How to use:
- Choose a mode, a difficulty level and how many rounds you want
- Each round shows one set of records that all follow the same baseline
- The baseline is printed above the records — read it first, then scan
- Click the single record that breaks the baseline
- If every record is consistent, click All Clear instead
- Rounds are dealt out evenly across the four tasks and, on Mixed, across the three difficulty levels
Test and Training:
- Training: the true anomaly is revealed after every answer with a short explanation, and you continue when you are ready
- Test: nothing is revealed during the session; answers are logged and every incorrect round is explained in the results
The four detection tasks:
- Access Logs: entries from a secure site, in three forms. Rule rounds hide a badge in the wrong format, a door that does not exist, a scan outside shift hours, a timestamp out of order, or a badge that enters twice without leaving. Pattern of Life rounds show each badge's usual door and usual hours, then hide one scan that is perfectly legal but does not match that badge's routine. Roster Cross-Check rounds pair an access roster with the entry log, and one entry disagrees with the roster on badge, name or cleared door.
- Surveillance Grid: a grid of identical markers. One differs in count, angle, shade or code.
- Signal Baseline: hourly traffic counts sitting inside a normal range. One hour sits far outside it — a spike or a dropout.
- Identity Documents: ID cards that must be internally consistent — an expiry before the issue date, a country code that does not match the stated nationality, a validity period that is not ten years, an ID number of the wrong length, or a holder under sixteen at issue. Cross-check rounds add a central file: every card is internally valid, and the anomaly is the one that disagrees with the file on nationality, ID number or date of birth.
How scoring works:
- Anomalies Found: rounds where you clicked the correct broken record
- Missed: an anomaly was present and you called All Clear or ran out of time
- Timed Out: rounds the clock took before you called them; anomaly rounds lost this way also count under Missed
- False Alarms: you flagged a clean record — either on a clean round or by picking the wrong record
- Accuracy: correct rounds out of rounds completed
- Avg Decision: average time taken per round, with paused time excluded
- Detection Profile: your accuracy broken down by task, so you can see which kind of anomaly you miss
- Confidence Calibration: switched on from Advanced Settings, some rounds ask how sure you are after you call it. Results then show whether your certainty tracks your accuracy — being confidently wrong is its own problem
Controls:
- Click a record to flag it, or click All Clear
- Press C to call All Clear
- Press P to pause and resume — the round clock freezes while paused
- Press Escape to end the session early; an ended session is not saved
- On calibration rounds, press 1, 2 or 3 to log low, medium or high confidence
- Switching browser tabs pauses the session automatically
Training tips:
- Read the baseline before looking at the records — scanning without a rule in mind is how anomalies get missed
- Check one column at a time rather than reading row by row
- Resist flagging something just because a round feels like it must contain an anomaly; clean rounds exist to catch that habit
- On Station Chief, some records are deliberately odd but permitted — back-to-back scans, a badge on a boundary minute, an ID full of zeros. Unusual is not the same as inconsistent
- Learn each task in Training mode, then measure yourself in Test mode with a 20 second limit
- Use the Detection Profile to pick your weakest task, then run a single-task session on it from Advanced Settings
- Ending a test early does not save a result, so finish the session to track progress