Invoker Eul Combo: Timing, Slot Preparation and Practice

The Eul setup combines three different tasks: preparing the right spells, aiming at a predictable landing point, and releasing each spell at a suitable time. Learning the keystrokes is useful, but it does not make those tasks interchangeable. A fast invoke can still produce a missed Sun Strike, and a correctly timed Meteor can still leave a surviving opponent.

This guide follows Sun Strike, Chaos Meteor and Deafening Blast from preparation to review. The queue examples describe reproducible inputs in Invoker Game. The timing example is an explicitly simplified calculation, not a claim that we measured the current Dota patch or an unavoidable kill threshold.

Start with a repeatable test, not a kill promise

In Hero Demo, choose one stationary enemy target and keep its level, items and defensive state unchanged between attempts. Set Invoker's abilities and inventory, then record those settings. Start without a Blink approach, allied damage or multiple targets. Removing those variables makes it possible to tell whether a miss came from your release timing or something else.

Test Sun Strike with the lift first. Add Meteor only after the first spell lands consistently. Add Blast last. If the complete chain fails immediately, you otherwise have three timing decisions and an aiming decision to diagnose at once.

A useful goal is ten attempts with the same conditions and a written count of hits, early impacts and late impacts. Ten is a convenient practice sample, not a proficiency requirement. Repeatability matters more than your single best-looking clip.

Calculate the release time from the landing time

Define the start of your timeline as the moment the target enters the cyclone, rather than the moment you begin reaching for the item key. Those events can differ because of movement, targeting and command processing.

For a simplified landing calculation, define:

The planning relationship is input time = L + epsilon - d - c. This is a way to organise measurements. It does not model every engine interaction, and it does not justify claiming that two hits occur on exactly the same server tick.

A worked example with explicit assumptions

Suppose your test uses a lift lasting 2.50 seconds, a Sun Strike release-to-impact delay of 1.70 seconds, and a Meteor delay of 1.30 seconds. Ignoring input-to-release time and boundary adjustment for this illustration gives:

0.00  Target enters cyclone: timeline begins
0.80  Sun Strike RELEASE: 2.50 - 1.70
1.20  Chaos Meteor RELEASE: 2.50 - 1.30
2.50  Intended landing boundary

These are release marks under the stated assumptions, not universal keypress instructions. Meteor follows Sun Strike by 0.40 seconds in this example; it is not cast immediately after it. If your measured delays differ, substitute them into the equation instead of preserving the old marks.

Blast needs its own observation because its impact also depends on the wave reaching the target. Do not append an arbitrary fixed timestamp to the table and assume it establishes a gapless disable. Aim, distance, enemy responses and the current mechanics all matter.

Prepare the slots and the queue separately

A useful starting state for the execution example is:

• D: Sun Strike.
• F: Chaos Meteor.
• Queue, oldest to newest: EEW.

The active queue does not have to match the most recently invoked spell. You can change orbs without invoking again. That distinction lets you prepare the next transition while keeping two useful spells available.

To build this state from an empty trainer queue, invoke Meteor with EEW R, then Sun Strike with EEE R. The new Sun Strike occupies D and Meteor moves to F. Press W without Invoke: the queue changes from EEE to EEW, while the slots remain unchanged. In Dota, allow for the actual Invoke cooldown while preparing; the browser does not enforce that wait.

Do this before approaching the opponent. Preparing in range of an enemy turns a quiet setup task into a reaction test and increases the chance of an unintended cast.

Execute the sequence without losing the slot state

1. Lift and identify the point. Use Eul in Demo and keep track of the target's landing location. Start from the same position for the first set of attempts.
2. Release Sun Strike at its calibrated mark. It is in D in this example. Aim at the landing point rather than at the elevated visual model.
3. Wait for Meteor's separate release mark. Use F, and observe the direction in which the meteor travels after landing. In the simplified example above, its release is 0.40 seconds after Sun Strike.
4. Prepare Blast from the known queue. From EEW, pressing Q removes the oldest E and produces EWQ. Press R to invoke Blast into D. Sun Strike moves to F; Meteor leaves the two slots.
5. Cast Blast when its contact is useful. Use its current slot and account for the wave's travel. Then reassess the opponent rather than automatically spending another spell.

The two-key transition Q R is valid from EEW, not from every permutation of the Meteor recipe. Starting from WEE, one Q gives EEQ, which invokes Forge Spirit. Both queues originally invoke Meteor, but they are not equivalent starting states for the next input.

This is why reading only the spell icon is insufficient when learning orb economy. The spell depends on orb counts; the next queue depends on order. If you lose that order, deliberately rebuild the required recipe instead of repeatedly invoking the wrong one.

Measure damage under stated conditions

Replace a universal “HP you can kill” table with your own controlled observations. A target's resistance, barriers, avoidance, nearby units, defensive effects and time spent in the Meteor path can change the result. Even nominally pure damage is not a promise that every protection is irrelevant.

For each trial, record starting HP, ending HP, whether all intended spells connected, and whether a defensive effect triggered. Separate an initial burst reading from later damage over time. Do not compare one attempt measured immediately after landing with another measured after the target has remained in the burn area.

If you calculate damage, document which components receive each multiplier and which were actually observed. Do not turn an estimated number of Meteor contacts into a guaranteed kill threshold. The practical decision is whether the commitment is worthwhile with your information, not whether a table promises success.

Add positioning and defensive responses one variable at a time

Once the stationary test is repeatable, change your starting angle, then add distance, and only then add a Blink approach. An approach can alter facing, targeting and whether you are already within cast range. There is no universal turn-time saving that can be inferred just from a keyboard sequence.

Against spell block, verify which action actually consumes the block before committing Eul. Against a defensive trigger such as Aeon Disk, watch for the effect and decide whether to hold follow-up spells. Against debuff immunity or an ally's save, test the relevant interaction in the current client rather than assuming that a damage type determines every outcome.

Do not apply the generic formula “duration × (1 − status resistance)” to every lift. First establish whether that specific effect is modified in the current version. If the target's landing changes, measure it and rebuild the timeline. If it does not change, moving your releases earlier can create the very miss you were trying to avoid.

Diagnose the miss before repeating the whole combo

A failed trial is useful when it changes the next trial. Repeating the whole combo faster without identifying the miss usually hides the source of the error.

Train the keyboard portion on Invoker Game

Open Combo mode and use the available Meteor–Blast combination for orb reuse. Its task is to invoke the shown spells, not to cast Eul or aim a projectile. Begin with EEW R → Q R: three initial orb presses, one transition orb and two Invokes. Compare it with rebuilding Blast using three fresh orbs. Review the trainer's efficiency result, not an imaginary landing-time score.

Next, use Combat mode for slot reading. It starts from its own generated state, so it is not a configurable replica of the preparation table. Read the shown queue, D/F and targets before acting; do not assume the tutorial's opening is present.

Finish in Hero Demo with the client timings, aiming and resources restored. Keep a short practice note with the conditions, what failed and what you will change next. The browser makes your inputs more deliberate; the in-game experiment determines whether the actual setup works.