The short version
This is where you get better at the game. Why you lose fights you thought were even, how to work out what your opponent is building before you see it, where your money and your clicks actually go, and how to review a game so the next one goes differently.
At the heart of it is a course in game theory, taught through the game: fourteen lessons from the class UC Berkeley ran for credit in spring 2009, Game Theory with Application to Starcraft. Every guide, lab and lesson ends with something to try in your next match, and every number is checked against the damage rules in the game's own engine.
- Weekly lessons
- 14 one a week
- Price
- Free no sign up needed
- Theory guides
- 15 ten minute reads
- Interactive tools
- 7 test any fight
- Replays to study
- 13 in your browser
Four ways to get better
-
Theory
15 guides on how the game actually works, at two depths. Every page ends with something to do in your next game.
-
Labs
7 tools backed by the real engine and the real unit data. Run the fight, measure the claim.
-
Game theory through StarCraft
The UC Berkeley course in fourteen lessons: resources, information, risk, signalling and deception, each turned into something to do in your next game.
-
Matchup guides
Build orders and what to do in a given matchup. Theory explains why they work; the guides tell you what to press.
Start here
Two pages, one of each kind. If you read nothing else on this site, read these.
-
The guide to start with
Everything is a resource
Minerals and gas are the two you can see. Supply, map control, time and your own attention are the four that actually decide games, and the class spent its whole first night on that idea. This is the guide the other fourteen rest on.
Read it -
The lab to start with
Damage Lab
Pick an attacker and a target, set the upgrades, and see how many attacks the kill takes and which single upgrade removes one. The arithmetic is read out of the engine, not paraphrased from a wiki, so the halves and quarters are the game's own.
Open the lab
What do you want to fix?
Pick the part of your game that is costing you the most. Each of these is a ten minute read with something to try at the end.
| If you keep... | Read | Then try |
|---|---|---|
| losing fights you thought were even | Dead units do not shoot | Battle Lab |
| getting stuck in chokes and ramps | Flux: chokes, arcs and the surround | Flux Lab |
| being surprised by what they built | Scouting is subtraction | Scouting Lab |
| floating minerals or getting supply blocked | Macro is units you can use | Economy Lab |
| running out of hands in big games | Your APM is a budget | APM budget |
| losing to upgrades you never noticed | Units are your eyes, ears and hands | Damage Lab |
| falling for the same trick twice | Deception is bought | Replays to study |
| not knowing why you lost | How to read a replay | your own last game |
Pick a route
-
Ten minutes
Read Everything is a resource, then play a game and name the resource you wasted most of. That is the whole first lecture, done properly.
-
An evening
Watch the short cut of lesson 1, read What micro actually is, and open the Damage Lab on the two units you lose fights with.
-
The whole course
Fourteen lessons in order, one a week: the problem sets, two of your own replays a week written up, and a final project you publish for other people to pull apart. It is the most thorough way through, and the one that makes you better fastest.
Where the numbers come from
Unit and weapon numbers come from one data file every page and lab reads, and the damage rules come from the engine this site runs on, so a lab's halves and quarters are the game's own. A number the 2009 lecturer gave from memory is marked as the lecture's, not stated as fact. Spot a mistake? Tell us and we will fix it.
Not affiliated with UC Berkeley or any person named on these pages.