Item mechanics · build 6742 · research note
Lucky Shot: follow the bonus, not the tooltip.
A separate damage packet explains more than “double damage.” But it does not, by itself, prove that armor, evasion and every kind of reduction are ignored.
DLTimes Research ·
01 / What actually procs?
Lucky Shot · 6,400 Souls
25% chance · 100% bonus damage · +30% ammo. Those are unupgraded build-6742 declarations. The native Lucky callback constructs a second packet from the incoming hit’s damage and the bonus percentage; it is not simply a permanent doubling of your weapon stat.
Armor Piercing · 6,400 Souls
55% proc chance · +8% weapon damage · +60% bullet velocity. AP is also proc-based. Its matching-shot callback replaces processed damage with a separate “after-buff” report. That is different from subtracting 55 percentage points of armor on every bullet.
Lucky rolls before the hit. Its configuration requests one roll for all bullets in a shot; the pre-roll code has a shared-shot branch and a separate per-bullet branch. Do not turn a shotgun’s pellet count into independent 25% chances for Lucky. A 25% declaration also does not promise exactly one proc every four shots.
The bonus packet inherits the hit’s damage category and carries flags preventing another ordinary proc and another critical calculation. For a Bullet input, “bonus damage” is still a Bullet packet—not a conversion to Pure damage.
What the extra packet does—and does not—tell us
The ordinary callback uses the hit amount available at that event. A second matching-shot recovery callback can use the original amount when the incoming packet is marked “do not proc,” unless the source is Lucky itself. This prevents a simple “main hit was blocked, therefore Lucky must be zero” inference. The complete ordering of ordinary hits, defensive callbacks, recovery and secondary hero packets still needs validation. Headshot scaling, falloff and pellets must not be inferred from the 100% bonus alone.
Upgrade declarations change the base item settings. This comparison deliberately excludes upgrades rather than mixing those states into the shop’s base price.
02 / Does it really bypass armor?
The blanket claim is not established. In the traced current damage consumer, a newly created Bullet packet still encounters Bullet resistance and Bullet damage reduction. Lucky’s bonus and no-crit flags are not the separate ignore-resistances flag. An end-to-end bypass would need another callback, replacement or eligibility rule—not merely the existence of bonus damage.
| Defense | Current evidence | Safe conclusion |
|---|---|---|
| Bullet armor / resistance | The Bullet packet consumer applies resistance unless an actual bypass is present. | No universal Lucky armor bypass proven. |
| Plated Armor | Base settings are 30 for main-hit deflection and 50 for the broader proc check. The same stored sample drives both. | Not a flat 30% damage subtraction. Lucky’s recovery path must be included before concluding that its bonus is blocked—or unaffected. |
| Bullet evasion | Lucky and AP declare “can be evaded: false.” This value is copied into their pre-roll record. | Supports intended non-evadable procs, not yet a complete proof of damage after an evaded main bullet. |
| Bullet damage reduction | A separate factor follows Bullet resistance in the current consumer. | Not the same as armor. An armor claim cannot automatically be extended to this factor. |
| All-damage reduction | Applied later, after the examined capacity-absorption path. | No generic Lucky flag exemption found. Later replacements still matter. |
| Shields / absorption | The examined capacity path can consume a bonus packet too. | Plated’s deflection is not this capacity mechanic; not every shield’s provisioning or eligibility is reconstructed. |
Why Plated Armor is the important exception to a simple damage formula
Plated can mark a shot “do not proc” without zeroing its main damage; only the narrower threshold also zeroes that damage. Lucky has a matching-shot recovery callback for that marker. Plated also consumes its stored shot identifier. These are stateful operations, so applying an independent 30% reduction to both main and bonus damage is not justified. The exact combined result remains unresolved here.
Likewise, evasion can prevent the ordinary hit while a non-evadable proc remains eligible. Its declaration is useful evidence, but it cannot replace tracing the full event sequence. We have not independently reproduced either interaction in a match.
03 / When does Lucky make more sense?
At equal cost, compare incremental expected damage, not the biggest proc number. In the simple resisted-bonus scenario, Lucky favors lower resistance; AP’s modeled benefit grows with positive resistance. If the reported Lucky armor bypass applies, its curve changes substantially. Use the toggle to see why resolving that claim matters.
Conditional comparison, not a validated purchase breakpoint. The solid curves assume Lucky’s bonus follows Bullet resistance and a successful AP hit restores the unresisted amount. The alternate Lucky curve tests the reported bypass instead. Neither scenario models the unresolved full event sequence.
Both cost 6,400 Souls. Axes stay fixed across all three weapon-damage presets.
40% resistance
Resisted-bonus scenario
- No item
- 120.00
- Lucky Shot
- 150.00
- Armor Piercing
- 170.56
Expected normalized damage per shot, not DPS or integer HP loss.
- Lucky / 1,000 Souls
- 4.69
- AP / 1,000 Souls
- 7.90
Model leader: Armor Piercing. Not an unconditional in-game recommendation.
Exact assumptions, formulas & accessible comparison table
Let B = 100 × (1 + existing weapon damage / 100), and r = resistance / 100. The 100 is a normalization unit, not a hero stat. Both choices start at the same investment tier. AP adds 8 percentage points to the existing weapon-damage pool; it does not multiply an already-buffed gun by 1.08.
- Baseline: B × (1 − r).
- Lucky, resisted scenario: B × (1 − r) × 1.25.
- Lucky, reported-bypass scenario: B × (1 − r) + B × 0.25.
- AP scenario: (B + 8) × (0.45 × (1 − r) + 0.55).
- Efficiency: (item result − baseline) ÷ 6.4.
Unupgraded items purchased separately. Positive resistance only, body shots, no falloff, evasion, Plated deflection, shields, damage reduction, vulnerability, Fixation or other separate packets. No ammo/reload benefit, projectile-speed benefit, selling, opportunity cost, headshots or joint-item proc correlation is priced. Nominal proc chances are used as expected frequencies, not guaranteed short-run outcomes. The view is not a resistance cap.
| Resist | Lucky / 1k | AP / 1k |
|---|---|---|
| 0% | 7.81 | 1.25 |
| 20% | 6.25 | 4.58 |
| 40% | 4.69 | 7.90 |
| 60% | 3.13 | 11.22 |
| 80% | 1.56 | 14.55 |
Reasons to consider Lucky
A gun already doing substantial damage, low-resistance targets and value from its larger magazine. More existing weapon damage also makes AP’s additive +8 percentage points a smaller relative gain. These are model and stat considerations, not a guaranteed hero-specific ranking.
Reasons to consider AP
High positive resistance in the comparison model, plus a use for projectile velocity. Its 55% chance is not a universal DPS multiplier; its value depends on what a successful replacement actually restores in the situation.
Averages do not guarantee a short burst or a kill. Ammo can improve sustained firing without raising firing DPS, and projectile speed can help practical hit rate without changing the damage of an already-landed shot. Neither benefit is assigned an invented Soul value here.
04 / Fixation is not the same kind of stack.
Haze · Fixation
The traced Fixation path reads existing stacks and a per-stack damage property, scales by the hit’s effectiveness and creates a separate Bullet packet. It marks that packet “do not proc,” then handles stack addition separately. This is not simply extra weapon damage folded into the gun stat.
Lucky’s matching-shot recovery path makes this separate packet relevant. But it does not justify “Lucky gives double Fixation stacks,” “every Lucky shot doubles all Fixation damage,” or a universal extra-stack multiplier. Callback order and eligibility still need to be joined.
Vindicta · Assassinate kill stacks
The selected kill-bonus callback contributes through the weapon-damage-increase channel, with a stack count multiplied by the evaluated bonus. The captured property is 6 per kill, with an upgrade declaration adding 4.
That is ordinary weapon-stat scaling, not evidence of a special Lucky-only stack multiplier. Flight’s separate Spirit damage and Assassinate’s ability damage must not be folded into the same bullet number without tracing their own paths.
Drifter · Hunger stacks
Hunger’s kill bonus also contributes through the weapon-damage-increase channel. The captured isolation-kill property is 3, with assist weighting represented separately in its callback. Its isolation damage amplification is another channel, not another copy of the kill-stack stat.
Practical distinction: a weapon-damage stack can increase the bullet amount supplied to both item comparisons. A separately emitted damage packet may interact through callbacks instead. “Both heroes stack damage” does not make their Lucky interactions equivalent, and this article does not claim current live validation of those combinations.
05 / Evidence, without the overclaim
Frozen client/server build 6742, revision 11077051, dated October 2 and inspected October 3, 2026. Authored definitions establish the displayed settings. Hash-bound static inspection establishes the selected packet, property and defense branches. Neither is an official explanation from Valve or an independent live-game measurement.
Scope, unresolved questions & what would settle them
- No game or server was run; no replay or match was acquired.
- The reported armor, Plated and evasion bypasses are investigation questions, not findings adopted from the request.
- Outstanding: full ordinary bullet dispatch, evasion-to-proc ordering, Plated recovery ordering, Fixation-to-Lucky ordering and joint-item behavior.
- A controlled comparison would need identical body shots and stats, logged ordinary/proc packets, known defense states, and enough shots to distinguish expected rates from burst variance. That experiment has not been performed.
- The efficiency chart is a sensitivity model. Passing website tests validates its arithmetic and rendering, not the game mechanic.
- Other modes, upgrades, arbitrary shield sources, immunity states and exact integer health mutation remain outside scope.
Curated implementation evidence is recorded in the repository’s docs/research/lucky-shot-6742.md. Raw binaries and extracts are not included in the article.