Market data and event reconstruction

Match each question to an observation. Use the applicable price series for the financial question, metered or explicitly defined interval-average quantities for the energy calculation, and unit targets, availability, bids, constraints and flows for operational mechanisms. A dispatch target alone does not prove delivered metered energy; an installed capacity number does not prove available supply in the interval.

AEMO's MMS documentation provides a route to current table definitions and market-data reports. Before extracting real records, write down the table, primary keys, field definitions, interval convention and revision treatment. Some participant information has different access conditions. Do not invent missing bids or treat a public dashboard as if it contains every settlement or operational field.

Join tables only after deciding how region, interval, intervention/run identifiers and revisions distinguish records in the selected data model. Reconcile row counts before and after every join. Preserve absent observations as missing until their meaning is investigated; converting a missing price to zero invents a free-energy interval. Topics 31 and 32 provide a synthetic cleaning lab before this real-data extraction task.

• AEMO — MMS data model and access routes

Use the case's separate original synthetic dataset, not June 2022 observations: four five-minute intervals with average customer loads 120, 180, 240 and 180 MW, and prices A$50, 100, 300 and 150/MWh. Each energy quantity is average MW × 5/60 hours: 10, 15, 20 and 15 MWh. Energy-only costs are A$500, 1,500, 6,000 and 2,250, totalling A$10,250 for 60 MWh.

The customer-weighted unit cost is 10,250/60 = approximately A$170.83/MWh. The simple mean price is A$150/MWh; multiplying it by 60 MWh gives A$9,000 and understates cost by A$1,250. Higher consumption coincides with higher prices in this example. That explains the arithmetic difference, not why the market price changed.

Declare the financial boundary. This example excludes contracts, losses, network and retail charges, ancillary services, taxes, prudentials and compensation. It is neither a complete retail bill nor a participant settlement. Before adding any of those components, identify their quantities, reference prices, applicable rules and cash-flow direction. Never multiply a regional price by a unit dispatch target and present it as an audited invoice.

Compare the observed teaching profile with a stated reference: 120 MW and A$50/MWh in every interval. Reference energy is 40 MWh and cost A$2,000. The observed cost increase is A$8,250. This reference is a chosen analytical baseline, not a forecast or a claim about what would really have happened without the event.

For each interval write Q = Q0 + ΔQ and P = P0 + ΔP. Then QP − Q0P0 = Q0ΔP + P0ΔQ + ΔQΔP. Summing the rows gives A$4,000 of price change at reference volume, A$1,000 of volume change at reference price and A$3,250 of interaction. The three terms add exactly to A$8,250. Omitting interaction would leave a large unexplained difference.

This is an accounting decomposition, not a causal market model. Changing price first and then quantity allocates the interaction differently from changing quantity first; disclose the method. To argue that an outage caused a price movement, investigate availability, ramp limits, bids, demand, imports, constraints and interventions, and test a plausible alternative explanation. A constrained rerun requires a validated model and inputs; the cost identity does not replace it.

AEMO’s review covers 10–24 June 2022. Administered pricing began in Queensland on 12 June and NSW, Victoria and South Australia on 13 June. The spot market suspension began on the afternoon of 15 June amid interacting operational and commercial pressures. Its contemporary summary and later incident report should be read together. The dated media release records the formal lifting at 14:00 AEST on 24 June. These are historical facts, not a claim that today's price settings, procedures or compensation arrangements are unchanged.

Use the report to extract three categories: documented system conditions, operator actions, and subsequent findings. Locate the sequence involving fuel pressures, generator availability, administered pricing, directions and suspension. Record the page or notice supporting each link. Distinguish the return of normal pricing during the staged process from the formal end of suspension; do not compress different milestones into one timestamp.

Your historical task is to identify what each source establishes and what it leaves unresolved. Prepare two competing mechanism explanations and the evidence needed to discriminate between them. Do not infer participant intent from an availability change alone. The four-row CSV is a separate teaching exercise and contains no reconstructed prices, volumes, losses or compensation from June 2022.

• AEMO — June 2022 events and reporting context
• AEMO — NEM market suspension and operational challenges, June 2022
• AEMO — suspension lifted, 24 June 2022

Write an event note with six parts: decision required; verified sequence; quantified exposure; competing explanations; action; and review trigger. Label each material statement as observation, assumption, model output or judgement. An appropriate first response can be to obtain a missing constraint or settlement record, not to increase a trading position on the basis of a visual correlation.

Independent transfer task: in the synthetic profile, move 5 MWh from the third interval to the first, keeping total energy fixed and assuming prices do not change. Recalculate the load profile and cost, then explain what operational flexibility would be needed. The arithmetic is not proof that a real customer can shift the load or that prices remain fixed after a large demand response. Identify rebound, service requirements and connection limits before recommending action.

For a real-data follow-up, select one bounded historical interval window, retrieve the applicable public records and preserve the original files. A reviewer should check your timestamp convention, joins, revisions, mechanism evidence and alternative explanation. Submit a reproducible workbook, a short timeline and a one-page memo. Neither a correct cost bridge nor a quiz pass alone establishes independent event-analysis competence.