Spawning Tool: FrogBuildWeekly#10 Making your own Game Plan

Help us find the best build orders by creating an account and voting!

Description

Foreforewords

I made this a year ago but was unable to complete it and upload it, but I just finished it and nothing really changed in this year. The only exception is that upgrades take longer so consider chrono boosting them. I also did not review it since its so long so if there is something significant missing feel free to let me know.

Forewords

Writing these has been fun, but I don’t have the time to continue, at least not on a weekly basis. That said I think we went through many of the game fundamentals. The only ones I really wanted to do but haven’t done is focus on baserace and camera hotkeys, and perhaps mechanical drills. Depending on demand I might do a few more at some point in the future. Thanks for reading, and thanks to those that gave me motivation to continue, and thanks to the people that helped me develop the builds by letting me do the builds repetitively! Hopefully I have inspired a few people to also make strategies/builds with the purpose of training a specific rather than winning.

Introduction

This week the topic will be on how to make your own good game plan. You can make your own plans by refining a strategy that you worked on for a year, or it can be directly improvised in-game. The preparation time per strategy with the method I will discuss is about 2-4 hours when you are experienced with it and maybe 10 hours for the first time. As a person that has been doing my own random stuff for years I feel confident that I can make anything into a viable strategy in this time, and my goal this week is to try to throw that knowledge out to you. I have never really thought about how I do it before or talked about how other people do it though so this guide will be a bit of an experiment and much longer than the previous ones. I will try to make the explanation into recipe-form, but the truth is a lot of the tasks take experience and trial and error. Being creative is never easy. Don’t expect to make something amazing on the first attempt, instead learn from your mistakes after making a bad strategy by asking “Which step went wrong and why?” Also note that a good game plan is far from any of the strategies in the previous guides, the previous builds are simply a collection of builds that usually lead you to be able to train a single aspect of the game, knowing that they are not safe or good plans. An actual solid and good plan often relies on much more fundamentals and tactical knowledge, and much more precise scouting, which I hope will become clear after this week. If you don’t have the fundamentals ready to perform this week’s example strategy you could simply just read instead and then making something that you can perform yourself. If however you want to actually play my example strategy this week you will need the following from previous weeks guides:

split off / trash control group + counter attack control group from week 2
decent force field usage and nonstop probing from week 3
multitasking ability / controlling blink dts week 5 (this week there is no phoenix/oracle though so it is less intense than that one, but the macro is more difficult instead)
rapidfire week 6 (used this week with prisms)
understand my view on win conditions and how to make a timing week 8, we will be balancing different win conditions this week.
be comfortable with warp prisms and have a control group setup that can control a main army consisting of 2 different groups of units week 9. (This week you can put secondary prisms in oracle group since there is no oracle). This brings it to a total of minimum 5 army control groups (including trash group) for this week’s game plan.

link to all weeklies: https://lotv.spawningtool.com/build/?contributor=18649&sort_by=r

That said let’s get started! The first step to making your own game plan is the initial idea. This can be so many different things, some examples are:

a build order or a part of a build order
a single tactic (for example force field ramp like week 1)
achieving a goal (for example deny creep)
a certain type of timing (eg disruptor stalker timing before brood lord)
finding a solution to a problem you had and adapting a different game plan to that (for example doing a certain strong carrier play against fake mutalisk since you already had stargates and a fleet and adding this plan as a tree-branch to a different game plan that struggled against fake mutalisks)
trying to copy someone else's game plan based on watching their games. Note that this is never exactly possible since a single game is always only a small part of the plan and contains flaws, so what you are always doing is making your own strategy which is highly inspired by theirs. Because you’re also good at different things it can be good to make certain changes anyway. Also note that if you simply do 95% of things exactly like they do then it might work against an opponent that does not scout, but it is likely that you’re creating a huge weakness in that case. A great example of this is the Deepmind AI that is an expert at copying human play but does not understand a lot of the intentions behind certain moves. In other words even when you copy others you need to get your own understanding of what you’re doing. If someone explains a strategy to you then that will give you much more than watching someone’s games, but there are always gaps that you need to fill in by yourself.

For this week’s example my idea was based on the win condition to counter attack with chargelots while defending at home. I want to emphasize that the main idea can be anything, if you do the following steps properly I truly believe that you can make almost anything into a viable strategy. So you don’t have to make something as dull as i did, you could base your strategy around nukes, observer speed, mass stasis ward or whatever you dream of. Playing what I want to play has always been my motivation to make my own game plans.

The second step is to try to generalize the requirements for the idea. In other words it is easy to find ONE way of doing it, for example to open with a prism instantly and making 8 gateways and charge and dt and hiding the prism across the map and scouting with an observer to see when the opponent moves out. When they do you defend with some dts and counter warp in lots of chargelots. But this is not what we want, we don’t want to use the first idea that comes into our head, we are not that intelligent :p. One massive flaw in the plan above is that if the opponent scouts the gateways then they will never attack and your army scales super poorly. So we want to generalize what we actually have to do. Our actual requirements are to at some point in time have charge + prism across + ability to defend while warping in at their side of the map + a plan if the opponent does not attack (final requirement is optional, depends if we want a solid plan or a gimmick that relies on opponent attacking, I want a more solid plan this time so I keep it). This is much more general and something we can keep in mind and then work out the details for later. The less requirements the better and the more general the better.

The third step is a creative one, here we use creativity to find tactics or macro tactics that are efficient in achieving our goal while satisfying the requirements from step 2. And we want these to synergise well with each other. Quite often the synergies you come up with are strong enough to hatch a new better idea giving up the original one. First we just list the ideas, any idea is great, we want these to be fresh in our head for later steps. Throughout we also list potential problems and see if they can be solved, maybe we need to add some requirements. Just as a sidenote - this step shows how the more tactics you have mastered the more options you will have and the better the strategy will be in the end, so even when we prepare a strategy the experience of knowing thousands of tactics helps even though only a handful will be used in the end (Wanted to say that because I think improving tactics is a great way to improve long-term and this is just one reason why.)

The counter attack is stronger the more gateways we have.
Because a big part of our army will be on the other side of the map we want the rest of our army to be strong defensively. (Examples of that is sentries, disruptors, colossi, shield batteries, staying on lower base count or sacrificing a base if they attack, against no lair dt, …)
Mutalisks start sounding scary with the prism completely being shut down. (potential problem)
As a plan if the opponent does not attack a higher gateway count would synergise with hitting a timing. (since we also want the gates to warp in lots of counter zealots), this might also solve the mutalisk problem. At the same time this can have bad synergy with the defensive unit composition.
+1 zealots is kinda nice to 2 shot lings, armor is also better than usual.
If we hit a timing that can synergise with warping in zealots in their main at the same time.
Addressing the plan if opponent does not attack, if the timing is an all in we don’t have to care about the late game.
If the opponent does not attack we can keep adding prisms to make the counter attack threat bigger.
Building a base that can be sacrificed is great if you can deal more damage than a base snipe to the opponent if they move out, this is a great way to make sure you have lots of strength for later if the opponent does not attack.

A great way to find more tactics and problem is to just go into game and improvise something and then look at replay, or continue theorycrafting. The more the better. To be honest I am writing this part after doing this creative process and I did not write my ideas down, I am sure I had much more ideas and many of them are forgotten by now. Because I forgot the ideas that I did not use at all the process will look more linear than it is; most ideas that you come up with will likely not be used or even tested, but the more you have in the back of your head the more efficient the later steps will be, the brain is good at retrieving the appropriate ones when you need them.

The fourth step is to address some important stuff that should always be considered, this list might differ from person to person depending on how you break the game down, but I give an example. We don’t solve this yet, again we just want to keep these things in mind and at the end we go make sure everything is solved. The purpose of this step is to find potential problems and inspire more ideas to add to step 3.

Are we scouting continuously?
Can we defend against the common all ins?
How much can the opponent scout of what we do and can it be abused? (first answer question if there is a hard counter to what you are doing)
Can we tech at a normal rate and grow economy quickly at a normal rate?
What if the opponent is super greedy?

None of the above are strict requirements, but if something is missing it will typically put harsh limitations to the use of your build which will make it hard to do anything good. Ideas from the above:

oracle openers are good against all ins and they give some scouting
phoenix shuts down their scouting and also helps with scouting
oracle also lets us take a 3rd at a normal time

Most of these come down to experience, just having an oracle doesn't mean you automatically defend everything, you also need solid scouting patterns and macro etc, but this is what most people are comfortable with, and usually you need something you are comfortable with to satisfy all of the above. As can be seen from week 9 the prism + sentry can achieve the purposes above also but requires different experience. The more styles and openers you are comfortable with the more options you have.

Step 5 Here I have some ideas and it’s time to test them. I went into game where I just did something that makes sure that the requirements of the idea are met to the best of my ability while trying to increase my experience with the tactics I just listed. This also allows you to find synergies that you did not predict and see how the macro lines up, for example you might need a base faster than you expected or lack gas at a certain time etc. I will go through each of the games I played in preparation for this guide and explain my thought process as much as I can. It can be good to play someone that is normally a bit lower ranked than you, you want to win like 75-85% of these games, it is easy to find weaknesses yourself at this early stage so you don’t need the opponent to point them out to you, you’re looking for experience and positive synergies.

The vod will show each of the games and I will talk about the experience. After writing the details about the games in text form I realize that it is very boring and hard to follow in text form. I don’t recommend you read it, instead watch the vod part about the creative experimenting game phase. However, since I already wrote the text I will keep and put it at the end of the written guide for anyone that thinks it’s worth reading.

[watch vod about creative experimenting game phase]

Before I continue I want to stop and write about where we are heading here. I started out with an idea, followed by a collection of tactics. Then I tested some of the tactics, and in the process each game I also tried different openers trying to find something good and gain experience with different openers. I also spent a lot of energy trying to find a synergising way to solve the requirements for the idea, with extra focus on the “a plan if the opponent doesn’t attack” which is the most difficult one. Of course the easiest way to make a new strategy would be to use an opener that you already know and also use an end-game style that you are familiar with already. To you the opener and endgame that I chose here might sound like a new invention; they were new but also very similar to things I’ve done before (see week 9) and the lategame is based around the speedprism which I worked on for half a year by now. All I really did is choose the best opener for the job and the best endgame for the job and made some slight modifications to the endgame to synergise better with the midgame idea. Making a completely brand new opener or endgame from scratch when the main idea is in the midgame will just force you to take 2 steps back with every step forward for a very long time because it is too complex. I could have just as well chosen the oracle opener which I don’t even have to modify and it had some synergies as well, and instead of transitioning to archons I could have made tempest storm to solve the “what if the opponent doesn’t attack” and prepared phoenixes to solve the mutalisk problem and worked on the details of those instead. This would have been much easier, but I wanted to make it harder for the sake of this guide, because the more freedom you have with what to make the more fun the game is in my opinion.

Getting back to my strategy, we are now at the point where I know the strategic outline, I want to open with a fast prism and scout with hallucinations and then observers. I have sentries for early defence and try to get shield upgrades to prepare for the endgame plan while setting up a counter prism. To defend I use robo bay units, and then I want to hit a timing with charge tech, after which I enter the endgame phase.

Step 6 Now we want to play against better players to search for problems that you missed. If it goes poorly don’t be too fast to throw it all away, oftentimes there is a small adjustment that can be made to solve problems. That said also don’t ignore a problem, if you can’t find a solution to it the entire strategy might need to be changed. I will show my game 10 in the vod, which was a ladder game against Cham. Note that it is good to play against a few different people because they all view the game differently, but at the same time you want to let people adapt to it and try to play against it several times, rather than catching 15 different people off guard once with a surprise and then thinking that you have something solid.

At this point I can stop, if I want to be flexible and trust in my opening intuitions and in-game macro decisions my plan is ready. The other option is to instead make a complete build around it to make sure that you’re not making mistakes or hitting a bad timing. The downside to that is that you will be less flexible and much more predictable.

I say I can be “done”, but the truth is you can always improve it, maybe you forgot about some problems with it or you later find new tactics that can be implemented, so while this has reached a state that I want to call “viable” you should still make any improvements you think of while playing it.

I will continue by writing about how to make the build order also, at least a semi-strict build order. If you don’t care about super refined builds you can stop reading here. In order to make it easiest to focus on macro this can be done against AI, and to be safe against all ins you can set benchmark times for when they hit and make sure that you are ready or test it against a friend. I am doing this while writing, so I will now hop into games against AI and explain what I learn from games similarly to when I did the tactics against humans.

Idea before game 1:

Open nexus first with a probe scout and a zealot in the wall, chrono boost one adept to get the sentry faster. Also make a shield battery and keep building gateway units to stay safe. Prioritise minerals early to easily afford a 3rd and +1. Harass with prism and build observers. Here is a list of things I want and some things I might need.

Want (in list of priority ish):

nonstop probes
4-5 sentries
robo bay
twilight in time for +2 and +2
2nd robo
charge
gateways for the counter to time out with charge
lots of observers and a 2nd prism to hit the counter harder
save chrono
colossus/disruptor

Might need based on scouting (also kind of in priority order):

immortals
shield batteries
warpin units
extra gateways faster than counter attack is setup
chrono immortal
colossus/disruptor chrono vs a later attack.
recall prism

Because you’re always going to have to balance these against zerg that is so flexible it is hard to make 1 refined build, but I will make 1 that hits a strong timing against a zerg that plays without an attack threat, so that is what I will do. The more safe it is the more often you can do it, but the less powerful it is.

After this phase I want a strong timing, the logic for strong timings I wrote about in week 8. This time the slow unit is the robo bay units, so we want to move out with our last wave of those and then spend money on warpins, which means we need quite a few gateways. If we get templar archives early we can let our gateways be on this long cool-down after warping ht while adding infrastructure rather than building them late. With this plan in mind I go into game 1 against AI to try to make something safe/smooth that hits a strong fast timing. Again I don’t recommend reading this in text form, instead see vod on the build refining phase against ai.

All that is left to do now is to see if it is cheese-proof when executed correctly. The way I memorize the build is:

gas mining minimal after nexus first with 1 chrono before nexus

Gas spending as follows:

2 adept
robo
sentry
immortal
+1

3rd base afforded after immortal, at this point spend gas on senties + robo bay + twilight. spend any spare minerals on gates until you have 8 and get the templar archives as soon as it is afforded. And finally the robo order without breaks:

immo
prism
obs
coloss
dis
prism and gogo.

Cheese-proofing

As it opens with nexus there are several blind counters that the zerg can do, but I am only interested in non-blind cheeses. Therefore I asked a friend to do hatch first ling flood and hatch first baneling bust. You’re always looking for the worst case scenario here. Ravager attacks on a big map against an immortal and early scouting did not feel like enough of a problem for me to even test it, and anything that hits later is easily scouted. Of course maybe I forgot something, you can never be 100% sure.

Game 1 - Ling flood on 20 ish drones.

The 2 adepts find the flood of lings coming across and are recalled home after which I defend quite easily with a full wall, I am confident I can do it again.

Game 2 - Baneling bust on 20 ish drones

Similarly to game 1 the adepts find a flood of lings and I can make a full wall. I failed to recall the adepts and lose them both, but still defended with the help of the finished shield battery, sentry, and extra walling and an immortal that was almost done. Since I held without the adepts I am quite confident this is not a problem. Maybe it could be a problem if lings go around the adepts, then I would scout it maybe 10 seconds later. For the sake of explaining the process of making a plan let’s pretend that is the case. In that case we need to find a compromise in our clean build that patches up this problem as cheaply as possible. In this case the cheapest solution might be to simply make a blind full wall 10 seconds before those adepts scout and then cancel it when the adepts scout no all in 10 seconds later, that only costs around 50 minerals. Now we can say that the build is done, but if we had to make many compromises the previously clean build might now be messy and need to be refined against ai again with the new compromises in mind, which is an iterative process going between refining and cheese-proofing. But this time I would settle and call the build finished since the compromise is so small.

Now all that is left is learning execute the build. The cleaner and the better the build is the harder it is to execute, because it can fall apart if you swap the order of a single building - even a pylon, or an entire timing can be delayed by 5 seconds because you forget a single structure/upgrade by 5 seconds. It can be an interesting learning experience macro-wise to refine a build to perfection, but it is actually really hard to play it then, but feel free to experience that for yourself rather than taking my word for it. First of all we want to make sure that we can recall the important information from memory, so just hide the information I wrote after “The way I memorize the build is:” and make sure that you can quickly recall that information. Note that a build order can usually be remembered in many different ways. Breaking down the build order in a way that is easy to remember depends from person to person, sometimes you can list gas-usage (like I did), sometimes you can list order of things from a production facility (also like I did), but other times we might need to build something at a certain exact supply (often when we need to cut workers this is the easiest way of knowing exactly when), and other times when we hit timings the order of gateways/gases/forge might be a little bit flexible, but we need +1 at a certain time so the best way to perform the build then is to start the forge at an exact time so we would need to memorise a time and look at the clock in that situation. Or something might need to be built at a certain time after something else rather than looking at the clock. Looking at the clock can also help with for example sending out an hallucination at 75 energy if you know exactly when that will happen. The more time you spend preparing the easier it is to play the build, for example memorising the order “immortal - void ray - observer - phoenix - phoenix - colossus - disruptor - carrier” is only a matter of preparation time and once that is memorised it is easy to perform in game. An alternative is to say “from robo make immo - obs - coloss - dis” nonstop + from sg make “void - phoenix - phenix - carrier” nonstop. This is easier to memorise because there is usually some logic to the exact orders, but it is harder to perform in game because you don’t know if the next unit should be built from the sg or the immo and you need to constantly keep track of when each unit finishes while playing. I don’t recommend playing extremely clean builds except in tournaments since the meta changes and eventually it will need to be thrown away, but it can be a good experience to try to do it a couple of times to experience what I just wrote about first hand. The skill to quickly be able to make a build is valuable, and the skill to quickly learn to execute a build is also valuable, it’s just that the builds themselves are not valuable long-term. After breaking the build down into information that is easy to execute and memorising it you want to play it against ai. If it involves micro you need to do that micro also because you want everything to become a habit, for example “after the adept scouts their natural base build a robo” will become built into your spine. You already have the memory of the build, this is not what you’re training, what you are training is to make these habits. Play against ai and if you mess something up instantly play from replay and go back 15 seconds so you don’t start repeating a bad habit. After you can execute it against ai just do it against some people and then it should be fine. After you are done write down the things you keep in memory on a paper note so that you can look at it a week from now and execute it still.

I will here summarize the final game plan/Build overview for those that want to play it. Only do it on a big map. See build order or the “the way I memorize the build is” for the opening. After that hit a timing as planned after making the 1 disruptor and 1 colossus but scout continuously throughout this period with hallucinations/prism/obs/adepts to see if you need to play more defensively and have the 1 prism ready to counter attack with chargelots. During the timing clear creep and force out units and take a 4rth and start shield upgrades, go to a high probe count. And finally go into a lategame where you deny creep and split your army in two consisting of plenty of +3+3 archons and prism speed, possibility to transition to tempest and storm as always but there is no rush, if the opponent attacks counter with zealot/dt in the lategame, the counter prism is a threat even in the lategame. Aim for a very high probe count.

As an end note, I never really addressed the mutalisk problem in detail. What I have found in the past is that instead of getting phoenixes or blink you can get prism speed and archons. I am not saying this always works but I am confident with it in this exact situation due to my experience which is honestly too hard to explain in this guide. The important thing is that you are confident that the problem is solved for you against a good opponent, being able to explain it explicitly with some exact detailed decision tree for how to play exactly is not necessary. And because I can’t do that I am still happy with the strategy, while I understand that others might struggle with it. In terms of writing a guide for making a plan I am happy with that, but for those that actually wants to test my plan for themselves I will add an attempt of an explanation. The idea is that archons can defend decently, but more importantly your main army is much stronger than usual because you don’t invest money in phoenixes and their upgrades (fleet beacon is so expensive!). Archons in prisms are as fast as mutalisks so the opponent can also never jump on top of your main army without prisms being there shortly after even if the army is on the other side of the map. And as a bonus if the opponent does not look at their mutas for 2 seconds they can all die if the archon drops gets on top of them

Analysis SECTION

Learning how to do it

As mentioned make your own strategies and look for what the weaknesses are. Which step went wrong and why? Or perhaps you actually lack strengths rather than have too many weaknesses. If you don’t already have a million ideas for things you always dreamed of doing then I suggest doing ICYFAR (pigs funday monday) and attempting to make a good game plan for it every week (Use that as the step 1 idea).
Maybe you have a different method for making game plans already, in that case perhaps you can maybe just take some of the above as inspiration to throw into your previous method.

GL HF

NOTES TO SELF: in vod talk about why i know i need to make changes in the build, how to tell if something is good or bad. before going into creative phase explain what we are doing, testing different openers and followups, and gaining experiance with the relevant tactics.

[end of guide]

Detailed written notes on every game that was played to create the plan, I figured this might be insightful to see my exact thought process throughout, but it was also pretty boring in text form, so it is not recommended to read, instead watch vod. But since I wrote it so I will leave it here in case anyone wants to read it. You could read only the summaries also.

Game 1:

I open with something I know works most of the time, something very similar to week 9’s build but without the full wall. This gives me a fast prism after which I take a 3rd and instantly try to get charge, +1 and a higher gateway count than usual so that I am ready to counter. I also made lots of sentries to have a strong defence at home. In other words I rush towards the idea in a fairly safe manner so that I don’t waste 10 games dying to all ins before I get to experience it. After that I problem solve in-game, the opponent did not attack when my setup was ready so I was forced to think about the “what if opponent does not attack”. Based on the macro setup I had I decided to go for disruptors and a 4rth. Since I spend most energy trying to figure out the best way to do things, eg what to chrono boost, I lose the prism and have to set things up again.with a new prism. The opponent attacks after my prism dies and I am surprised by how decent sentry shield battery is against roach (better than I expected). I built 2 observers to scout since the robo was idle and that helped me out tremendously, then I can more easily stay safe at home by building units at the right time and having the optimal composition against the opponent, but also I know exactly when to send in the counter prism. At a later point that game I am forced to recall home 13 ish zealots because the opponent had enough to defend, and they helped me kill his lurker attack on my side of the map instead.