1
State machine
Represent flat, candidate, active, exit-pending, and cooldown states explicitly.
Lesson · Algorithmic and Automated Trading
Signal Engine and Strategy State
Convert normalized events into deterministic strategy-state transitions while preventing repeated, contradictory, or stale signals.
Core concepts
What to understand
Keep the boundaries, state, and authority explicit.
2
Transition signal
Emit a decision when state changes, not on every event where a condition remains true.
3
Determinism
The same ordered inputs and prior state must produce the same next state.
Interactive lab
Signal State Machine Lab
Toggle stateful processing, confirmation, cooldown, and stale-event handling. Observe duplicate and contradictory signals.
Signal State Machine Lab
Deterministic synthetic educational data · no broker or live-market connection
SYSTEM MODEL
Process
Process one event at a time
Every transition should reference the triggering event and previous state.
Evaluate
Calculate conditions from normalized data and current state.
Transition
Apply one valid state transition with a reason code.
Persist
Store the new state before publishing the resulting signal.
Knowledge check
Check your understanding
Select one answer for each question. Feedback appears immediately, and a wrong choice is answered on its own terms before the rule is restated.
1. Why can a true condition produce no new signal?
Signals should usually represent transitions, not repeated observations.
2. A cooldown primarily prevents:
It enforces a defined period before the same setup can trigger again.
3. Strategy state should be persisted before publishing because:
A restart after publication but before persistence can repeat the decision.
4. The entry condition is true on 12 consecutive events. How many entries does a stateless engine with confirmation 1 emit, and how many does a stateful one?
Stateless: 12 entries, 11 of them duplicates. Stateful: one transition into LONG, then nothing, because a condition that stays true is not a new event.
Practical rules
- Model strategy states explicitly.
- Attach event IDs and reason codes to every transition.
- Reject stale or out-of-order events according to a fixed policy.
Common mistakes
- Emitting an entry on every qualifying bar.
- Keeping state only in process memory.
- Allowing late events to reverse current state silently.
References
Official reading
These sources cover algorithmic controls, market access, broker APIs, system operations, or security.
FINRA · Regulatory Notice 15-09
Official guidance on supervision, testing, controls, documentation, and monitoring for algorithmic trading strategies.
Alpaca · WebSocket Market Data
Official example of receiving real-time stock, option, crypto, and news updates over a streaming API.