Abstract
Problem: How did the critical hit and critical failure systems actually work under the hood in the original Fallout?
Approach: Tim Cain walks through his original design notes, explaining the exact formulas, data structures, and design decisions behind Fallout's critical system.
Findings: Criticals were a layered system combining roll margins, luck stats, location-based outcome tables, stackable effects, and secondary stat checks — with a clever psychological trick disabling all criticals on the first game day.
Key insight: Early QA testing revealed that players who got a critical failure on their very first combat roll perceived the RNG as broken, so Tim disabled all criticals for the first game day — and nobody noticed the missing critical successes either, proving the asymmetry of loss aversion in game feel.
1. Terminology and Structure
In Fallout's code, a critical hit or failure was called an outcome. Critical successes were based on where you hit the target (body location), while critical failures were based on the type of weapon you were using.
Both critical successes and failures shared the same core calculation for determining whether they triggered, but diverged in their outcome tables and modifiers.
2. The Critical Chance Formula
The chance of getting a critical was based on how far your roll was from the threshold, divided by 10:
- If you needed an 80 to hit and rolled a 70, that's a 10% critical success chance (10 ÷ 10)
- If you rolled a 60, that's a 20% critical success chance (20 ÷ 10)
- If you rolled a 90 (a miss, since you needed 80), that's 10 points into failure territory, giving a 1% critical failure chance (10 ÷ 10)
For critical successes specifically, your Luck score was added directly to the critical chance. Since the range was 0–100, the theoretical maximum critical chance with perfect conditions (100% hit chance, rolled 0, Luck 10) was 20%.
In practice, with a typical 80% hit chance, the roll-based component would max around 8%, making Luck the dominant factor in critical success frequency. Luck was not a factor in critical failure chance in the same way, though it adjusted the severity roll downward.
3. Damage Multipliers
Every outcome had a damage multiplier stored as an integer in half-units to save bytes (integers vs floating point):
| Stored Value | Actual Multiplier | Effect |
|---|---|---|
| 1 | 0.5× | Half damage |
| 2 | 1.0× | Normal damage |
| 3 | 1.5× | 50% bonus damage |
| 4 | 2.0× | Double damage |
Critical successes commonly used multipliers of 3 or 4 (1.5× or 2.0× damage).
4. Critical Success Outcomes
When a critical success triggered, the attacker rolled 1–100 and added the Better Criticals perk bonus (+20 per perk level). This number mapped to six outcomes per body location, roughly divided into 20-point bands, plus a seventh tier above 100 (only reachable with perk bonuses).
Every creature had a critical success chart divided by hit location (head, torso, groin, etc.), with six outcomes per location. Each outcome specified:
- A damage multiplier
- A list of effects (stored as flags, so multiple could apply simultaneously)
- A stat check the target had to roll against (e.g., Endurance for body shots, Luck for head shots)
- A secondary effect list applied if the target failed the stat check
For example, a critical hit to the head might: bypass armor AND knock the target down (primary effects), then force an Endurance check — if failed, also knock them unconscious. All three effects could stack.
5. Critical Failure Outcomes
Critical failures worked similarly but were organized by weapon type rather than hit location. A pistol would fail differently than a dagger, which would fail differently than a missile launcher.
The player rolled 1–100, adjusted by Luck relative to 5: points of Luck above 5 reduced the roll (better outcome), points below 5 increased it (worse outcome). High rolls were bad for failures.
These mapped to five outcomes in roughly 20-point bands. Unlike critical successes, there was no special sixth outcome above 100. Some weapons had no effects for the lowest (best) roll, meaning you'd "critically fail but nothing happens." Effects could stack here too — you might hurt yourself AND fall down.
6. The First Day Psychology Trick
Tim disabled all criticals — both successes and failures — for the entire first game day (the starting caves and likely the first random encounter on the way to Shady Sands).
The reason was purely psychological. During QA testing, testers who got a critical failure on their very first combat roll would march into Tim's office insisting the random number generator was broken and filing it as a bug. No amount of showing them the code or explaining the math would convince them otherwise.
The elegant solution: just turn criticals off at the start. The telling detail — nobody ever noticed the missing critical successes. Tim's takeaway: "That psychology thing works both ways." The game shipped with this feature.
7. Creature Immunities
Some creatures were flagged as immune to critical successes, critical failures, or both. Tim specifically recalled the Overseer being immune to both. Notably, the Overseer in his rising console with machine guns was a different entity from the Overseer standing in the hallway at the game's end.
8. Effect Duration
Critical effects lasted for varying durations:
- Knockouts were controlled by Endurance — higher Endurance meant shorter unconsciousness
- Crippled limbs and blindness lasted until healed via the Doctor skill
- For NPCs, permanent effects rarely mattered since they were usually killed or never encountered again — this system primarily affected the player character
9. Design Consistency
NPCs used the exact same rules as the player. There was no behind-the-scenes cheating or simplified NPC combat — the same critical tables, the same formulas, the same chance calculations applied to everyone.
10. References
- Tim Cain. YouTube video. https://www.youtube.com/watch?v=CvVWp8DmD9I