EPTPart 1: Introduction to BoX style
What is BoX playstyle?
BoX referes to Best of X, and X represents the number of times you have to play someone in a series. It is a playstyle that is designed to excell against someone if you have to play them numerous times. This is particularly handy if you play at the high level as you'll get matched against a small number of people at your end of the bellcurve. It is also handy if you happen to play on the SEA server, because only about 3 people play on that server at any one time.
Instead of doing a guide on a build order which i realized would become out of date in a short time, i decided to do it on my playstyle, which results in my use of certain types of build orders. This playstyle however has seen use since the days of brood war, especially in zvp, as we saw a plethora of build order variants that could be done off a 9pool speedlings opening. Well the idea is the same here.
Foundations of a good BoX playstyle.
The first idea you need for this playstyle is for all of your buildorders to look the same up until a point where they can't get any worker scouts into your base(this is the end of your foundation build). This is so your opponent cannot tell what build order you are going for. This is powerful in the 2nd game of a BoX series, when your opponent sees your doing exactly the same thing, and assumes you will be doing the same build order after their worker dies.
After your foundation build has finished, you pick a build order variant. Variant builds are the builds you can choose from your foundation, and your variant builds cannot be seen until they scan(only works sometimes), scout with air or you attack them. Unlike the foundation build, which you can only have 1 of, you can have as many Variant builds as possible, which your'e opponent has to guess. Here's a diagram to show how it works.
![[image loading]](http://img265.imageshack.us/img265/1764/box1uw.jpg)
example of a Foundation build:
9 pylon
14gate
15gas
18 cybernetics
19zealot
22gas
24stalker.
the foundation build stops as soon as the enemy worker is dead or gone, and no more wokers can enter your base. This is usually 10 seconds after your first stalker spawns.
example of variant builds:
how many variant builds you can do is down to your foundation build. I've decided to do a foundation build with 2 quick gasses, this allows for sentry heavy gateway builds, and any tech off 1base, (blink stalker, dt, voidray, collosus)
1)once sentry done, add citadil and 3 more gateways with continuos probe production. stop probe production at 32 and make only stalkers (4gate stalker)
2)once sentry done, add gateway and stargate with continuos probe production(2gate voidray)
The black line up to the yellow circle is all that your opponent sees with a ground scout. If we changed anything in the build there would be another black line leading into your build order variants for each thing we changed. And with the ground scout, your opponent can see these two different black lines*. The most easily seen difference is if your have got 1 gas geyser, or two.
*wether your opponent can tell the difference between these two black lines is actually the only thing that matters. If you build 1 more probe or chronoboost at a different time, then the other black line is very close to your original one, and is likely not noticed by your opponent.
Now, the reason we have the build order variants is for three main reasons:
1) to make ourselves hard to counter, and predict.
2) to predict and counter our opponents.
3) to force players to take many precautions and make sure their build is safe to all of your builds OR to make your opponent take risks by not following precautions.
Here's an example for the third reason: Your in a bo3 vs a zerg player. You go 2gateway voidray into expand the first game. They go 2 bases, then roach tech with lings and queens. they see your voids, and are now forced to make more queens and switch to hydra tech. In the 2nd game they expect the same build, as your foundation build is the same. Instead they go straight to hydra tech and get owned by a 4gate blink stalker all in build order, because their hydras did not get out in time and they had no roaches. In the 3rd game they go roach and hydra, to be safe against any of the builds you can do. This results in a weaker economy from them, and you go 3gate expand and beat them with a collosus push later on. BoX style makes your opponent choose between risk and safety.
If you only have 1 Build order variant, well you might as well only have 1 black line. Since you'll be doing the same thing every game anyway. This makes you easy to predict and counter.
example: lets say that you go for a 1gate FE build order every game vs a terran player. they go for a 2 thor/scv all in rush every game, you have great difficulty vs this build and only win 20% of the time versus particular player. resulting in a 20% winrate vs the player. If you had another variant like 2 gate robo which had an easy time beating the thor rush, then your winrate will increase greatly if you use this build instead.
The more variants you have, the harder you are to counter and predict. But it is important that you don't go overboard with these variants, because practicing a lot of build order variants takes a lot of time. and some two build orders essentially have the same effect on your opponent.
part 2: picking your variants, deduction and adaptation.
Choosing your variants: Choosing what variant build order to do is down to map positioning, your opponents playstyle and build order variety, the overall trend of game style at the moment, and recent games with your opponent.
game trends: The more popular a build order is, the more practice people have against it. It's a well known fact that everytime someone uses a new build in the gsl, people start copying it and doing it on the ladder everywhere, making it immensely popular. You should pick variants that have not seen much use as people will have less experience against it. As long as it's a good build order though
Map positioning: farily obvious, but in cross positions you'll want to do expanding builds more often as it's safer to in pvt, their 1 base timing attacks will have further to walk.
Opponents playstyle: This is a bit more complex. Im going to make up some values to explain. lets say you know your opponent has 2 different buildorders. It's the pvt matchup. One build is 1/1/1, a tank/marine/banshee/raven timing attack. And the other is 1rax expand into 2more raxs, and then to starport. an MMM based play.
lets assume that your variant that is best at beating 1/1/1 is a pheonix build, which you know can beat it about 75% of the time, but pheonix only beats 1rax expand 30% of the time. And your best build for beating their 1rax expand is a 1gate expand from yourself, which beats it 60% of the time, but only beats 1/1/1 20% of the time. Now lets imagine that they do the the 1rax expand in about 80% of the games you see, and 1/1/1 in about 20%.
pheonix > 1/1/1 at 75% winrate
pheonix < 1rax expand at 30% winrate
1gate expand > 1rax expand at 60% winrate
1gate expand < 1/1/1 20% winrate
1/1/1 done 20% of the time
1rax expand done 80% of the time.
.
phoenix wins 39% of the time (.75x.2) +(.3 x.8)
1gate expand wins 52% of the time (.2x.2) + (.6x.8)
Therefore, you should do the variant of 1gate expand every game and you will win 52% of the time. However only some opponents will be stubborn enough to do the same build over and over against you. Lets assume that you're a midlevel master league player, and that when you are laddering and you play someone twice in a row, they are 10% more likely to do a buildorder that is strong versus the build that you did the previous game. lets say you did 1gate expand the first game and they went 1rax expand. Now this game they have a 30% chance of doing the 1/1/1 build. (previously 20% of the games they did) Now doing the pheonix build would grant us a 43.5% winrate (.75x.3)+(.3x.7) and 1gate expand only 48% winrate. (.2x.3)+(.6x.7) This means the terran player can gain a 52% winrate over the protoss, as the protoss's best option is still to 1gate expand, which now only yeilds them 48% winrate because 1/1/1 is victorius more often. Lets assume we are the terran player and we want to mix our 2 build orders enough so that no matter what protoss build order they do, it will have the same chance to win versus us.
x= 1/1/1
y=1rax expand
z1=phoenix win rate
z2=i gate expand win rate
z1=.75x+.3y
z2=.2x+.6y
if z1=z2 then
.75x+.3y=.2x+.6y
y=1.83x
note also that x+y=1
i.e. 100% of games
then y=1.83 x also y=1-x, by substitution
1.83x=1-x
x=1/2.83=.353
by substitution y=.647
therefore by substitution z1=z2=.75x.353+.3x.647=.2x.353+.6x.647=.459
the maximum winrate the terrran can acheive is 54.1%, assuming the protoss will react and adapt to your build orders as well. (1-.459), this is done by going 1/1/1 35.3% of the time, and 1rax expand 64.7% of the time.
part 3: Examples of BoX playstyle.
I have covered the strengths of this build and how it functions, now it's time to see examples of it.
My favourite example is the match between MorroW and IdrA in IEM. Although this weighs more heavily on pyschological manipulation, we see MorroW do the same reaper build three times in a row, but in the last game, using the same foundation as his reaper builds but he goes for a hellion/maruader push variant instead of the reaper variant. While IdrA has finally adapted his build to the reapers, MorroW switches in a variant that completely counters idrA's adaption.
game series can be seen here.
My pvt builds:
foundation build:
all chronoboosts on probes until 24supply(goes on stalker)
9pylon
14gate
15gas
17 pylon
18cybernetics
19zealot
22gas
23pylon
24stalker
26(after stalker starts) warpgate research
sentry once stalker is done.
-end of foundation build-
*start of 1gate1robo expand variant*
robo once 100 gas accumulated
probe production stops at 31 supply.
build another sentry once the first one is finished
31-33 Nexus then pylon straight after
31-33 after nexus is built, add 2 more gateways.
build 2 observers consecutively out of robo,
send 1st to their base.
end of variant once 1st observer reaches their base. This is when my build adapts to what it scouts, and there are variants of my variant -_-.
It's important to have a wide enough range of variants so that if you know what your opponent is doing, you will have at least one variant that will be strong against it. And it's also a good idea to have a variant that is fairly 'standard', in other words it can take on almost all builds the opponent does if you slightly outplay them.
Here are 3 replays of different variants i use in pvt including the one above. They are all played against a 3100+ master league player, DrWaaagh.
http://screplays.com/replays/stroggos/16281
http://screplays.com/replays/stroggos/16283
http://screplays.com/replays/stroggos/16285
All criticism is welcome.
id on NA: kiwistroggos.330
id on sea: stroggos.825
