Descent playbook: filling the meter without feeding

Descent is a seasonal heretic-path system, and the Descent and Daos chapter explains where it sits among the game's combat systems. This is the deep-play article: the point economy, the transformed-state arithmetic, seal slotting and the habits that empty a window before it pays. Every figure below comes from a single corpus entry, so treat them as values to test in your own build rather than settled truth.

How do you build Descent Value without taking damage?

Every fight opens with 15 points already on the meter. From there the largest single income line is casting: every divine ability or fabao used outside Descent adds 15 points. A standard six-slot rotation therefore pays 15 × 6 = 90 points, and with the opening 15 the meter crosses 100 inside one rotation. The defensive triggers exist, but they are not required to reach the threshold.

Trigger	Points gained
Fight start	+15 (flat)
Cast a divine ability or fabao outside Descent	+15 per cast
Lose 15% of maximum HP	+5 per trigger
Your shield breaks	+5 per break

That ordering changes how you should open. A heretic cultivator who waits for the opponent to commit is using the slower half of the economy; a player who opens with a full rotation is using the faster half and arrives at the threshold first. Losing 15% of maximum HP pays 5 points and a shield break pays 5, so a bad trade is worth roughly a third of a single cast. The cast line dominates every other input, which is why an aggressive opener reaches Descent without ever being hit.

What does the transformed state actually change?

At 100 points the character transforms. Damage dealt rises by 6% and damage taken falls by 6%, a 12% effective differential between the two sides of the exchange. The reduction is a damage-reduction layer rather than a shield: it modifies incoming damage, and a shield absorbs whatever remains. Running Descent with a shield up therefore gives two separate defensive layers instead of one redundant one.

The meter drains at 8 points per second. From a full bar that is 100 ÷ 8 = 12.5 seconds, or about three 3.5-second ability cycles. No listed effect pauses the drain. The window is long enough for a short combo and too short to rebuild inside, so the decision is what to spend during it, not how to squeeze more actions into it by playing faster.

How much does Yin Deity dread add?

Yin Deity dread (阴神威煞) is a relative multiplier layered on the base state. For every 1% your dread exceeds the opponent's, Descent adds 0.5% damage dealt and 0.5% damage reduction. At a 20% advantage that is 10% and 10%, which stacks onto the base 6%/6% for a 32% effective swing during the window.

Because the check is relative, the stat has no useful absolute target. The same dread value is strong against one opponent and neutral against another, so the number to read is the gap before the engagement rather than your own sheet. That makes it a pre-fight decision: if the gap is large, the Descent window is a win condition; if it is not, the state is a flat 6%/6% buff with a short duration. In PVE the enemy value is fixed, so the multiplier is whatever the matchup allows and the state behaves as a high-uptime flat buff.

Which Descent Seal matches your kill window?

You can equip up to 3 Descent Seals (降神刻印), and each changes how the state functions rather than adding a new ability. Evil Lion extends the duration, which buys more time inside the 6%/6% window. Wicked Tiger drains the opponent's Descent Value on interaction, which is a denial effect rather than a personal buff. Demon Bear grants additional attributes during the state, concentrating extra value into the same 12.5 seconds.

The corpus lists no ranked best seal; the slot decision follows the kill window instead. If the build cannot kill inside 12 seconds, Lion extends the attempt. If the matchup is another heretic cultivator and the race to 100 decides the fight, Tiger attacks that race directly. If the combo already deletes a health bar, Bear makes the deletion more certain. All three can be run at once, but each slot spent on a seal that does not answer the actual failure is a slot that answers nothing.

Which habits empty a Descent window before it pays?

Five patterns account for most wasted windows, and all five are visible in the point rules:

• Holding abilities for a perfect moment. Every cast outside Descent is 15 points; a held ability is a delayed trigger, not a saved resource.
• Watching only your own bar. An opponent at 85 points against your 40 is about to transform. The answer is a shield or distance, not another trade.
• Treating a big hit as a mistake. A 15% maximum-HP loss pays 5 points. The hit is income as long as it does not kill.
• Breaking a shield for points on purpose. The break pays 5, but the shield is gone; a shield-reliant build should price the exposure before chasing the point.
• Skipping the dread comparison. A large dread deficit turns the 32% ceiling into the 12% floor and makes the whole window a defensive formality.

The common thread is that Descent punishes passivity twice: once by delaying the trigger and once by shortening the usable window relative to the opponent's. A player who is behind on the meter is not behind on damage yet, but is behind on the timing of the next exchange.

Can Descent trigger more than once in a fight?

Yes. The state ends when the meter decays to zero, and the meter can be rebuilt to 100 and triggered again. From zero, 7 casts at 15 points each is 105 points, roughly one full rotation. A second window is therefore a rotation-length project, not an accident, and over a long fight that is why the cast line matters more than the damage-taken line.

Two interactions are worth keeping straight. The 6% reduction and a shield are separate layers, so a shield does not reduce the value of the transformation. Casting during the state does not add points, because the income line is explicitly outside Descent; a player who keeps casting inside the window is spending the window, not refilling it. Re-triggering therefore requires leaving the state, which is a real cost in a fight that is already close.

When is Descent the wrong system to plan around?

Descent carries a season tag in the corpus, so a long-term plan that assumes it remains live is unsupported. Within a season, two builds get less from it. Shield-reliant builds pay the shield-break trigger at exactly the moment the defensive layer is lost, which converts part of their income into a penalty. PVE builds that expect the dread multiplier to scale run into fixed enemy values, so the relative check stops being a win condition. In both cases the state is still a 6%/6% buff with high uptime: useful, but not something to rebuild a character around.

For the rotations that fit inside the 12.5-second window, see the build library. For the 3.5-second cadence the window is measured against, see the combat chapter. Both pages draw on the same source values listed here, so re-check them together after a balance patch.