EPT"Why can't I just build stuff, get upgrades, make an army, expand, create a "death ball" with colossi etc and just go kill the opponent."
And yes, that does work. You don't actually need to follow a specific build to do that. Why many of us like builds so much is because they teach us one specific way of progressing through the (early) game that works well, i.e. where we take actions that fit together (almost) perfectly.
If you do things in any way where they fit together perfectly, then you can do whatever the heck you want. Figuring out how to understand the game well enough so one can play just like that is what I think most of us are after. I am at least.
I like to think of builds as "skeletons", i.e. they provide a basic structure and idea of progression to my play, and I will take that and adapt to how the game goes and the kind of ideas I get while I'm doing stuff in the game.
That is how builds are useful to me. They let me know about the strengths and weaknesses that every race has and that other races can exploit or fall prey to, and I take that understanding, learn from it, and am able to start coming up with my own builds (or "progressions") that incorporate these learnings.
An example would be:
Terrans have this really strong research called "Stimpacks". The earliest a T can get it done and have a decent sized army with it would be around 6:30 ingame time. In order to not lose to pushes that exploit this strength of the T race I need to understand what it relies upon and how I can defend it. Examples are: The push needs to be able to reach my base and units, so I can get some sentries out and force field my ramp - which also means if I had expanded before 6:30 I would need either many more sentries to forcefield a larger part of the map (the natural choke is usually a lot larger than the ramp into the main base) or some other way of defense or I would have to be able to give up my expansion in this case without losing the game right there.
So imagine almost every T would go for this Stimpack 6:30 build against a P player every time. That's when builds come in: For everything you would be doing in the first 7 minutes against a T opponent you would have to ask yourself the question "would I be able to defend against a 6:30 Stimpack push?" and you could only go down those paths (builds / progressions) that would indeed be able to defend against this kind of opening from your opponent. What specific build you go wouldn't matter, and it wouldn't need to be very specifically timed out, as long as it would meet the necessary criteria of lasting beyond the first push from T. Your understanding of the game determines the degree of freedom you have in making the choices in the game up to that point (and beyond of course).
This train of thought reminds me of an idea I remember day[9] expressing, namely that some of the questions we need to ask in order to determine a build's viability are about whether or not they are able to many of the popular and most commonly employed builds.
Of course this gets a little complicated as it requires observation and understanding of the current meta-game (= the way most people play currently, i.e. which builds they are usually going etc), but I think it's still a very useful way of looking at a build and beginning to understand its strengths and weaknesses. From there on it can be adapted and mixed and matched, during a match or beforehand.
I think there's more on builds, but I'll leave it at that for now.
