Enter with a written question

Testing everything at once produces a collection of impressions. Start with a sentence that can be answered: does this item reduce the combination by one shot, can the ability reach from this corner, or how much stamina remains after the route? The answer should affect a purchase or action in a match. If it does not, the test may be interesting but not urgent. A narrow question also reveals which settings must remain unchanged between attempts.

Build a baseline before adding items

Reset the hero, choose the level and ability upgrades, then record damage or timing without the proposed change. Use the same dummy, range and starting health. After buying one item, repeat the exact input order several times. Critical hits, moving targets or inconsistent distance can distort a single result. The baseline turns a large damage number into a comparison. Without it, a player may credit the newest item for power that actually came from levels, range or a different opening hit.

Use the map objects that match the problem

Static hero dummies are suitable for raw damage and range. The moving dummy is better for projectile timing, while the garden contains allied and enemy Guardians with trooper waves. Neutral camps of different sizes let players compare clear speed and health loss. The gym dummy attacks with melee and supports parry practice. Choose the closest match to the live situation. A stationary target cannot prove that a slow projectile is reliable against a player who changes direction.

Use slow motion to study an action, then return to normal speed

Sandbox controls can reduce game speed, making an animation, parry cue or input order easier to see. Use that setting to identify the moment, not to declare the sequence learned. Return to normal speed and repeat until the action works with real timing. Fast cooldown and no-death options also shorten practice, but they can hide resource pressure. A combo that works only with one-second cooldowns is a demonstration, not a live build. Reset the rules before the final test.

Imitation and direct-control tools reveal both sides

A spawned hero can repeat actions or be controlled directly. This helps test whether a defensive response can interrupt a sequence and how the target sees the approach. Clear the control state when the comparison ends, because an old control setting can create confusing behavior. If a bug leaves movement disabled after returning to the Hideout, the documented remedy is to clear direct control or restart. The state of a training tool should never be mistaken for a balance problem.

Movement tests need fixed landmarks

The Sandbox includes a zipline, rope, jump pad, Cosmic Veil and teleporters. Mark a start and finish using visible edges, then count stamina and total time. Repeat the route with one movement change. Small gains matter only when they survive several runs and leave enough stamina for the next fight. The fastest route in an empty map may be unsafe in a match, so record an alternative that uses cover. Practice should produce choices, not one fragile speed record.

Convert the result into a rule for real matches

Finish by writing one short conclusion and its limit: buy this item when the combo needs the extra hit, use this angle only inside the measured range, or save one stamina charge after the jump. Then test the rule in a normal match and review whether enemy movement changed the outcome. Sandbox evidence removes mechanical uncertainty, but it cannot reproduce visibility, pressure or human reaction. A good practice result is a practical first decision that can still change when new match information appears.