Choose a weapon
for its contribution.
Fit the system first. Then refine the weapon.
Mark and rarity help weapon performance, but are only part of the result. A weapon also needs suitable firing arcs, a compatible firing mode, adequate power where applicable, and supporting bonuses.
Choose useful weapon effects and interactions, then use actual weapon results in parses to guide refinements such as critical-chance versus critical-severity modifiers. Keep the full build and available modifier choices in mind.
A powerful effect still needs to activate.
A proc is an effect that activates when its trigger succeeds. Many standard 2.5% weapon procs get a chance at the beginning of a firing cycle, rather than on each shot in that cycle. Some special weapons follow different rules.
Illustrative comparison: one weapon, four eligible shots per cycle, a 2.5% independent chance at each eligible check, and no internal cooldown or other restriction.
| Over the same firing | 2.5% per cycle | 2.5% per shot |
|---|---|---|
| Checks in one four-shot cycle | 1 | 4 |
| Expected activations per cycle | 0.025 | 0.100 |
| Chance of at least one activation in that cycle | 2.5% | About 9.63% |
| Expected activations over 100 cycles / 400 shots | 2.5 | 10 |
In this example, per-shot checks produce four times as many expected activations. One cycle-based weapon averages one activation per 40 eligible cycles; the per-shot model averages one per 40 eligible shots.
Optional time example: a five-second cycle
Assume the same four-shot cycle takes five seconds, with continuous firing for 60 seconds: 12 cycles and 48 eligible shots. Expected activations are 0.3 per minute for per-cycle checks versus 1.2 per minute for per-shot checks.
Under these assumptions, the chance of no activation during that minute is about 74% per-cycle versus 30% per-shot. One weapon could go the whole fight without activating.
These are teaching assumptions, not a prediction for every firing mode or weapon. Haste, number of weapons, uptime, target eligibility, and effect restrictions change the result.
Expected values are long-run averages, not guaranteed schedules or literal fractional activations. If an effect permits at most one activation per cycle, the four-shot model instead averages about 0.0963 successful cycles per cycle, or 9.63 per 100 cycles. The table’s 10 counts all successful independent shot checks.
What to check: how often does the effect activate, what does it contribute, and does it support your build? A low-chance effect is not automatically worthless, and a dramatic tooltip is not proof of a large DPS gain.
Evidence & calculations
Jayiie: Weapon Procs and Calculating Chance Based Events documents cycle-based weapon checks and exceptions. The comparison above is independent illustrative arithmetic, not a recreation of a particular historical patch.
Expected activations = eligible checks × 0.025. Chance of at least one = 1 − 0.975 raised to the number of checks. The probability of zero = 0.975 raised to the number of checks.