Alertness & target selection
An enemy being alert, remembering you, and actively targeting you are different states. Losing one does not always clear the others immediately.
- Current target
- 1.0×
- New target
- 0.9×
- Ranking distance
- 0 at 50 m
Alert state vs target knowledge
Making an enemy alert is not the same as giving it your exact position or making you its active target.
A nearby alert does not automatically reveal your position
One confirmed 20 m relay can raise nearby enemies' alert state without giving them your identity or exact coordinates.
| Mechanic | Range / condition | LOS required? | Carries exact target/location? | What it does | Player implication |
|---|---|---|---|---|---|
| Hearing propagation | Sound's base range × enemy/live modifiers | No LOS test in examined path | Sometimes source + event position; sometimes position only | Updates hearing/perception information and listener response strength. | Walls can break sight without blocking hearing. |
| Nearby alert relay | 20 m 3D sphere | No | No | Raises a qualifying nearby alert state without supplying a target position. | Nearby enemies can react through cover without being handed your coordinates. |
| Shared squad knowledge | Group/perception-owner relationship; not a single radius | N/A | Uses shared target information | Several enemies in the same AI group can use the same stored information. | Shared information does not force every enemy into the same alert state or target. |
| Target-reconsideration relay | One confirmed producer uses a 25 m member gate | Not established as a general LOS rule | No copied target record | Makes the recipient reconsider target information it already has. | An enemy can suddenly act on information it already had. |
| Other group relays | Confirmed ranges include 8, 10, 20, 25 and 30 m for different reactions | Path-dependent | Path-dependent | Different group reactions use different distance limits. | There is no single universal 'shared aggro radius'. |
| General target priority | Choosing among already-known targets; distance factor reaches 0 at 50 m | N/A | Selects among targets the enemy already knows about | Uses distance, facing, current-target advantage, and other validity checks. | 50 m is not a universal de-aggro distance; it is one ranking factor. |
| Projectile Danger | Closest trajectory distance d relative to warning range R | No ordinary LOS test in recovered calculation | Can be entity-backed + cached position; distinct from direct visual confirmation | Full contribution at d ≤ 0.5R; linear falloff to zero at d = R. Strong Danger can trigger a direct under-fire response on enemies configured to return fire. | A quiet near miss can make some enemies react to an attributed source before normal visual spotting completes. |
| Breaking LOS | Visual channel only | Yes for visual sensing | Visual loss clears the visual channel, not the whole source record | If no other usable way to track the target remains, its priority becomes 0. If another tracking source remains, the target can stay eligible. | Breaking sight can make an enemy switch targets, but only after it loses its last usable way to track you. |
Alert modes
The state machine has ordinary awareness modes plus specialized return-fire/contact modes.
| Mode | Current functional meaning | What distinguishes it |
|---|---|---|
| Baseline | Unaware | Can transition through suspicion, investigation, nearby alerts, or a direct under-fire response. |
| Suspicion | Visual spotting in progress | Accumulates visual spotting progress before ordinary combat awareness. |
| Investigation | Checking a remembered source position | Can be driven by hearing or lower Danger without requiring a currently visible target. |
| Combat alert | Standard combat awareness | Can be reached through direct confirmation or relayed alerts, so it does not mean the enemy currently sees the Helldiver. |
| Search | Last-known-position search | Uses remembered position, returns to combat alert on fresh evidence, and eventually falls back toward investigation. |
| Direct under-fire response | Urgent return-fire behavior | Available to enemies configured to return fire after sufficiently strong attributed Danger. |
| Relayed under-fire response | Urgent peer reaction | A nearby enemy can relay its own endangered state and position without directly transferring the original shooter coordinates. |
| Post-contact persistence | Extended return-fire persistence | Can persist after fresh evidence is lost. Its confirmed 16-second threshold is conditional rather than an unconditional total duration. |
Under-fire responses are separate from normal combat alert.
Strong Danger can trigger a direct under-fire response on enemies configured to return fire. Nearby allies can enter a relayed response, and some enemies can remain in a longer post-contact persistence mode after fresh evidence is lost.
| Mode | What source information is established | Important limitation |
|---|---|---|
| Direct under-fire response | Danger can install an entity-backed source record with a cached source position even without ordinary direct-perception evidence. | A selected perception source should not automatically be described as a continuously visible weapon target. |
| Relayed under-fire response | The relay builds a fresh source record from the nearby state-5 ally's entity ID and position, then installs that ally as the receiver's active perception source. | The transition does not copy the original shooter's source record. A later replacement through ordinary perception and target ranking remains unresolved. |
| Post-contact persistence | All three captured state-7 enemies retained an installed selected entity source while the ordinary direct-perception flag was absent. | A generic source-property flag and the separate state-7 timer gate are functionally proven, but their original internal names remain unresolved. |
The direct under-fire response can begin from low awareness without completing normal visual spotting. Danger can retain an attributed source and a cached source position, so the enemy can react to incoming fire without continuously seeing the Helldiver.
The direct and relayed under-fire responses use an 8-second timeout threshold. Post-contact persistence uses a 16-second threshold, but another confirmed condition can hold its timer at zero. Fresh direct evidence returns these modes to normal combat alert; timing out returns them to search behavior.
Group-reaction ranges
Each range belongs to a specific reaction. There is no single shared “aggro radius.”
- 8 m: one direct-alert and coordination path.
- 10 m: another short-range coordination relay.
- 20 m: the confirmed nearby alert pulse.
- 25 m: one target-reconsideration trigger.
- 30 m: additional coordination and membership gates.
Target reconsideration
It rechecks its own known targets; the sender does not copy a new target into it.
- Distance term
- 0–10 m
- Facing term
- 0–180°
- Score
- distance × facing
- Source set
- Existing records only
Linear from 1.0 at 0 m to 0.0 at 10 m.
Linear from 1.0 at 0° to 0.0 at 180°.
Highest strictly positive score can be installed as the selected source.
This relay does not copy a new target record from the sender.
One target-reconsideration relay has a separate 25 m membership/proximity gate. That trigger distance is not the same thing as the 10 m source-scoring curve inside the recipient.
Distance × facing score
This calculation ranks already-known target information. It is not the formula for whether an enemy initially sees you.
0.600 × 0.750 = 0.450 Brighter cells have a higher calculated score. The marker shows the current slider values.
Distance factor = max(0, 1 − distance / 10). Facing factor = max(0, 1 − angle / 180). The relay score is their product.
Target ranking
The game first checks whether a target is still valid, then applies distance, facing, current-target advantage, and other factors. If a required check fails, that target's priority becomes zero.
Current targets have an 11.11% priority advantage
The current target receives 1.0 while a new target receives 0.9. A new target therefore needs more than 11.11% better scoring from the other factors to replace an otherwise-valid current target.
| Factor | Input | Priority factor | Evidence note |
|---|---|---|---|
| usable target information | at least one usable way to track the target remains | 1.0 | The target remains eligible. Sight is only one possible source of information. |
| usable target information | no usable tracking information remains | 0.0 | Priority becomes zero even if this was the current target. |
| valid target | target is valid for this enemy | 1.0 | Otherwise its priority becomes zero. |
| current-target advantage | already the current target | 1.0 | The current target keeps the full factor. |
| current-target advantage | a different target | 0.9 | A new target needs more than 11.11% better scoring from the other factors to replace an otherwise-valid current target. |
| large target class | mapped large-unit classes | 0.5 | Mapped large-unit classes halve target desirability. Ordinary Helldivers do not receive this penalty in examined samples. |
| 3D distance | ≤ 5 m | 1.0 | Full distance priority. |
| 3D distance | 25 m | 0.8 | Priority falls gradually from 5–25 m, then more quickly from 25–50 m. |
| 3D distance | ≥ 50 m | 0.0 | Distance contributes zero here; 50 m is not a universal disengagement distance. |
| facing angle | 0° / directly ahead | 1.0 | Maximum facing priority. |
| facing angle | 45° | 0.9 | The first defined facing step. |
| facing angle | 180° / directly behind | 0.8 | Facing changes preference but does not hard-exclude a target. |
The number of tracking sources is not added into a threat score here. A target either has usable tracking information or it does not. Damage, noise strength, weapon visibility, health, armour, stance, and recent attacks do not directly change this target-priority calculation.
Distance weighting
Priority stays full through 5 m, falls slowly to 25 m, then falls faster to zero at 50 m.
| Distance | Priority factor |
|---|---|
| ≤ 5 m | 1.000 |
| 10 m | 0.950 |
| 15 m | 0.900 |
| 20 m | 0.850 |
| 25 m | 0.800 |
| 30 m | 0.640 |
| 35 m | 0.480 |
| 40 m | 0.320 |
| 45 m | 0.160 |
| ≥ 50 m | 0.000 |
Priority stays at 1.0 through 5 m. It falls by 0.01 per metre from 5–25 m, then by 0.032 per metre from 25–50 m. At 50 m, distance contributes zero to target priority.
This is a target-priority boundary, not a universal combat, memory, or disengagement distance.
Facing weighting
Facing changes priority gradually, but much less strongly than distance or the 0.5 large-target penalty.
| Relative direction | Angle | Priority factor |
|---|---|---|
| Directly ahead | 0° | 1.000 |
| 45° off-axis | 45° | 0.900 |
| Perpendicular | 90° | ≈ 0.867 |
| 135° behind | 135° | ≈ 0.833 |
| Directly behind | 180° | 0.800 |
A target directly ahead receives 1.0. At 45° the factor is 0.9, then it declines gently to 0.8 directly behind. At equal distance, a new target directly ahead can only overcome the current-target advantage once the current target is more than 45° off-axis.
Target persistence
Losing visual contact removes sight as a reason to keep you targeted. Other usable tracking information can still keep you eligible.
| Situation | Current target | Other target | Result |
|---|---|---|---|
| Current target can still be tracked | 1.0 priority factor | 0.9 priority factor | A new target must overcome the current target's 11.11% advantage through distance, facing, size, or another factor. |
| Current target loses sight, but another tracking source remains | Still eligible × 1.0 | Eligible × 0.9 | The current target can remain preferred despite no direct visual contact. |
| Current target loses its last tracking source | 0 score | Any positive score | Any other valid target with a positive score wins. It still receives the 0.9 new-target factor during that update. |
| No target has a positive score | 0 score | None | The selected target is cleared. The enemy can still remember information or continue investigating separately. |
Memory is not target eligibility
When sight is lost, visual tracking can disappear before the enemy forgets you entirely. If no other usable tracking source remains, target priority becomes 0 and the current-target advantage cannot save it.
A replacement target still receives the normal 0.9 new-target factor during that update. Any positive score can still beat the former target's zero score. Once selected, the replacement becomes the new 1.0 current target.
If no target has a positive score, the selected target is cleared. The enemy may still remain alert, investigate a last-known position, or remember information about you; those states are separate from actively targeting you.
A target switch is not guaranteed on the exact frame that vision fails. The confirmed update paths refresh every 0.25 s or 1.0 s depending on an unresolved state, so a short delay can occur.
Lure Mine perception
| Case | Established result | Important limitation |
|---|---|---|
| Lure Mine | The deployed Lure can appear as a real target. | The confirmed 2 m / 90° values control detonation, not the distance from which enemies are attracted. |
Unsupported claims
| Common claim | What the evidence shows |
|---|---|
| Enemies share your exact position inside 20 m. | The confirmed 20 m relay raises nearby enemies into combat alert without carrying the Helldiver's identity or coordinates. |
| There is one shared-aggro radius. | Different group reactions use 8, 10, 20, 25, and 30 m ranges for different purposes. |
| 50 m is the de-aggro distance. | 50 m is where distance stops contributing to one target-priority calculation. It is not a memory or disengagement boundary. |
| Breaking line of sight always makes an enemy change targets. | Losing sight removes visual tracking. If another usable tracking source remains, the current target can stay eligible and keeps its 1.0 priority factor. |
| Enemies keep targeting a hidden Helldiver because of the current-target bonus. | If the hidden Helldiver loses every usable tracking source, target priority becomes zero. The current-target advantage cannot keep a zero-priority target selected. |
Open questions
| Question | Why it remains unresolved |
|---|---|
| When is a known target forgotten entirely? | Losing the last usable tracking source can remove target eligibility before stored information about that target is deleted. The full forgetting process is not mapped. |
| What determines the 0.25 s versus 1.0 s target-update interval? | Both timing values are confirmed, but the state that chooses between them is not yet identified. |
| Does the relayed under-fire response later identify the original shooter? | The initial relay identifies the nearby endangered ally. A later transfer of the shooter identity has not been proven. |
| How does the Lure Mine attract enemies from farther away? | The Lure can become a target, but the mechanism that creates long-range interest has not been isolated. |
Evidence standard
How the OTL separates code confirmation, controlled live testing, and behavior that remains unresolved.
How the OTL rates evidence
The OTL rates code evidence and live-game evidence separately. A mechanic can be clear in code but still need a controlled gameplay check for hidden state, ordering, or another runtime factor.
Live confirmation means controlled gameplay or instrumented runtime observation, not player impression alone.