The Concept
Timing is the timed production and control of all units such that a player can maximize his advantage at a certain relative game time. Having good timing means capitalizing on your current situation in the game - for example, a numerical, positional, or tech advantage.
 Static defence timing
Static defense timing is best understood through a simplified example scenario: imagine you (playing Terran) know your opponent is going 3 Hatch Muta, and you also know exactly when they will arrive at your base. Because you did an FE build, you are going to need Turrets. Assume (ignoring Mutalisk micro and Turret placement) that Mutalisk and Turret power scale according to the same linear function. Also assume each Turret kills 3 Mutalisks.
Obviously, to defend against the Mutalisks with Turrets, you will need 3 at each base. When should you build them? You want all 3 Turrets to finish exactly when the Mutalisks are within range of your base, because this will guarantee your base's safety while also delaying incursion of the Turrets' construction cost as long as possible. This is the most basic type of timing. To delay Turret construction further at one of your bases (main or natural) you can predict which base the Mutalisks will strike second and build Turrets at that base slightly after you build Turrets at the base the Mutalisks will strike first.
By waiting as long as possible to build Turrets, you can afford to build more SCVs and more infrastructure before sinking resources into static defense.
 Static defence buildup
In the above example scenario, you had perfect information (you knew exactly when the Mutalisks would arrive at your base). Timing with perfect information is nothing to brag about – you can see how long it will take the enemy's forces to reach you and can prepare accordingly. Timing with limited information is much more difficult.
With no initial scouting, should a Zerg build 5 Sunken Colonies (enough to stop a Sparks attack), or 2 Sunken Colonies (enough to stop a 1 Barracks Academy rush)? To a player with proper timing, this is not an either/or question. Timing static defense buildup means constructing 2 Sunken Colonies at the 1 Barracks Academy rush timing, and adding Sunkens such that their completion coincides with subsequent potential Terran attack timings (for example, to total 5 Sunken Colonies at the Sparks timing). This concept also applies to a Protoss player defending his Forge FE from various early Zerg attacks: rather than add many cannons at once (as in the simplified Turret example above), the Protoss player should add cannons gradually according to the Zerg attack timings.
 Attack timing
Attack timing is more complex than static defense timing. When a player is praised for "great timing", "timing" usually refers to attack timing.
Consider a standard Zerg versus Terran game: Terran expands with 1 or 2 Barracks and 1-6 Marines while Zerg goes 12 Hatch 11 Pool 13 Hatch, making all Drones and heading towards Lair. Terran applies pressure until Mutalisks are out, when he returns to his base to defend. Zerg continues to expand, gets Lurkers, and techs to Hive and Defilers. Terran eventually moves out with Tanks, Marines, Medics and Vessels to fight Lurkers, Zerglings and Defilers. Eventually one side is pushed back, loses an expansion, and concedes the game.
Now, let’s identify when each player is weak, starting with Zerg. Once Medics and Stimpacks are researched, Zerglings can no longer deal with the Terran army. Similarly, once the Muta harass is over, in the time before Zerg can defend his expansions with Lurkers, he is weak. He is weak once again during the transition when he is waiting for Defilers. These are all key times for the Terran to attack. Watch for it in games of strong players, I guarantee Terran will move out at exactly all three of those times, because they are so crucial times when Zerg is weak.
We can also see when Terran is weak. Once Mutalisks are out, Zerg has free reign of the map and Terran is weak. Similarly, once Lurkers are out Terran often doesn’t have Tanks or Vessels to deal with them, so they are again weak. Terran is weak again when Defilers initially come out, because he shouldn’t have a great deal of tech to be able to easily fight them at that point. Unlike Terran, these are not necessarily points when Zerg should attack. Terran’s “weakness” may come from loss of map control, and Zerg should use that to expand. In other cases, such as Defilers, it is an opportune time to attack.
You should always be using timing to your advantage. If you FE as Terran, you can assume Protoss will double Nexus in response. If you time your army to attack at the critical moment when he has paid for his expansion but has not received any benefit from it, you may be able to break it.
This is the critical idea to timing attacks – you attack someone when they have spent resources but have not yet gotten any benefit from it. If you mistype your attack too soon they will not have spent the resources and will be able to crush you, but if you wait too long they will have the added benefit of whatever they spend resources on (new expansion – more minerals and hence more units, or new tech – able to counter your attack because you haven’t teched because you were doing a timing attack). Players relying on timing attacks walk a thin line between attacking too early or too late, but if timed right timing attacks are the most devastating attacks in the game.
 Production timing
Production timing is the balancing act of infrastructure, tech, economy, and army size. The theory of it is far too complicated to go into extensive theoretical detail, so instead we will rely on vague theories and examples.
Production timing is like static defence timing but with more options and is heavily tied to attack timing. If you are planning on attacking in 3 minutes, should you cut workers now and build only units, or cut them in 2 minutes and build units? Which gives you the largest possible army 3 minutes from now? If you know when you are going to be doing a timing attack, production timing makes that army as strong as possible. An example of this is FireBatHero versus sAviOr on Katrina. FBH cuts units and Turrets in order to get seven Barracks producing units by the time Mutalisks are out. The result is a midgame timing attack before Lurkers that sAviOr’s Mutalisks cannot hold back. This is a true timing attack, but the main thing to take awhile from this game is the timing of production. FireBatHero maximizes his economy by making as few units as possible, and then transitions into maximizing his army by making few workers and static defence as he approaches the time to attack.
 Counter-production timing
StarCraft often works in a Rock-Paper-Scissors cyclical nature. If you look at what I’ve said earlier, Zerg is weak after the Mutalisk harass is defended and they don’t have Lurkers out. Why? Because they’ve spent time and resources (on buildings, research and morphing Lurkers) and they haven’t got any benefit from this. So a Terran timing attack at this point would be strong.
But as a Zerg player, you know you’re weak to this timing attack. So what if you time your army to be maximized when his timing attack comes? This is the essence of counter-timing. As difficult as timing is to master, counter-timing is that much more difficult.
Let’s examine the possibilities in the Zerg versus Terran scenario. If all goes according to plan, Terran moves out with his timing attack, but Zerg hasn’t even made a Hydralisk Den. He has instead decided to focus on pure MutaLing. Terran moves out, expecting to find a dwindling Zerg economy stalling for time while Lurkers are morphing, and instead finds a massive MutaLing army that crushes him.
Let’s replay the scenario with a smarter Terran. Terran first notices that Zerg is still making Mutalisks through this early midgame. After scouting, he realizes there is no Hydralisk Den, so Zerg must have shifted his production for a counter-timing attack. As a result, Terran shifts his timing. He will wait until Zerg places an expansion or Hydralisk Den and then attempt his timing attack yet again.
- [G] Timing by Chill.