Starcraft vs. Epistemologist

 For my sins I'm watching Starcraft 2 videos. There's a ungood reason, don't ask. 

 They say the pros are bad. I have to agree. I would in fact rather love an opportunity to coach a pro, because I'm pretty sure I can make them wreck everyone else. First, they don't appreciate how all the units work. If something wasn't immediately good they dismissed it and never looked back. Second their in-battle priorities need severe retraining. Go go go will not win you a fight against someone with faster hands; stop and think sometimes. Stop and look around your base sometimes. They also have problems with their micro priorities. Spend the 1/10th of a second to make the detection follow the army rather than a-moving them into the killbox. Use f2 for retreating, not attacking. 

 

 So, e.g. void rays. Nobody uses void rays. Why not? Because you need a fleet beacon for flux vanes. Upgraded void rays are ridiculously fast, but you wouldn't know this because no pro player gets that upgrade.

 E.g. forges. Using chrono boost, the protoss can get 3/3 or even 3/3/3 long before the other two races. Aim for a late-game timing attack, very thematic. Nobody tries fast forges. Indeed they instead don't upgrade ship armour, then wonder why they die so fast to marines. "Why are all my interceptors going poof." Truly, it is a mystery. They probably don't even know that ship armour affects interceptors.
 Why don't they do this? Because, back when you started with six workers and nearly every game was low-eco, upgrades were basically a trap. Now, four bases are standard and upgrades relatively cost half as much compared to the total expenditure. Or less.

 The pros keep trying slow void rays with no weapon or armour upgrades, and for some reason they think they're bad. Again, a deep mystery. 

 Historically low-eco games were won in approximately one engagement. If you lost the army you lost the game. A fleet beacon with flux vanes costs nearly two void rays, and if you use it to get +2 +2 that's another couple void rays. You'd have like two ships feeling the power. Plus you wouldn't want to run from the decisive engagement, that just gets you shot in the back. Now nobody has gone Socrates on that assumption. Back then carriers weren't even usable. Now, getting shot in the back is something every player does every game.And yet, nobody questions...

 (They should turn and attack when it's clear they can't escape, but don't.)

 They can afford to lose whole-ass bases and lose all their observers/overseers/ravens every battle. They can afford to throw 6-8 zealots into suicide attacks, but they can't afford +1 shields. Math, how does it work?

 Protoss vs. terran probably wouldn't feel so desperate if they didn't let the terran have an upgrade advantage literally all the time.

 I just saw the "best" protoss, Maxpax, try to get tempests without +2 air attack. +2 attack is a huge breakpoint...and he just doesn't know about it. Four tempests can one-shot all the terran things you want one-shot, if you have +2 attack. But he didn't have +2. And didn't have any armour. Surprise surprise, they all got shot down and he lost the finals.  

 I saw someone try to exploit recall to expand like zergs do. Just get all the bases, so you can have all the gas. Worked great. Didn't lose. Hence, it was just the one time. Made too much sense, repelled them like crosses repel vampires.

 I mean, it makes sense. If you're a real athlete, if you don't have something seriously wrong with you, you don't try to play Starcraft 2 professionally. Serral is obviously higher class than I am, but he plays Starcraft 2, like, on purpose, meaning he must have some terrible internal scarring. No wonder they're basically bad at the game, they would be bad at anything they tried. Revenge is sour again, anyone who would be good is good enough to know better than to try at all. 

 For this sort of thing I could in theory ruin the ladder myself and inject the strategies offensively as it were, except I'm slow as balls. Asperger's is partially a nerve disorder. I misclick several times a minute due to nerve glitches, unless I slow way the hell down and double-check every button press as I'm pressing it. I need someone fast to execute my strategies for me. This sort of thing doesn't happen because, to modern mortals, taking advice feels like losing. Better to lose with your own strategy than to win with someone else's, see. 

 I have perhaps overlooked something with void rays in particular, as I haven't tried it myself. Void rays do about as much damage as two stalkers, cost as much as two stalkers (but build faster) and have lower HP and lower non-armour damage.  However, it's not just void rays, it's a whole bunch of things. The odds I'm wrong about even one of them is tiny. Void rays do stalker-level damage before prismatic alignment, they don't have concave issues since air doesn't have collision, with flux vanes they're significantly faster than stalkers, and siege tanks can't blow up your void rays. Odds I'm wrong about all of this bunch of things is basically 1 in infinity. I only need to be right about one of them to enable a pro to win a tournament. 

 Small list:
Build infinity gateways instead of suicide zealot runbys. Get bank and instantly re-max.
Every zerg should get every tech building and macro hatches, then scout and surge their weakness.
Get burrow, burrow zerglings at all their expansions. Use banelines like widow mines - cover your retreat.
Void ray strike teams. Pick up queens with graviton, blow up static D with prismatic.
Use disruptors to cover carriers against marines instead of storm.
Use disruptors with adepts instead of zealots. (Get resonating.)
Phoenix/void ray/carrier, use the excess minerals on cannon instead of dead zealots.
Void ray/carrier, use it on zealots anyway, so if they focus on killing interceptors, they get zealot'd.
Set up DT strike forces on every expansion, blink in to kill the turrets, pulling the army, then retreat and hit the next expo with the next team. Try to run terran out of scans, then send all teams in at once.
uThermal tried DT blink ambushes (huge alpha strike) and it worked great. Just needs refinement.
+2 tempest assassin quads. Blow up tanks, blow up vikings, blow up liberators. Five with +1 for zerg targets.
Defend expos with a high templar garrison. Use high-energy templar on assaults, replace with fresh garrison.
Always load the templar into a warp prism to shield against EMP and because they're slow.
Stop letting all your disruptors die, wtf fgs. Can afford multiple dead disruptors but not forge #2...
After storming, plant another storm behind the first so they retreat into it.
Use force fields to break up their concave instead of trying to box it in. Choke it a bit.
Use sentries with adepts and archons instead of stalkers. Maybe even zealots. Try to prism micro the sentries.
Garrison a sentry in the main whose sole job is scouting hallucinations. Stop losing observers constantly.
Bring a probe with your warp prism push, and build a pylon behind the battle in case the warp prism pops.
Hey since prisms are important, maybe get 2 and prism speed. Plan for it to go boom, have a backup.
Can you pop banelings by shading single adepts into the baneling ball?
Stop getting archons without shield upgrades. They just go pop.
With a large adept harass, always shade out and simply cancel the shade if they don't respond adequately.
Skip warp gate to get two stargates, or two robos, or one each, or very early twilight. Get warp gate later.
Try mass DT archon against mass phoenix? Zealot/archon. Maybe just fuckin' kill 'em. Macro nexus + cannon.
Forge/gate wall instead of gate/cyber wall. Maybe spook them into thinking you're cannon rushing. Get +1.
Blink DT + observer tumour-hunting teams. Blink away if they bring an overseer.
Final form is AllTech protoss. Four stargates, infinity gateways, four robos, research literally everything including observer speed and +3 shields. Unscoutable: can do anything at any time.

 By the way, hatches have 1250 HP and graviton lasts for seven seconds. 1250/7 = 179 DPS = 4 void rays, equivalent to eight stalkers. Voids are slower than corruptors but only barely - they can run as long as the corruptors haven't already arrived.

 I want to try a DT ambush against unburrowing lurkers. What if patience is a weapon?

 Hmm. Macro nexus purely for probe production and chronoboost on upgrades? Get enough to constantly chronoboost three forges and two cyber cores. Terran can afford many random CCs. Orbitals are 550.
 Macro nexus for scouting? They're called probes. Cheaper than observers - have them hang around watching expansions for expo timing. And more chronoboosts again.

 Videogames are epistemically important because they don't use judges. They're toy-level real worlds, which let you check your predictive power in an unmistakable way. Bonus: fast turn-arounds. If I predict e.g. that Haier's management system is the best, I have to wait around for someone to make a Haier to validate the prediction. If I predict that void rays can be good with certain conditions, I can just build some void rays and try it. Meanwhile, folk can argue that there are successful companies which aren't like Haier; there's many ways for them to easily reject sound evidence, whereas they can't argue that I didn't just blow up their expansion with my void ray strike force.