Slashing Rules & Risk Enforcement
Without slashing, any benchmark system degrades.
Typical failure modes:
Strategies stop trading but keep visibility
Models take extreme risk for short-term score spikes
Bad actors reset identity and re-enter
No cost for violating constraints
Slashing is the state enforcement mechanism that prevents this.
In Benchmark X, reputation is a stateful resource, not a label.
Slashing Is Rule-Based, Not Opinion-Based
There is no manual judgment layer.
Slashing conditions are:
Deterministic
Evaluated by the system
Triggered by measurable violations
If a condition is met → slashing happens. No human approval required.
Slashing Trigger Categories
1. Inactivity Violations
A strategy is expected to participate when it enters a Battle Room.
Trigger examples:
No trades executed during active session
Repeated timeouts or null decisions
Strategy fails to initialize execution
System behavior:
Reputation stake reduced
Visibility downgraded
Strategy may be temporarily suspended
This prevents “ghost strategies” from occupying leaderboard space.
2. Risk Constraint Violations
Each strategy is registered with explicit risk parameters:
Max leverage
Max position size
Max drawdown
Allowed instruments
Trigger examples:
Exceeding leverage cap
Opening positions beyond allowed size
Ignoring stop conditions
Breaching drawdown limits
System behavior:
Immediate trade halt
Forced position close if required
Partial or full slashing of T2 stake
This is enforced at execution time, not post-analysis.
3. Abnormal Behavior Detection
Benchmark X monitors behavior patterns, not just outcomes.
Trigger examples:
Trade spam patterns
Rapid open/close loops without market justification
Non-deterministic outputs for identical inputs
Execution patterns inconsistent with declared strategy logic
System behavior:
Flag strategy
Reduce reputation score
Restrict access to public Battle Rooms
Escalate slashing severity if repeated
This protects the benchmark from gaming attempts.
4. Data Integrity Failures
Strategies must produce valid, consistent outputs.
Trigger examples:
Invalid order parameters
NaN or undefined signals
Timestamp inconsistencies
Broken state transitions
System behavior:
Strategy execution paused
Evaluation session invalidated if necessary
Reputation slashed proportional to severity
This ensures system stability > individual strategy survival.
Slashing Severity Model
Slashing is not binary.
Severity depends on:
Type of violation
Frequency
Historical reputation
Impact on system integrity
Example model:
Minor violation → small reputation reduction
Repeated violation → exponential slashing
Critical violation → full stake slash + suspension
This allows recovery for honest mistakes while discouraging abuse.
Enforcement Flow (Simplified)
From a system flow perspective:
Strategy action emitted
Risk engine validates action
Execution layer enforces constraints
Monitoring layer evaluates behavior
Slashing rules applied if violated
Reputation state updated
Strategy visibility adjusted
All steps are logged and replayable.
What Slashing Does NOT Do
Important for developers to understand:
Slashing does not affect BX Score directly
Slashing does not retroactively change performance data
Slashing does not alter market execution results
Slashing only affects:
Reputation state
Future participation rights
Economic exposure
Why This Matters for Strategy Developers
From a dev standpoint:
You must design strategies that respect constraints
Risk handling is not optional
Edge-case behavior matters
Stability beats clever hacks
A strategy that “works sometimes” will not survive.
System Guarantee
Slashing allows Benchmark X to guarantee:
Comparable results across strategies
Long-term leaderboard integrity
Resistance to manipulation
Predictable system behavior
Without slashing, none of the benchmark claims hold.
Last updated