EPT[D] SC2 Notes: How To Plan A Strategy - Page 4
| Forum Index > StarCraft 2 Strategy |
|
Tonymoi
France2 Posts
| ||
|
Salivanth
Australia1071 Posts
In TvZ, I want to kill my opponent with marines and SCV's before he gets enough defenses out to stop me. Step 1: What is "enough defenses"? Based on what I've seen of VODs, if my opponent has a spine, 2 queens, and a handful of lings out, I have to have either hurt him very badly or have a bunker up that I am able to repair before this happens. So I need to achieve one of these goals. It seems easier to get a bunker at my opponent's natural before X:XX rather than cripple him before X:XX. Step 2: What is the timing I have to hit? I remember from my Protoss experience that a 15 hatch goes down at about 2:15. A hatchery, according to the SC2 wiki, takes 100 seconds to build. This means it finishes at 3:55. A queen takes 50 seconds, and a spine takes 50 seconds. Given that you need time for creep to spread and for the main-base queen to move down to the natural, I estimate 4:55 as a timing where I have a repairable bunker up in range of my opponent's natural hatchery, or my attack has failed. This will be refined through testing. Step 3: How do I get a bunker up before this timing? 4:55 is ridiculously early. It's clear I have to do this off 1 base. I can't hide the bunker and have it in range of my opponent's natural. I probably can't hide it in a way that I can leapfrog over there, either. That means my opponent can and will pull drones to stop me, and zerglings if he has them. Therefore, I need marines and/or SCV's to protect them. Based on this logic, and the VODs I've seen, it's clear I have to proxy-rax. That's what the pros do, and now it's obvious why 2-rax pressure into expand is no longer viable at high levels. 11/11, 11/12, and 12/12 are all options for the proxy. I'm not sure what the relevant differences are between them, so I'll have to test this. A bunker takes 40 seconds to build, so I'll need to have my defensive forces prepared to stop the bunker being denied at 4:15, 4:05 to be safe. This means I probably need to pull SCV's from my base at about 3:15, estimating a 50-second rush distance. I know to pull 5 SCV's if I'm all-in, because I read it on a forum thread. A barracks takes 65 seconds to build, and I'd like to have at least 1 marine (25 seconds) from the second barracks at the critical 4:00-4:05 timing. That's likely to be a total of 3 marines. This means my first rax has to start at 2:05-ish, the second rax at 2:30-ish. This means my first SCV should move out to proxy at about 1:25, and the second at about 1:50. This matches up with the timing I saw on a VOD, where the first SCV leaves the base as soon as the Depot it's building finishes/ Step 4: How do I optimise this timing? Based on VODs and logic from previous steps, I should start with the following build order: 10 Depot, send an SCV out as soon as it finishes. At about 1:45-1:50, send another one. Two raxxes, proxied, at 2:05ish and 2:30-ish.. Make constant marines. At 3:15, pull 5 SCV's. At 4:05, use my 7 SCV's and ideally, 3 marines to buy time for a bunker or two to go up, just off creep. I need one to finish, while retaining at least a couple of SCV's. This puts me in a good position to win the game if the bunker goes up. And thus, I have a build order! I may have been too focused on the timings and less on the overall strategy, though. | ||
EsportsJohn
United States4883 Posts
On November 01 2013 19:04 ApocAlypsE007 wrote: I have a question. What is your thought process when dealing with mirror matches where your opponent plays exactly the same as you and tries to achieve the same goals as you? How do you gain an advantage in that situation? This an older daily, but contains some relevant information regarding mirror matchups and learning to identify your "edge" and use it to your advantage: http://day9.tv/d/Day9/day9-daily-325-pvp-where-is-your-edge/. I want to stress that you shouldn't take the specific details of strategy in this daily to heart (the first example game is a WoL 4-gate vs. 4-gate on Tal'Darim Altar), but there are definitely some good general thoughts about finding your edge and exploiting it in a useful way. Something Teo mentioned earlier in this thread was the "greed > safety > aggression > greed" wheel: Oh i forgot, saying "cheese and allins can be countered, but macro builds can't" is also somewhat incorrect. No matter what, you always have the wheel of aggression>greedy>safety>aggression. Assuming you are both playing in a macro game, sometimes it's possible to scout the particular "brand" of macro build the opponent is using and immediately switching into the style that will counter it. The best example i can come up with right now was a WoL GSTL match where the P went fast 3nex, the T instantly scouted it thanks to a lucky scv, went 5rax marine and won. If he had gone for the standard 2medivac push he wouldn't have been able to punish his opponent as heavily. At the end of the day, every sc2 build has some kind of "counter", or, another build that does extremely well against it. On November 01 2013 22:16 Millicant wrote: Now, that seems to bring up a great question. With a strict adherence to a "plan" you essentially have two players with two plans, and whichever happens to work better wins. Where does Adaptability/Flexibility come into it? When does a player deviate from his plan to exploit his enemy? Within a perfectly refined strategies, there should be a lot of flexibility and pre-planned adaption available. For instance, in ZvP, if you scout your opponent going for a 7:30 3rd nexus, you know you can cut unit production, get a faster 4th, get extra tech up, and hit with a stronger attack slightly later. This is all because you know your opponent is going for a 3rd base and can't attack you. If he didn't try to go for a fast nexus, you would do your normal game plan: build a swell of units (or at least have gas saved up) around 9:00 to stop any 2-base timings, and then take a 4th and extra tech up. Protoss can opt to build a 3rd nexus before additional gateways based on scouting. Terran too (barracks and CCs). For zerg, it's a tradeoff between additional units and tech/hatcheries. In this way, it's kind of just a rearrangement of minerals. It doesn't actually change your whole game plan. Naturally, the amount of variation ("wiggle room") in your strategy is dependent on the game length. In shorter games like cheeses or 2-base all-ins, there's a lot less variation available to you to get ahead and still achieve your goal while in a lategame-focused strategy there are nearly infinite variables. This correlation is the reason why lategame-focused strategies have a higher likelihood of winning. Cheeses and all-ins can still be beaten with good scouting, defense, and decision making. In other words, a godlike macro player is much more likely to win tournaments compared to a "master cheeser". | ||
|
EsportsJohn
United States4883 Posts
On November 02 2013 01:20 Salivanth wrote: Just today, I was thinking of trying to learn Terran through cheesing, but didn't have any build orders. I guess this is a perfect opportunity to practice goal-oriented play. Let's give it a try: 10 Depot, send an SCV out as soon as it finishes. At about 1:45-1:50, send another one. Two raxxes, proxied, at 2:05ish and 2:30-ish.. Make constant marines. At 3:15, pull 5 SCV's. At 4:05, use my 7 SCV's and ideally, 3 marines to buy time for a bunker or two to go up, just off creep. I need one to finish, while retaining at least a couple of SCV's. This puts me in a good position to win the game if the bunker goes up. And thus, I have a build order! I may have been too focused on the timings and less on the overall strategy, though. I think this is a really good strategic breakdown of a very simple strategy. The next big step is trying this out a few dozen times to find out which maps work best for it and where the best proxy locations are. You would probably delve into overlord scouting patterns on each map and how to exploit them. On November 01 2013 10:34 CakeSauc3 wrote: I love to play mech with Terran, as bio is too difficult for me to control vs. Zerg. I want to analyze Flash's game vs DRG on Derelict Watcher from IEM New York: http://us.esl.tv/video/3dca06cc715c9cdc/ Goal: Move out at +2/+2 with tank, thor, hellion, and banshee when Zerg is transitioning to Hive tech, in order to kill them before Vipers are produced, which would render our mech army useless. Working backwards: Use constant hellion/banshee production to establish map control, pressure Zerg, and roast drones while setting up a third base, beginning double armory upgrades, and working up to 5 factory mech production. The pressure prevents Zerg from teching up quickly, buying us time to produce an unstoppable mech deathball. How to open? Use a reaper expand to scout the Zerg's opener while we take our natural and prepare for early game aggression. Am I getting the concept? Is there something I'm missing? Thanks  I think it's a good goal and you've got some good ideas thrown in there. To be honest, my format in the OP can be better revised to read "GOAL -> lategame -> mid game -> early game" instead of all the "steps". Using this type of method, we want to look at:
These are all the kinds of considerations we need to make when just stating a goal as simple as "I want to kill the zerg player with a big mech army before vipers are out." But yeah, essentially you've got it. Just dive right into all the details and focus on trying to make everything as tight and powerful as possible. Again, it's important that every single move should connect from beginning to end; every single thing you do in the game should be working toward the ultimate goal of making the strongest attack possible before vipers are out. | ||
|
CakeSauc3
United States1437 Posts
On November 03 2013 00:19 SC2John wrote: MAP CHOICE What map works best for this? Map choice is very important to strategy. If we're looking at how we've structured this build, we don't need a 4th base so we could potentially do it on any map which can get 3 bases up easily (Akilon, Derelict Watcher, Bel'Shir Vestige, Yeonsu, Whirlwind). Another thing to note is the general terrain of the map. Some maps are smaller and more choky, making them easier to push across like Bel'Shir and Akilon. Others like Derelict or Whirlwind maybe have wider, open areas or long rush distances that will make pushing across the map dangerous or time-consuming. These are all the kinds of considerations we need to make when just stating a goal as simple as "I want to kill the zerg player with a big mech army before vipers are out." But yeah, essentially you've got it. Just dive right into all the details and focus on trying to make everything as tight and powerful as possible. Again, it's important that every single move should connect from beginning to end; every single thing you do in the game should be working toward the ultimate goal of making the strongest attack possible before vipers are out. Huge thanks to you! That's a lot to digest. I have just a few questions: 1.) This is a lot to get down. In order to learn it quickly, do you suggest I practice this build on my own while I refine it before bringing it to the ladder, or do you think it's best just to dive into the ladder with these things in mind and then make adjustments as I observe my opponents' play? 2.) Concerning map choice, I'd always heard that mech was best on maps with tight chokes and nearby locations. However, one reason I got so excited with this build was I saw Flash do this on Derelict Watcher, a map which seems to not favor mech much in the slightest. Do you think that makes it a safe build for any map, or was Flash just pulling a mind game that would only work at the very highest level? 3.) Lastly, what you mentioned at the end concerning every move, every action should go towards making this build as tight and strong as possible... are you mostly talking about just making sure my build order gets refined to the point that my attack hits the soonest possible? Or are there other actions/moves that I need to consider? If you can elaborate on that, that would be awesome. I appreciate your willingness to help! | ||
|
B-rye88
Canada168 Posts
| ||
|
EsportsJohn
United States4883 Posts
On November 03 2013 02:28 CakeSauc3 wrote: Huge thanks to you! That's a lot to digest. I have just a few questions: 1.) This is a lot to get down. In order to learn it quickly, do you suggest I practice this build on my own while I refine it before bringing it to the ladder, or do you think it's best just to dive into the ladder with these things in mind and then make adjustments as I observe my opponents' play? 2.) Concerning map choice, I'd always heard that mech was best on maps with tight chokes and nearby locations. However, one reason I got so excited with this build was I saw Flash do this on Derelict Watcher, a map which seems to not favor mech much in the slightest. Do you think that makes it a safe build for any map, or was Flash just pulling a mind game that would only work at the very highest level? 3.) Lastly, what you mentioned at the end concerning every move, every action should go towards making this build as tight and strong as possible... are you mostly talking about just making sure my build order gets refined to the point that my attack hits the soonest possible? Or are there other actions/moves that I need to consider? If you can elaborate on that, that would be awesome. I appreciate your willingness to help! 1) Absolutely. If you're already quite familiar with a build or the way a particular building grouping works (for instance, 3rax/factory/sp on 2 bases) and your macro is decently strong, you can attempt to do something on the fly. But usually this doesn't happen unless you're in masters. Therefore, I strongly recommend loading up a dozen games against the AI until you have your opening build down perfectly with no hiccups. 2) "Mech" as a whole is based off of the tank, which naturally does well with tight chokes and short walking distances and does horribly when surrounded 360 degrees on open ground. That said, it might be worth looking very closely at the VoD to see what Flash does differently to allow this push to happen. One might say that the hellion/banshee control in the mid game slowed down creep spread significantly and that the major emphasis Flash puts on banshees allows him to have a really high DPS despite the terrain. Also, as pointed out in your goal, he does a good job of hitting a pre-hive timing, which allows him to hit before vipers and blinding cloud. I'm not fully convinced mech can work on this map all the time, but perhaps there are some elements to Flash's play that allowed him to do this. The best way is to just copy Flash's game plan the best way possible and find out for yourself! 3) I mean that all of your timing should be as sharp and crisp as possible. But I also am trying to point out that everything should have a use toward making your attack stronger. For instance, the hellion/banshee harass not only kills off stray drones/queens/units, but also gives you strong map control and slows down the creep spread. It's important to keep your banshees alive as well; each one gives your final hellbat/tank/thor/banshee attack much more force. There are smaller, more subtle things too: does it make sense to do an early marine poke? Possibly, but it slows down our hellion/banshee phase by a lot, which means creep spread will be harder to stop. Does it make sense to try and do a mine drop before hellion/banshee? Not really, it achieves the same kind of goal hellion/banshee harass does. Does it make sense to get factories 4 and 5 before our armories? No, because our +2/+2 timing will be much much later and we probably can't afford to produce off of 5 factories yet anyway. It's important to ask yourself these questions and relate it back to the idea: "does this ultimately help me with my goal?" I suggest listening to this older Day9 podcast. In it, he talks about "redundancy and purpose" and basically explains that you should iron out your strategy to include only the things necessary and give those things as much purpose [towards your main goal] as possible. On November 03 2013 04:51 B-rye88 wrote: So given this post, would you recommend a learning pattern whereby we take the simplest builds (ie. as a bronze player I'm currently practicing and using a 1-base 3-rax early stim/concussive timing) and gradually learning more advanced and possibly divergent builds going forwards ala 'the Steps' learning process? Definitely! To be honest, the best types of builds are strong 2-base timings. They give you a little taste of "macro" then allow you to win the game early with an attack that relies more on "micro". Just make sure that you're hitting all of your benchmarks and starting your attack on time. I learned almost all of my basic mechanics from practicing things like 2-base muta or 2-base stim timings. | ||
|
mau5mat
Northern Ireland461 Posts
| ||
|
EsportsJohn
United States4883 Posts
On November 03 2013 10:26 mau5mat wrote: Sorry if I sound rude, but I would really appreciate a response from my post on page 1. Great posts and information as usual from this thread! I've been meaning to get to it, but my terran knowledge is a little lacking so I was going to pass it off to vader. Vader's been out busy for the past week, but I'll put him on that post as soon as he gets back. In short, though, I really liked your division of the game stages, and I thought it was a good format to put general TvZ into. I'll get vader to guide you better! | ||
|
mau5mat
Northern Ireland461 Posts
On November 03 2013 10:41 SC2John wrote: I've been meaning to get to it, but my terran knowledge is a little lacking so I was going to pass it off to vader. Vader's been out busy for the past week, but I'll put him on that post as soon as he gets back. In short, though, I really liked your division of the game stages, and I thought it was a good format to put general TvZ into. I'll get vader to guide you better! Much appreciated! | ||
|
PinheadXXXXXX
United States897 Posts
PvZ Goal: I want to kill zerg with a colossus-gateway unit based army before hive tech upgrades or units kick in. Step 1: What is the composition I want to deny? What is the comp i need for myself? I would ideally like to kill zerg before they get ultras, brood lords, or 3/3 on their units. I want to aim for a composition with colossi, sentries, blink stalkers, and perhaps archons or high templar for feedbacking vipers. Step 2: What timing do I want to hit? From watching pro games, i know that the earliest hive finishes at around 13 minutes, but a hive timed for actual hive tech finishes more around 15. (my observations show that the 13 min hive is only for vipers and doesn't allow zerg enough money for much else) I want to hit a timing with the highest tech i need to get; which in this case is probably archons. I should hit with maybe +3 attack and charge on the zealots, and a high enough number of sentries to prevent roach/hydra from kiting me too hard. I should be able to get all that out by 15 minutes. Step 3: What do I need to do to get there? In order to get a large army by 15 minutes, it would probably be advantageous to get an early third; i can take that from an FFE to get earlier upgrades and tech. I want to get colossi first, probably 2 out by 12 minutes because I have seen strong roach or roach hydra pushes that hit around then. I will need somewhere around 6 sentries around that time, maybe more. And when I push I would like a strong infrastructure of 10-12 gateways with a robo. Step 4: What do I have to do early game to achieve my goals? I want to open with a forge fast expand to get my +1 started early, and then follow it up with some tech. After a lot of trial and error I have decided that robo first doesn't offer enough map control, so I should probably open up stargate with phoenixes to control the zerg's greed a little and make my midgame defense easier. After that, I can secure a third. To properly secure this third I need somewhere around 4 sentries and a couple zealots, and i need to get 2-3 gateways beforehand to get those units and from there I can make a third with a cannon walled off by more gateways. I should probably follow that up with a twilight and robotics for more upgrades and colossi, and then I can go up to 8 gateways on 3 bases, with thermal lance, blink, sentries, and cannons. I can add on 3 more gateways and a templar archives to finish up when I attack, and then make archons for the actual push. | ||
|
CakeSauc3
United States1437 Posts
On November 03 2013 09:17 SC2John wrote: 1) Absolutely. If you're already quite familiar with a build or the way a particular building grouping works (for instance, 3rax/factory/sp on 2 bases) and your macro is decently strong, you can attempt to do something on the fly. But usually this doesn't happen unless you're in masters. Therefore, I strongly recommend loading up a dozen games against the AI until you have your opening build down perfectly with no hiccups. 2) "Mech" as a whole is based off of the tank, which naturally does well with tight chokes and short walking distances and does horribly when surrounded 360 degrees on open ground. That said, it might be worth looking very closely at the VoD to see what Flash does differently to allow this push to happen. One might say that the hellion/banshee control in the mid game slowed down creep spread significantly and that the major emphasis Flash puts on banshees allows him to have a really high DPS despite the terrain. Also, as pointed out in your goal, he does a good job of hitting a pre-hive timing, which allows him to hit before vipers and blinding cloud. I'm not fully convinced mech can work on this map all the time, but perhaps there are some elements to Flash's play that allowed him to do this. The best way is to just copy Flash's game plan the best way possible and find out for yourself! 3) I mean that all of your timing should be as sharp and crisp as possible. But I also am trying to point out that everything should have a use toward making your attack stronger. For instance, the hellion/banshee harass not only kills off stray drones/queens/units, but also gives you strong map control and slows down the creep spread. It's important to keep your banshees alive as well; each one gives your final hellbat/tank/thor/banshee attack much more force. There are smaller, more subtle things too: does it make sense to do an early marine poke? Possibly, but it slows down our hellion/banshee phase by a lot, which means creep spread will be harder to stop. Does it make sense to try and do a mine drop before hellion/banshee? Not really, it achieves the same kind of goal hellion/banshee harass does. Does it make sense to get factories 4 and 5 before our armories? No, because our +2/+2 timing will be much much later and we probably can't afford to produce off of 5 factories yet anyway. It's important to ask yourself these questions and relate it back to the idea: "does this ultimately help me with my goal?" I suggest listening to this older Day9 podcast. In it, he talks about "redundancy and purpose" and basically explains that you should iron out your strategy to include only the things necessary and give those things as much purpose [towards your main goal] as possible. . Thanks for your detailed response! I'll be sure to check out that podcast, and I think I'm going to spend a lot of time on this build. I'm excited to see if I can use it to get a high-percentage win rate vs Zerg on the ladder, something which I've never really been able to do with Terran. Here's hoping! | ||
|
vaderseven
United States2556 Posts
1) Absolutely. If you're already quite familiar with a build or the way a particular building grouping works (for instance, 3rax/factory/sp on 2 bases) and your macro is decently strong, you can attempt to do something on the fly. But usually this doesn't happen unless you're in masters. John, I disagree. Rather strongly to be honest. I find the very best way to learn the strongest ways to play is to actually play in a very rigid way in game. When I am learning a new style I will come up with I feel is a logical game plan vs the most common thing I face and I will then plan out everything macro and army movement wise. I then will go play that exact plan vs as many people as possible on that map. I will group all the replays (regardless of win or lose) by enemy game plan. Once I have a good group of replays vs a type of enemy play style, I watch them all back to back looking for patterns. Seeing different people do similar styles will often make patterns REALLY jump out as while every player has hugely different takes on styles there is core concepts in every style that carry over from player to player. As such, those core concepts will stick out as very similar actions, timings, movements, etc regardless of the player. Only once I have seen these patterns emerge do I attempt any type of modification to my play. This keeps my modifications focused on the enemy's plan instead of of cute things that happen in single games. Often, when you have tweaked your play to handle a style in a logical way, small cute things become non factors when they before seemed like huge roadblocked that had to be addressed. This kind of approach to learning a style leads to a huge number of loses at first due to you blindly following a plan. Your army will be snipped because it was at a dangerous location. Your main will fall because you had no defense to his harass. The game will suddenly end because three DT walk into your main. Thats PERFECT. If I have 19 replays of my 4M vs two base muta players I can watch them all back to back, find the common part of why I lost to each of them, address that sole factor, and have the smallest and easiest to understand adjustment that works vs all two base muta plays. This leads to very smart reactions such as burrowing two mines in very specific spots in reply to seeing a late gas opening, no fast third base, and only two queens followed by adding two very specific turrets at two very specific locations once the mutas are seen. The mines will be placed at the only two locations that I have ever seen 2 base muta players send their first mutas in at and the two turrets will be placed at two locations that I have simply never been able to get a marines from my rally to fast enough to defend though they were always just a second too late. This adjustment might allow me to be safe vs the mutas as the two mines buy me just enough safety to get the two turrets up and by the time a muta ball big enough to deal with that is in play my game plan has moved to a point where he has to react to it or die. On the fly adjustments or even adjustments based on only one reply would have probably led me to scanning at a certain time, spamming 2 turrets at each base, and delaying some timing. The turrets location would be based on fear of where I perceive weakness instead of what the other player has tested vs many Terran players is actually weak. I would end up investing more resources in game to a on the fly reaction that will not possibly contain logic created through pattern observation and will instead contain fear motivated by guesses. Basically, take huge value in replays where you get completely crushed because you didn't know the enemy would do something. Gather as many as you can and organize them and then apply logic to patterns you see. By keeping your play a constant, you will be able to actually craft exact deviations that require the least investment for the largest payoff. The more times you have done the exact same thing, the more exactly you know its actual weakness. By knowing your plays actual weakness you can address the real problem. If you freestyle, you practice at guessing and perfecting the art of fear. I prefer to perfect the art of problem solving. Treat the game like a science experiment, don't let your own play be an un-measurable variable when it could be the ultimate control. A strange and wonderful result of this approach is that what sounds like memorization actually becomes intuition. You just have to separate the active questioning part of the mind from the active playing. Once that is done, the things you decide as logical will be revealed slowly as truths in your play. As that happens, you learn that logic and how to play it at a much deeper level than some kind of written out build order or list of reactions. For more on this, see http://www.teamliquid.net/staff/zatic/Day[9]0002-BuildingTriggers.mp3 (discusses how getting large sample size of replays where you play the same is trully crucial in developing good decisions) and read this great wonderful quote: + Show Spoiler + Have confidence in your execution, and skepticism and doubt in your analysis. When you're playing the game, you should be convinced that what you're doing is absolutely correct, so that way you can maximize your efficiency with whatever you're doing. Whether it'd be a really strong play or an absolutely flawed play. After the game is when you should look at your play and say "mhm, what did I do right and wrong here, what's a way that I could improve? What's a build that he could have done that would have crushed this a 100% of the time?" That is when you think of what adjustments you want to make for the next game. And of course, in that next game, you'll play with 100% confidence as though everything you're doing is absolutely correct. A huge issue that a lot of players deal with is that when they are in the game, and they get thrown off just a little bit - maybe they get storm-dropped or maybe they see a strategy they havent seen before -, that's when they start to analyze their play and to make these huge, large-scale adjustments. Unless making that adjustment was part of your plan, don't do it! You need to make sure that what you're doing is something you are comfortable and confident with. That is what's going to maximize your chance of winning, not some last minute, second-guessing strategy you pulled out of nowhere. -day9, podcast on Having a Good Mindset | ||
|
vaderseven
United States2556 Posts
On November 03 2013 04:51 B-rye88 wrote: So given this post, would you recommend a learning pattern whereby we take the simplest builds (ie. as a bronze player I'm currently practicing and using a 1-base 3-rax early stim/concussive timing) and gradually learning more advanced and possibly divergent builds going forwards ala 'the Steps' learning process? I suggest learning a game plan that you feel you understand the logic behind the winning move OR to blindly play out a macro build you stole from someone very good at the game. If you choose the latter, pay attention to what you lose and carefully group the replays. Once you have played and lost enough, you can begin to make statements of logic about how you feel the game can be won. Once you have blindly macro'd to a lose vs a maxed protoss army as terran 500 times you start to see the logic in killing the toss before they get to that stage and you even start to see the logic of killing them at specific times. Personally, I choose the first choice. Find some kind of logic, to hell with WHEN that logic leads to a win, and build a plan based on the logic. Then gather replays and analyze and modify logically while executing confidently. The blind macro method can be useful for very low level players to jump start their mind with some metagame as you will see a wide variety of enemy styles which will help to motivate logical plans on how to kill said enemies. I would say, anyone under the rank of Plat can use that method as a valid approach if they wanted too. That's a generalization though. I should say, anyone who is at or below the rank of Plat due to lack of understanding in how are won or how matchups work could benefit from that approach. Anyone can benefit from the choose a logical way to win and try to make it work approach. It is most likely the overall superior way. | ||
|
EsportsJohn
United States4883 Posts
On November 03 2013 18:19 vaderseven wrote: John, I disagree. Rather strongly to be honest. I find the very best way to learn the strongest ways to play is to actually play in a very rigid way in game. When I am learning a new style I will come up with I feel is a logical game plan vs the most common thing I face and I will then plan out everything macro and army movement wise. I then will go play that exact plan vs as many people as possible on that map. I will group all the replays (regardless of win or lose) by enemy game plan. Once I have a good group of replays vs a type of enemy play style, I watch them all back to back looking for patterns. Seeing different people do similar styles will often make patterns REALLY jump out as while every player has hugely different takes on styles there is core concepts in every style that carry over from player to player. As such, those core concepts will stick out as very similar actions, timings, movements, etc regardless of the player. Only once I have seen these patterns emerge do I attempt any type of modification to my play. This keeps my modifications focused on the enemy's plan instead of of cute things that happen in single games. Often, when you have tweaked your play to handle a style in a logical way, small cute things become non factors when they before seemed like huge roadblocked that had to be addressed. I may have been a little unclear when I answered that question. I meant to say "absolutely practice the strategy a dozen times against the AI first". The other part there is when you're making a modification to an already established strategy. For instance, on a high level when you've already had tons of experience with reaper/hellion pressure into 4M pressure, you could modify it to hellion/banshee within a couple of games and experiment with that. It's a minor change for someone who already has the mechanics and the general game plan down. However, someone trying to learn void ray-based PvZ after they've played exclusively phoenix/colossus is BAD from a learning experience. Even someone who is has good macro and has played phoenix/colossus extensively will need to start from nearly the ground up in order to make it a working strategy and understand it fully. Also, vader, if you could get to this post, it would be much appreciated: http://www.teamliquid.net/forum/viewmessage.php?topic_id=433514¤tpage=3#53 | ||
|
vaderseven
United States2556 Posts
| ||
|
Millicant
United States78 Posts
On November 03 2013 18:34 vaderseven wrote: I suggest learning a game plan that you feel you understand the logic behind the winning move OR to blindly play out a macro build you stole from someone very good at the game. If you choose the latter, pay attention to what you lose and carefully group the replays. Once you have played and lost enough, you can begin to make statements of logic about how you feel the game can be won. Once you have blindly macro'd to a lose vs a maxed protoss army as terran 500 times you start to see the logic in killing the toss before they get to that stage and you even start to see the logic of killing them at specific times. Personally, I choose the first choice. Find some kind of logic, to hell with WHEN that logic leads to a win, and build a plan based on the logic. Then gather replays and analyze and modify logically while executing confidently. The blind macro method can be useful for very low level players to jump start their mind with some metagame as you will see a wide variety of enemy styles which will help to motivate logical plans on how to kill said enemies. I would say, anyone under the rank of Plat can use that method as a valid approach if they wanted too. That's a generalization though. I should say, anyone who is at or below the rank of Plat due to lack of understanding in how are won or how matchups work could benefit from that approach. Anyone can benefit from the choose a logical way to win and try to make it work approach. It is most likely the overall superior way. Now, this is fascinating. Not to stir up drama, but disagreement between you two really makes for awesome discussion and I (and others like me) benefit greatly from it! So this re-ignites my earlier question about reacting/responding to your opponent's play. I understand that for the process of developing a strategy or game plan, you are clearly saying "Don't modify it! Look at the replays and make adjustments after many many games!" That makes perfect sense - got it. However, does this mean that you NEVER react to what your opponent is doing? Even once your plan is developed and solidified, you blindly follow your plan and whichever player's plan comes out on top wins? This seems a little... disappointing. Lastly, you state "I would say, anyone under the rank of Plat can use that method as a valid approach if they wanted too." That method meaning the blind macro? But to the contrary you are saying that finding the logic behind the plan and executing is superior, yes? | ||
|
B-rye88
Canada168 Posts
Now, this is fascinating. Not to stir up drama, but disagreement between you two really makes for awesome discussion and I (and others like me) benefit greatly from it! So this re-ignites my earlier question about reacting/responding to your opponent's play. I understand that for the process of developing a strategy or game plan, you are clearly saying "Don't modify it! Look at the replays and make adjustments after many many games!" That makes perfect sense - got it. However, does this mean that you NEVER react to what your opponent is doing? Even once your plan is developed and solidified, you blindly follow your plan and whichever player's plan comes out on top wins? This seems a little... disappointing. Lastly, you state "I would say, anyone under the rank of Plat can use that method as a valid approach if they wanted too." That method meaning the blind macro? But to the contrary you are saying that finding the logic behind the plan and executing is superior, yes? My thoughts: Your plan can include divergent options. For me, this would be something such as "I will scout my protoss opponent with a reaper before going up the ramp with my 3-rax stim timing; if he is still on one-base, I will set up a light contain to keep him on one-base, and use the reaper / scan to find out whether he's teching or 4-gating, cut a round of production to expand, and either bunker the high & low ground in case of twilight, get missile turrets if we see dark shrine, get 3 bunkers on the low ground if I see immortal play, or tech to starport if I see a robo bay; If i see early tech and little production, I can also go up the ramp and kill the tech structure with my stim". The game will play out the same way early, with the same build, but I have decided to do various things based on what I find. So long as you have a plan, derived from logic, you can then learn to execute it. | ||
|
EsportsJohn
United States4883 Posts
On November 04 2013 04:48 B-rye88 wrote: My thoughts: Your plan can include divergent options. For me, this would be something such as "I will scout my protoss opponent with a reaper before going up the ramp with my 3-rax stim timing; if he is still on one-base, I will set up a light contain to keep him on one-base, and use the reaper / scan to find out whether he's teching or 4-gating, cut a round of production to expand, and either bunker the high & low ground in case of twilight, get missile turrets if we see dark shrine, get 3 bunkers on the low ground if I see immortal play, or tech to starport if I see a robo bay; If i see early tech and little production, I can also go up the ramp and kill the tech structure with my stim". The game will play out the same way early, with the same build, but I have decided to do various things based on what I find. So long as you have a plan, derived from logic, you can then learn to execute it. I wrote about this a page ago...I talked about the difference of playing to win vs. playing to practice.It's really important to understand that vader and I are talking about playing to practice. Playing to win means using your refined strategy "toolbox" in order to pick the best strategy and variation in order to best beat your particular opponent while playing to practice is means specifically playing one strategy over and over, changing little to nothing from game to game, in order to refine that strategy and learn the best possible way to play it. When vader says something like "Do not change your strategy at all during the game," he is not implying that you should ignore what your opponent is doing. What he means is to take a large sample size of the same exact build to see what damage it can do and how it can come crumbling down. After you have this sample size, make minor adjustments or add some considerations to your strategy to fix the weak areas or strengthen the strong areas. | ||
|
Millicant
United States78 Posts
On November 04 2013 09:52 SC2John wrote: I wrote about this a page ago...I talked about the difference of playing to win vs. playing to practice.It's really important to understand that vader and I are talking about playing to practice. Playing to win means using your refined strategy "toolbox" in order to pick the best strategy and variation in order to best beat your particular opponent while playing to practice is means specifically playing one strategy over and over, changing little to nothing from game to game, in order to refine that strategy and learn the best possible way to play it. When vader says something like "Do not change your strategy at all during the game," he is not implying that you should ignore what your opponent is doing. What he means is to take a large sample size of the same exact build to see what damage it can do and how it can come crumbling down. After you have this sample size, make minor adjustments or add some considerations to your strategy to fix the weak areas or strengthen the strong areas. Ah - that definitely clears it up somewhat. Thanks! | ||
| ||
