Masks are a way to **encode human behavior into a constrained search space**.
They are not guesses at individual passwords, they are hypotheses about **how people tend to structure passwords** under real-world conditions.
This page documents a set of empirically derived masks based on analysis of a **2.7 billion–record password corpus**, along with guidance on how and why these masks matter.
---
## **Why These Masks Exist**
When large password datasets are analyzed at scale, certain structures repeat with extreme regularity.
These repetitions are not accidental, they reflect:
- Minimal-effort compliance with password policies
- Cultural and regional numeric preferences
- Language-specific defaults
- Predictable placement of digits and characters
The masks below represent **high-density regions of the human password space**.
> **Key Result:**
> These masks alone recovered **~6% (≈162 million)** plaintext passwords from the dataset.
That recovery rate was achieved **without rules, dictionaries, or behavioral labeling** only structural constraints.
---
## **US / Western-Centric Mask Plans**
These masks reflect trends commonly observed in US, EU, and similar regions where:
- Alphabetic characters dominate
- Digits are appended rather than interleaved
- Lowercase is preferred unless policy forces otherwise
|**Mask**|**Description**|
|---|---|
|?l?l?l?l?l?l|6 lowercase letters|
|?l?l?l?l?l?l?l|7 lowercase letters|
|?l?l?l?l?l?l?l?l|8 lowercase letters|
|?l?l?l?l?l|5 lowercase letters|
|?l?l?l?l?l?l?l?l?l|9 lowercase letters|
|?l?l?l?l?l?l?l?l?l?l|10 lowercase letters|
|?l?l?l?l?l?l?l?l?l?l?l?l|12 lowercase letters|
|?d?d?d?d?d?d|6 digits|
|?d?d?d?d?d?d?d?d|8 digits|
|?l?l?l?l?l?d?d|5 letters + 2 digits|
|?l?l?l?l?l?l?d?d|6 letters + 2 digits|
|?l?l?l?l?l?l?l?l?d?d|8 letters + 2 digits|
|?l?l?l?l?l?l?d?d?l?l?l?l|Letters → digits → letters|
|?d?d?d?d?d?d?d?d?l?l?l?l|Digits followed by letters|
### **Observations**
- Digits overwhelmingly appear at **edges**, not in the middle
- Lowercase dominates unless forced otherwise
- Longer passwords often remain **structurally simple**, not complex
---
## **Asia / Digit-Heavy Mask Plans**
These masks reflect datasets where numeric passwords are significantly more common, often due to:
- Mobile-first authentication habits
- Numeric keypad familiarity
- Different cultural memorability patterns
|**Mask**|**Info**|
|---|---|
|?d?d?d?d?d?d|6-Digits|
|?d?d?d?d?d?d?d|7-Digits|
|?d?d?d?d?d?d?d?d|8-Digits|
|?d?d?d?d?d?d?d?d?d|9-Digits|
|?d?d?d?d?d?d?d?d?d?d|10-Digits|
|?d?d?d?d?d?d?d?d?d?d?d|11-Digits|
|?d?d?d?d?d?d?d?d?d?d?d?d|12-Digits|
|?l?l?l?l?l?l|6-Lowercase|
|?l?l?l?l?l?l?l|7-Lowercase|
|?l?l?l?l?l?l?l?l|8-Lowercase|
|?l?l?l?l?l?l?l?l?l|9-Lowercase|
|?l?l?l?l?l?l?l?l?l?l|10-Lowercase|
|?l?l?d?d?d?d?d?d|2-Lowercase + 6-Digits|
|?l?l?l?d?d?d?d?d?d|3-Lowercase + 6-Digits|
---
### **Context & Takeaways**
- These masks heavily favor **numeric-only passwords**, reflecting:
- Mobile-first input patterns
- PIN reuse
- Numeric identifiers doubling as passwords
- Mixed alpha–numeric masks are present but far less dominant than in Western datasets.
- Lowercase-only passwords remain common where Latin input is used, but digits still dominate.
These masks are **statistically derived**, not speculative.
They consistently recover large volumes of weak credentials when applied early in analysis pipelines.
If you want, next we can:
- Add **policy-aligned variants** (min/max length, required digits)
- Layer **increment strategies** (--increment)
- Convert these into a **.hcmask plan** ordered by expected yield
### **Observations**
- Pure numeric passwords remain extremely common
- Mixed masks often place letters **before** digits
- Length increases do not imply entropy increases
---
## **Why Simple Masks Still Matter**
Many of these masks look trivial.
That is precisely why they work.
- Users optimize for **minimum effort**
- Policy compliance does not equal randomness
- Length ≠ strength when structure is predictable
---
## **When to Use These Masks**
These masks are most effective when:
- Targeting **weak or legacy password policies**
- Performing **baseline password hygiene analysis**
- Establishing a **floor of expected compromise**
- Studying **human password construction behavior**
- Building **policy-aware variants** (e.g., forcing uppercase or symbols)
They also serve as an excellent foundation for:
- Hybrid mask + dictionary attacks
- Rule amplification
- Comparative analysis across regions or datasets
---
## **Key Takeaways**
- Masks represent **structure**, not guesses
- Large-scale data confirms that **simple structures dominate**
- Low-hanging fruit is always present, even at massive scale
- Mask attacks are one of the **purest measurements** of human password behavior
---
#sudad
[[Masks]]
[[Home]]