timings. I wasn't critizising your interpretation of the EXACT timings at all, I was critizing your approach of USING these exact timings without putting enough emphasis on the possible "time-stretch" of the respective build the toss player could be using......my main point was, that you could've simplified it by focussing on the timing"range" instead of certain points in time.
Hmm I tried to outline the timestretch, in saying that 5:31-6:10 is the realistic 1419 3 gate sentry expand range with 2 sentries preceding nexus (and 3 in the case of 6:10 nexus) and then adducing for this claim by showing the different combinations of probe and sentry supplies that lead to different locations along this range. I also said that a 5:09 is definitely a 1 gate expansion given 1419 and 2 sentries but 5:20 may or may not be, which was my way of discussing the vague 5:15-5:30 nexus placement time which is the range 1 gate and 1 gate twilight both occupy. I didn't mean to say "5:09 is THE 14g19g 1 gate expansion timing with 2 sentries and 32 supply." of course not what mumbo jumbo. I could've said "5:08-5:13 if he has 32 supply and 2 sentries" but I expect people to have a bit of common sense when they hear me drop times. Apologies for being a bit confusing there.
|
Wow. Ya know, it's nice to finally have all these things written down as I could never remember them.
Edit: What about on a map like Tal Darim Altar where, if I'm not lucky with my drone scout and I can't get in to check gases, I only see the rax and marines? What would his range of builds be from there?
|
On July 08 2011 00:46 Amaterasu1234 wrote:
Wow. Ya know, it's nice to finally have all these things written down as I could never remember them.
Edit: What about on a map like Tal Darim Altar where, if I'm not lucky with my drone scout and I can't get in to check gases, I only see the rax and marines? What would his range of builds be from there?
Perhaps a better player should answer this question. My main piece of advice would be to make sure that the next marine that pops out finishes at one of the 3:05 3:30 3:55 timings so you can know that it's a 12 rax you're looking at. No reaper, 1 rax and constant marine production are all tells and does have consequences for his range. You can already eliminate 12 rax 16 cc, 1 rax reaper, 1 gas 1 fact reactor hellion, and if he constantly makes marines numerous 6 rax variations. But as for your question with regards to his range and how we should play out this game I will not try to guess at this or give advice. Such a tough question.. and one for a much better player.
|
Not only is this a great thread with real specific timing examples but for me personally, the best part of this thread is thinking about your opponents buildorder in terms of ranges.
When we start thinking in ranges we can use game theory to formulate optimal responses against those ranges. We can build decision trees and prioritize information in a game of incomplete information. We can put numbers on concepts that look very abstract on the surface and use math in the game of Starcraft to improve our winrate.
Although Starcraft is a game were an almost infinite number of variables influence our decision making we're still able to improve on our decision making by putting these variables in a model (like you did here). The model will always be flawed but with the growth of the game and the number of games played steadily increasing, these models should become more accurate over time and even a flawed model is often better than no model and playing purely on instinct and experience.
Of course this is exactly what good players are doing already (whether it's consciously or simply because of experience) but to formulate these ranges and their corresponding optimal response we have a way of sharing this experience and making concepts like 'game sense' actually something we can learn. These are all steps in improving our knowledge of the game and maybe making the steep learning curve of Starcraft a little less steep.
Thanks for the realization! Now i can't wait to dive into some replays and extract some statistics to formulate my own models.
|
Thanks for all this wealth of information.
This all overwhelms me. Does the average masters/gm Zerg have all this stuff memorized?
I'm a high diamond zerg. I have never used in-game time in my scouting/reacting habits. Other than making queue-cards and straight memorizing this stuff for hours... whats the best way to incorporate this knowledge into my gameplay? Are there more important timings than others to learn first?
This is probably why you put "Advanced" in the title xD
|
Great thread, it helps my PvT too seeing the timings that Terran hits against Zerg.
Also, in regards to your ZvP scouting and Dark Templar, I can get 3 dts, (with 6 zeals and 1 stalker) at 6:45 and expand soon after, and in this case the expansion is planned but the zealot count chases all lings scouting natural out, so it's DT before expansion instead of expand --> DT and no scouting on if I've expanded or not (speedlings can find out, suicide lings can find out). However, a problem with DT first is it's more risky for both players, and most importantly, any DT before nexus strat (10p10g12gas15core17gas17pylon) can be significantly delayed by drone scout micro since they will wait to place TC before drone is pushed from base, at about 4:15...
Also, regarding "DTs do damage but don't cripple. I'm not sure the optimal response against the above information. I'm feeling mass ling, as this is good vs both elements of his range and is necessatated pre-6:45-6:55 anyway" mass ling is bad since P will pull his Dtemplar, group his zealots and warp in 3 sentries and have 1archon6zeal3sentry1stalker to defend. It completely mauls mass ling unless you catch in the open, but DTs can run back to base unharmed and the rest of the force never leaves the base anyway -- I suggest roach because P has to wait to push once he morphs to archon so that gives you time to get roach
|
So the way you see the game. Is the future. Personally my timings in ML are purely just feeling and experience. The stuff you are pointing out seems like GM material and if a player truly takes good macro/micro then using their additional APM to notice all the things you talked about in the OP, then that is the difference between playing casually and playing professionally. IMO
Not to mention all that info comes without the overlord that dies to the sentry but retrieves no information.
|
On July 08 2011 03:13 djengizz wrote:
Not only is this a great thread with real specific timing examples but for me personally, the best part of this thread is thinking about your opponents buildorder in terms of ranges.
When we start thinking in ranges we can use game theory to formulate optimal responses against those ranges. We can build decision trees and prioritize information in a game of incomplete information. We can put numbers on concepts that look very abstract on the surface and use math in the game of Starcraft to improve our winrate.
Thanks for the realization! Now i can't wait to dive into some replays and extract some statistics to formulate my own models.
...is it possible to actually make something like a mind-map as a visual of sorts? 'Cause not only would that be just amazing, it'd be infinitely helpful to every zerg player ever!
|
I think something important to note when playing vs T and seeing their production facilities is to check if they're building something or not (lighting up). Maybe we'll see some ridiculous high level mind games where players build a unit and cancel last second to create the illusion that they're producing units and expand instead.
|
Very nice write-up, though I must say it won't do much good for lower league players. They usually won't be doing exact builds and would more likely have the mentality of one thinking 'Hey, this unit looks cool, I want to make a lot of it!'
Can you please create a thread giving the timings of Zerg builds, to complete the set?
|
Is there anything like this for Terrans to know the times of Z and P?
|
With 12rax 13gas gas finishes a few seconds after 14th scv comes out (13th in mineral line) but much closer to 14th scv than rax completion time.
|
Very thorough analysis, thanks so much for taking the time to organize this and share.
|
On July 08 2011 05:26 Perfect wrote:
Is there anything like this for Terrans to know the times of Z and P?
I wouldn't know about P, but for Z, everything's way more generalized, since Zergs USUALLY react to what T and P is doing. Ofc there are times that we initiate the pressure ourselves, but not too often. The tells for Z are fairly easy:
- Check Pool timing, gas timing, expo timing. It's fairly obvious whether he's opening 14g/14p, 11p/18 hatch/ 15 hatch ? pool, etc. You could get into seconds but it doesn't really matter.
- Check when Ling Speed kicks in. Delayed Ling speed could be a hatch first opening with no speed, or fast Lair with no speed. If Lings are still speedless around 7 mins in, Z is either going roaches, roach/bane, fast Lair, Spanishiwa, or just forgot speed ^.^ 1-2 scans will tell you exactly which. You should probably scan around 6:30 for fast Lair (with speedless lings), if you don't see it, expect heavy Ling/Bane/Roach aggression. If there's no roaches or banes, then it's Spanishiwa build.
- Depending on Lair timing, scan at either 8 min (fast Lair) or 9-9:30 min mark for Infestation Pit/Spire.
I mean... Terran has scans. You should never feel blind in early game since you're dictating the early game.
|
On July 08 2011 05:44 Flonomenalz wrote:
I wouldn't know about P, but for Z, everything's way more generalized, since Zergs USUALLY react to what T and P is doing. Ofc there are times that we initiate the pressure ourselves, but not too often. The tells for Z are fairly easy:
- Check Pool timing, gas timing, expo timing. It's fairly obvious whether he's opening 14g/14p, 11p/18 hatch/ 15 hatch ? pool, etc. You could get into seconds but it doesn't really matter.
- Check when Ling Speed kicks in. Delayed Ling speed could be a hatch first opening with no speed, or fast Lair with no speed. If Lings are still speedless around 7 mins in, Z is either going roaches, roach/bane, fast Lair, Spanishiwa, or just forgot speed ^.^ 1-2 scans will tell you exactly which. You should probably scan around 6:30 for fast Lair (with speedless lings), if you don't see it, expect heavy Ling/Bane/Roach aggression. If there's no roaches or banes, then it's Spanishiwa build.
- Depending on Lair timing, scan at either 8 min (fast Lair) or 9-9:30 min mark for Infestation Pit/Spire.
I mean... Terran has scans. You should never feel blind in early game since you're dictating the early game.
Great! I keep losing to this ~12 min Mass roach after normal opening. I finally started to notice that when Z does NOT go for a gas around pool timing they usually end up doing a huge roach push, but was not 100% sure on that. I will also look out for these other tips.
|
how do you scout a 12 drone rush.
|
Great writeup! I'm looking forward to the addition of FFE since that's been giving me the most trouble lately.
|
On July 08 2011 05:53 Perfect wrote:
how do you scout a 12 drone rush.
You don't, it's largely irrelevant on the current maps. The only reason it was ever useful is because Steppes of War had a rush distance that was ridiculous.
I'm not saying those close positions don't exist still (thanks for not fixing Metalopolis), but they're on 4-player maps, and going for a 12 drone rush on a 4-player map is throwing the dice on a 33% win rate.
|
|
|
On July 08 2011 04:23 tehemperorer wrote:
Great thread, it helps my PvT too seeing the timings that Terran hits against Zerg.
Also, in regards to your ZvP scouting and Dark Templar, I can get 3 dts, (with 6 zeals and 1 stalker) at 6:45 and expand soon after, and in this case the expansion is planned but the zealot count chases all lings scouting natural out, so it's DT before expansion instead of expand --> DT and no scouting on if I've expanded or not (speedlings can find out, suicide lings can find out). However, a problem with DT first is it's more risky for both players, and most importantly, any DT before nexus strat (10p10g12gas15core17gas17pylon) can be significantly delayed by drone scout micro since they will wait to place TC before drone is pushed from base, at about 4:15...
Yea, against 1417 without sentry it's auto evo and you'll have spore in time no matter what. 10p10g will actually confuse me in a real game though, but when I see you haven't made a sentry I'll be ready because gas cheese is such a large part of your range.
Anyway you can't get a 6:45 DT warp if I drone scout because you can't autotwilight. For this you have to rely on me not drone scouting or a successful proxy. Also 6:45 is the optimal autotwilight off 1 chrono 12 gate, 14g17g. with 10p10g you should be able to get it earlier.
Also, regarding "DTs do damage but don't cripple. I'm not sure the optimal response against the above information. I'm feeling mass ling, as this is good vs both elements of his range and is necessatated pre-6:45-6:55 anyway" mass ling is bad since P will pull his Dtemplar, group his zealots and warp in 3 sentries and have 1archon6zeal3sentry1stalker to defend.
The information I'm talking about is the information that you're either
- blink
- stargate w/ or w/out gateway
Which is also the information that you're NOT DTs.
Is there anything like this for Terrans to know the times of Z and P?
Search "timings" in TL search function. There's a whole list of Z timings. Only problem with that thread is it hasn't been integrated into a nice easy to use model for you. All I know is that speed off 14g14p finishes at 5:02.
|
I didn't like this guide personally. I found it very convoluted and/or just plain poorly written.
I don't understand why 2 people said it was well-written.
How is this good English?:
On July 07 2011 14:08 arbitrageur wrote:
Despite ZvP having more timings to learn, timings aren't as important as his range is betrayed by the composition he shows in comparison to nexus and usually this with general gas timings and chrono allocation is enough.
And that's just the quality of the English sentence syntax, let alone the lack of structuring, and even formatting. Sure, some basic paragraphs are there, a bit of bullets, and titles, which is useful but it doesn't make it great, nor does it fix the problems.
For some parts that I actually did seem to understand the reasoning, I wouldn't even necessarily agree with them. The topic also seemed to cover the game in way too much of an analytical and narrow-visioned method. The problem is that different maps have different timings, different game actions change timings, and different players have their own differences in build style.
Not only that, but oftentimes players will not opt for the most efficient way of running a certain build since it's too much of a dead giveaway of their actions.
Aside from those things, there's nothing too wrong with writing and suggesting to play to analytical, but I would have to say that it's a small minority of people who excel at that. Overall I'd say this goes over too many people's heads, and is definitely not worth the time practicing and training the brain to deal with some of the stuff mentioned.
As a suggestion to the author, aside from just fixing up some structure here and there (and everywhere), I think it would be a good idea to explain and/or introduce various concepts that you talk about in a sort of explanation section (or preface). In addition, there should be some sort of introduction which can be related to, or part of the explanation of concepts, as well as overview (conclusion) paragraphs after a section.
|
I think most of the people in this thread understood the meaning of the OP just fine. Sure there are some errors, but idt we need to get supernerd about it.
The overall message i think is that there are a wide range of builds that just plain kill zerg if unprepared. It would be good to notice a rage of timings for certain tells of these builds so you can tweak your overall plan for that game. I think if we follow day9 advice, we all have a general "plan" on which we base our build. What the OP is saying that there are specific timings we should be aware of and they are of "X-Y" range, specifically at "ABC" time if the T or P build was optimal. No way do i just assume that the timings are for all maps all matchups. I appreciate the effort of the OP and this will definitely help my play rather than hurt it.
The idea is not new, but certain tells i've not seen anyone talk about openly even on TL and that is somewhat of an achievement. I think if we all keep thinking together like this the community can sustain some healthy growth.
|
On July 08 2011 12:46 Xapti wrote:
And that's just the quality of the English sentence syntax, let alone the lack of structuring, and even formatting. Sure, some basic paragraphs are there, a bit of bullets, and titles, which is useful but it doesn't make it great, nor does it fix the problems.
Are you really judging a guide by the use of the English language instead of its content? (And are you really using a poorly constructed sentence to do so?  )
The topic also seemed to cover the game in way too much of an analytical and narrow-visioned method. The problem is that different maps have different timings, different game actions change timings, and different players have their own differences in build style.
Not only that, but oftentimes players will not opt for the most efficient way of running a certain build since it's too much of a dead giveaway of their actions.
How can you become too analytical when discussing a strategy game?
By the way this isn't analytical at all, it's just a list of timings we can use to do analysis on the possible range of builds our opponents are doing. Apparently you don't understand the concept of ranges because by thinking this way it doesn't actually matter that there are variations in what our opponent is doing (because of a million factors including maps and play styles).
Isn't it great that we can deduct that there is a high probability of a proxy rax by looking at a clock and counting SCV's? Sure sometimes there won't be a proxy rax because our opponent missed his timings, is playing mindgames or is just plain bad but this doesn't really matter. We can account for such things by adjusting our ranges and probabilities.
After that we might still conclude that it's a good idea to prepare for the proxy rax even though it's not actually there in some cases. If the probability of a proxy is high enough and we die often enough if we don't prepare we will still see an increased winrate by preparing for it regardless.
What arbitrageur has done here is giving us a framework to construct these ranges based on what we do see.
Overall I'd say this goes over too many people's heads, and is definitely not worth the time practicing and training the brain to deal with some of the stuff mentioned.
Please speak for yourself.
|
These timings are WAY too specific. I am a 1600+ masters protoss player and I consistently have different timings for my 1 gate and 3 gate expands. It really depends on what I was making and how many chronos I have which in turn depend on what I scouted from the zerg player.
Sometimes I go zeal, stalker, stalker, sentry. Sometimes I go zeal, sentry, sentry, sentry. I change this based on the gas timing I scout. The earlier of those two examples makes my 1 gate expo about the same time as a 3 gate expo but with pressure.
Getting locked into these timings is only going to confuse you as a player. Unfortunately there is no quick answer to scouting. You have to be able to read into their build using finer details than the game clock.
|
fascinating. I have to wonder if these numbers will be correct if the opposing player is of a lower caliber. Other than that, this is an incredible collection and a great read. Thank you so much for accruing this data.
|
... personally I just want to give thanks to the original poster.
This is possibly one of the most useful posts here, especially the ZvT section. I've had real problems distinguishing builds from each other and preparing properly for the various 1 base all ins a terran can do (marines, marine tank, cloaked banshee, 2 port banshee, blue flames marine tank followup, marine scv, etc etc etc).
Once the game goes longer and you can scout the composition, I find it much easier. But a 1 base terran really feels too much like a coinflip for me, in that I have a hard time preparing properly for all the various timings / compositions, assuming a proper marine patrol on the edges to prevent scouting slow overlords.
|
Well done. I am Protoss, so please delete this whole thread It will be hard to go DTs when Zerg knows about the timings
Nice collection of timings, friend
|
This has to be one of the most genius and helpful threads out there.
This is going to help me a ton, TY again and please update with any new info! :D
|
On July 08 2011 15:16 djengizz wrote:Are you really judging a guide by the use of the English language instead of its content? (And are you really using a poorly constructed sentence to do so? )
No. I was talking about how some people said the guide was well-written. I already said that in my first post.
Secondly, I was not (superficially) judging the guide in any other way— I read it, hence I was making a fully-informed personal review of it. Presentation is an extremely important factor in communication.
On July 08 2011 15:16 djengizz wrote:
Please speak for yourself.
No. As a top 1% of players, I understand things that most people do not. As a Starcraft fan (like anyone), I also know that far more than 50% of players are bronze, silver, or gold. In other words, most players (bronze,silver,gold) will not be able to make use of this because of the timing and strategy variance involved in such low-level play. Of course, that's only if they understand it, and are willing to spend a load of time trying to train their brain.
I'm not saying I know what's better for someone. If someone likes this guide then fine they can do what they want with it, but I would not recommend it.
|
|
|
Very nice, well written guide. I've been meaning to find some common timings for these matchups but never got round to, so it's a major help to me. But indeed, i think you should use more range of times instead of specific down to the second benchmarks, but even so most people who will use this should know that there are deviations in timing with every indivudual player.
|
On July 07 2011 14:19 Asday wrote:
I take it there's like a 0% chance of these accuracies translating to the lower leagues?
Either way, excellently written, and boatloads of information, thank you.
you dont need to change anything at a lower level
that timings are just THE EARLIEST possible threat
for scouting its not too much of a difference because you can still just continue to poke until you see what he is doing and even if it is much delayed you still should be able to scout the same things by just constantly poking and being careful how to react to what you see
|
This is actually insane. Combine this with the information from the The Art of ZvT and The Art of ZvP threads and you know a whole lost about those matchups.
How long did it take you to write this guide?
|
Very, very helpful. Posting in this thread so I can easily find it again and use as a reference.
Liquipedia could use something like this.
|
Extremely well written by the OP, and I can see it being very useful for people in all leagues.
|
been waiting for this for a long time.... u r LEGEND!
timings may be a bit too specific so try to have a bit more leeway, and when will this be on liquipedia
|
Excellent initiative! I'll contribute with timings when I start playing again! Well done so far!
|
What a lot of people are wondering, is how to apply these at a lower level, when builds are hardly refined.
The idea is that the time stamps that the OP provides is the EARLIEST possible time at which something mentioned will pose a threat to you or when it will occur, and if you're worried about something in particular, that's when you should cater to it. If it happens much later than the given time then in that case it'll just be a shitty build and you'll be able to crush it relatively easily.
|
That's a pretty sick OP, I'm really impressed and have not read such an interesting post for a long time, hat's off sir.
I can say that the terran part is very accurate and useful. I'm not sure about the protoss part because I don't have any experience playing P, but this is still an impressive post.
|
On July 07 2011 15:31 whoopingchow wrote:
Fantastic post. I'm only a Bronze level Terran, so a) I know what to benchmark myself at; b) I know what to look for in my matchups now; and c) I know how to hide units and buildings to confuse scouts :p. Thanks for an extremely well-written post!!
At bronze level you can't really predict much with scouting. At bronze level you can't really tell for example if your opponent has 3 gateways, that he's going for a 3-gate sentry expand, because at that level people just play pretty much by the "ear", if you will. And you can't really "confuse" your opponents because most of the people at bronze level can't read opponents by early scouting because A) their knowledge about openings and such isn't very broad yet B) their opponent is not following a spesific build, so we're back at the start basically where the opponent is playing by the "ear" and not following a spesific build and there for nullifying the scouting somewhat. Of course you can figure something out, but you shouldn't expect that if you didn't see A, he definitely will not go for B because that would be stupid. Don't overestimate your opponents and expect perfect and crisp play. Scouting at that level is just pretty much consisting just that he's not going for something "cheesy" like dt rush or banshee rush, so that consists just checking does he go for an early double gas etc.
|
arbitrageur you are a smart and handsome scholar who is also a gentleman.
|
On July 08 2011 12:46 Xapti wrote:I didn't like this guide personally. I found it very convoluted and/or just plain poorly written. I don't understand why 2 people said it was well-written. How is this good English?: Show nested quote +On July 07 2011 14:08 arbitrageur wrote:
Despite ZvP having more timings to learn, timings aren't as important as his range is betrayed by the composition he shows in comparison to nexus and usually this with general gas timings and chrono allocation is enough.
And that's just the quality of the English sentence syntax, let alone the lack of structuring, and even formatting. Sure, some basic paragraphs are there, a bit of bullets, and titles, which is useful but it doesn't make it great, nor does it fix the problems. For some parts that I actually did seem to understand the reasoning, I wouldn't even necessarily agree with them. The topic also seemed to cover the game in way too much of an analytical and narrow-visioned method. The problem is that different maps have different timings, different game actions change timings, and different players have their own differences in build style. Not only that, but oftentimes players will not opt for the most efficient way of running a certain build since it's too much of a dead giveaway of their actions. Aside from those things, there's nothing too wrong with writing and suggesting to play to analytical, but I would have to say that it's a small minority of people who excel at that. Overall I'd say this goes over too many people's heads, and is definitely not worth the time practicing and training the brain to deal with some of the stuff mentioned. As a suggestion to the author, aside from just fixing up some structure here and there (and everywhere), I think it would be a good idea to explain and/or introduce various concepts that you talk about in a sort of explanation section (or preface). In addition, there should be some sort of introduction which can be related to, or part of the explanation of concepts, as well as overview (conclusion) paragraphs after a section.
yeah I realise the quality of writing is pretty poor. This post started as a brain dump without any clear structure on sc2sea.com then I went to my sc2 notepad and added in all the timings that I had verified multiple times in progames. I just didn't and don't have the incentive to spend 2-3 hours improving the quality, but i did put in 1 hour making it at least understandable.
Although i must dispute your other criticisms, namely that its overly analytical and too narrow minded. From the perspective of knowing these timings and playing good players, my experience of never dying to dts, never being caught off guard by 1 base stargate all ins, always knowing blink is a huge part of his range when he reveals it, never not nkowing proxy rax is coming, etc, i know how good these timings are when you're playing good players.
The problem is that different maps have different timings, different game actions change timings, and different players have their own differences in build style.
Not only that, but oftentimes players will not opt for the most efficient way of running a certain build since it's too much of a dead giveaway of their actions.
- Maps have the same timings. His range changes because different builds have different levels of effectiveness.
- Provide examples of "game actions change timings" and how this invalidates what I'm saying as something other than extremely useful.
- Do you realise that my approach that I've used through the entire OP - that of assigning a range based on timing tells and other information - is due solely BECAUSE ppl have different build styles?
- Since when does suboptimal execution of a build jeopardize a player who is sticking to these timings? Provide an example.
|
On July 08 2011 15:29 Packeteer wrote:
These timings are WAY too specific. I am a 1600+ masters protoss player and I consistently have different timings for my 1 gate and 3 gate expands.
You didn't read my post, or you forgot it when you wrote this reply.
Sometimes I go zeal, stalker, stalker, sentry. Sometimes I go zeal, sentry, sentry, sentry. I change this based on the gas timing I scout. The earlier of those two examples makes my 1 gate expo about the same time as a 3 gate expo but with pressure.
Use your brain. Your choice of units affect your range. In poker, a check raise does not MEAN that he's got 2 pair. It does not MEAN he's got a flush draw. You assign a RANGE based on a huge amount of variables, then you make an optimal play based off this range.
5:20 nexus does not MEAN 1 gate expansion. It means build (A), (B), or (C). The important thing is that (D), (E) and (F) cannot be in his range. You can then construct an optimal play, rather than blundering into the midgame with a very vague idea of what your opponent is doing. Yes, you can narrow it down with these "finer details" you're talking about, I presume unit composition, chrono allocation and the like, but if you sacrifice clock timings you have no idea w hat you're missing out on.
Getting locked into these timings is only going to confuse you as a player. Unfortunately there is no quick answer to scouting. You have to be able to read into their build using finer details than the game clock.
If you played zerg you would die to DTs some of the time, have 2 queens against 4 phoenix some of the time, think hes "probably" 1 gate expo but he has a twilight, think his build range composes of multiple elements but indeed it can only be 1 gateway, think terran is "most likely" 12 rax 16 CC but in fact hes 12 14 proxy, think terran is 1214 but he's 1111, think terran is gas first but you're actually not sure and as such you don't alter your play, etc. You simply have no idea what you're talking about.
"confuse you as a player". Lol. So knowing someone is proxy rax with 100% confidence is confusing? Knowing that somebody is gas first is confusing? Knowing that his nexus is uber late is confusing? What is it about this information is confusing to you?
|
Im not zerg, but this is the level of analysis that will really improve players decision making, excellent post.
|
I have been looking for a post to know timing to be warrey of, thanks this is amazing.
|
I hope people appreciate that understanding how you need to handle the game at the Master's and up level is done exactly as this is presented. Having taken T to mid-high Master's, then P, and now working on my Z, if you aren't doing this, you are making yourself suffer needlessly.
You don't NEED to remember every single timing, just the ones that present dramatic problems for you. Knowing the critical features, aka, an 11Rax12Rax has 5 marines at your doorstep 1:30 after the first marine finishes, inform your Build Order first and your decision making second.
ZvT
-Plan your build so your pool finishes early enough to stop 2 Rax pressure
-If no 2 rax, have defense in place for first Hellion(s) at 5:05 and prepare for a Banshee by 6:30 and cloak by 7:30
-What are the 2-3 key big midgame timings? Important Marine-Tank push, 2 medivac, and FE timings?
Knowing that BEST CASE 1 or 2 Hellions hit you right at 5 minutes, and 2 or 4 hit by 5:30, means that if you have your Spines or Roaches up then, you will always be safe, from Bronze to GM. How can people say this isn't helpful no matter the league? You get prepared, you drone more and continue your plan safely.
BO planning
-What is your target midgame? What number of bases, gasses, and tech structures do you need to achieve that end?
-When do you NEED the first/key upgrade done? In ZvP, Lingspeed around 6 min to stomp a 4 gate is critical, and earlier stops Stalker pressure. but in ZvT? If you aren't going to use them to stop a Terran player AT THAT TIME, why get it so early? I take my gas so I can finish Ling Speed at the right time given a particular Terran tech path. If he goes 2 rax, it's a few food earlier. If it's a Hellion opening, I take it later and use Spines.
ZvP
-How do you get your Hatch down by 25 food? If he cannons? 2 gates?
-How are you stopping 4 gate at 6min?
-No Stargate, <5 Gateways, and no tech? Prepare for DT by 8 min.
-5 gate FE cancel starts 100 seconds after the Nexus starts (5:30-> ~7 min it leaves his base)
-6 gate push will HIT as Forge finishes which, with constant CB it takes 1:50min, so guess 2 min from when you see it start spinning if a lot of CB, 2:40 is longest possible time.
You can answer all your planning questions based on the best case scenario timings in a MU and then know what you need when in doubt. If you see no FE by 6 min, you should be hyper vigilant the next 30 seconds and prepare for a big Gateway and maybe Stargate push.
You don't need to know all the times, just the ones that your build can't deal with naturally.
|
Very well-written, I can`t wait for FFE deconstruction. I have been playing with nestea`s 12 pool 19 hatch opening vs FFE (60 drones by 9 mins and lair at this time) with very high success rate at mid-masters lvl (no drone scout, but very early lings to send to all the bases at once). However, because a toss has a cannon at his front, I cannot do ling pokes and I have no idea if he`s going stargate, DT, 7gate blink, or just standard colli ball. What I`m doing is getting evo chamber at 6:40 and getting 3 blind spores (oh well, better than losing the game) and sacing in an ovi at 7:30. If it`s cross on some maps, I might have trouble getting an ovi there in time though.
|
|
|
|
|
|
EPT
