Timing windows: when to hit and when to wait

Lesson

12 of 14

Watch

A written lesson, no video

Practise

A checkpoint och a plan for your next game

A timing attack is not a build order. It is an interval: a stretch of the game where you are strong enough and they are weak enough for one fight to decide a lot. After this lesson you will be able to find those intervals on purpose, walk into them on time, and see your opponent's coming before it arrives.

What you will learn

• What a timing window is, and the four things that open one.
• How to find a window with build times and a production count, not a feeling.
• Why travel time is the number that kills most timings.
• How to see an enemy timing coming, and how to hold it cheaply.
• What a failed timing costs, and when to expand behind the threat instead.

What a timing window is

Your army and theirs both grow all game. Neither grows smoothly, because both of you keep stopping to buy things that do not fight: an expansion, a tech building, an upgrade. Every one of those purchases is a dip in the curve.

A timing window is a stretch where your curve sits above theirs by enough to win a fight. It opens and shuts on its own, whether or not you notice it.

There are two kinds, and they are mirror images.

• Your power spike. Something of yours just finished: an upgrade, a tech, a round of production. You are stronger than your army count suggests.
• Their dip. They just spent on something that does not shoot. They are weaker than their income suggests.

The best timings stack both: your upgrade finishes in the same minute their expansion goes down.

The four things that open a window

Once you know these you can predict windows instead of stumbling into them.

Two of those they choose. Two you can cause. If you never seem to find a timing, you are probably waiting for them to open one for you.

Finding a window with numbers

You do not need to guess. You need three numbers: when something finishes, how fast units come out, and how long the walk is.

Build times. These are the game's listed build times. On fastest speed, which is what everyone plays, they pass in about two thirds of the time in real seconds.

Production rate. One building makes one unit per build time. Three gateways on dragoons make a dragoon every 50 / 3, about 17 on the build clock. Count their production buildings and you know how fast their army grows. Count yours and you know how fast your window closes on its own.

Worked example: the first mutalisks. You scout a Zerg spire finishing. The mutalisks cannot exist yet. They need 40 on the build clock from the moment the eggs start, then the flight to your base. That is your window to move out on the ground, or your deadline to have turrets and anti air in place. Either way it is a number, not a hunch.

Worked example: the expansion. You see a Protoss nexus start. That is 400 minerals, which is four zealots that are not on the field. Every unit they would have had is missing from the fight you can take right now. Their army is smaller than their income says it should be, and it stays smaller until the nexus has paid for itself.

Worked example: your own spike. You have four siege tanks and siege mode is about to finish. Before it finishes, a tank is a slow unit with a small gun. The moment it finishes, the same four tanks can hold ground against almost any ground army. That minute is your window. Be walking before the research ends, not after.

The number that kills timings: the walk

A window three minutes wide with a ninety second walk is a window ninety seconds wide. Most failed timings are this mistake and nothing else.

So add the walk up front:

1. Name the window before the game. Decide which of the four sources you are hunting. "When they expand" is a plan. "When I feel ahead" is not.
2. Scout so that you arrive at the opening, not in the middle of it. What you need to know is whether they made the purchase. See lesson 7 for when to look.
3. Count only what exists. A unit still in production is on the wrong side of the line. It will be there in time for their defence, not your attack.
4. Leave early. Move out so you arrive as the window opens. The walk comes out of the window, not out of the game.
5. If the window has shut, do not attack anyway. The minerals you were going to throw at a ready defence are worth more on your own expansion. Lesson 9 is the same lesson from the other side.

Attack, or expand behind the threat

You find a window. Your advantage in it is worth some amount of their army. The other option is an expansion of your own, 300 or 400 minerals.

Attacking is the better buy when the damage you will do is bigger than what you will lose, plus the base you did not take. At most levels, most of the time, it is not.

So there is a third option, and it is often the best: walk your army to the edge of their base, make them sit at home, and put your own base down while they are watching you. You collect the value of the window without spending the army. It is the closest thing to a free lunch this game has.

Seeing their timing coming

Every timing you walk into, they saw before you did. Turn the method around.

• Count their production. Three factories and a single base means a push, not a macro game. Seven gateways on two bases means dragoons are coming, soon and together.
• Watch for missing things. No expansion when there should be one. No workers being made. Fewer units than their income should buy. That money is going somewhere, and it is usually a timing.
• Look for a finishing clock. If you saw a tech building start, you know roughly when it finishes. Write the time down in your head and scout again just before it.
• Watch the army leave. A single unit parked outside their base, or a scan of the middle of the map, tells you when they move out. Every second of warning is a second of production.

Holding a timing cheaply

You do not need to match their army. You need to make the fight bad for them until your own curve catches up.

• Fight at your own ramp or choke. A narrow entrance means fewer of their units can shoot at once. See the flux guide for why this matters so much.
• Use static defence you would have built anyway. A bunker, a cannon or a sunken colony at the right moment is worth several units.
• Pull workers only at the last moment. Every second they mine is income toward the units that will end the push.
• Keep producing during the fight. Units that pop out in the middle of the fight arrive exactly when the attacker is weakest.

Their window shuts as soon as your production catches up. Holding a timing is mostly about surviving until then.

What a failed timing costs

A timing that does not win still costs you everything you gave up to build it. You skipped an expansion, or delayed tech, or cut workers. If the attack trades evenly and pulls back, you are now behind in the economy with nothing to show for it.

This is why the choice is not attack or do nothing. It is attack, or expand behind the threat. Deciding that before you move out is what separates a timing from a gamble. And a timing that is only a feint, meant to make them buy the wrong defence, is really a bluff. That is lesson 11.

Common mistakes

• Measuring the window from your own base. Subtract the walk.
• Counting units still in production. They are not in the fight.
• Attacking because you feel ahead. Name the purchase that opened the window, or do not go.
• Staying after the window shuts. When their production catches up, pull back and take a base. A lost army closes your window and opens theirs.
• Never causing a window. Harass forces them to spend attention and money on defence. That spending is a window you made.

Practise

• Find a window you missed. In one of your own replays, find a moment where your opponent spent 300 or more on something that does not fight. Note the time, and what you were doing.
• Measure it properly. For that same moment, count only the units you had built. Take the walk out of the window. Say honestly whether the attack was on.
• Expand behind it. Play a game where you find a window and take a base instead of attacking. Compare how it goes.
• Cause one. Play a game where you open a window with a harass, and name what they bought in response.
• Test the fight. Put your army against theirs at the moment of the window in the Battle Lab, then again a minute later with their extra production added.
• Price the expansion. Use the Economy Lab to see how long a new base takes to pay back its cost.

Checkpoint