Our methodology

Data sources, calculation methods, editorial policy, and how we ensure accuracy. Most figures come straight from official Australian and international government feeds; some are FuelCrisis-derived from a documented formula; a few are editorial assessments where no live feed exists. Every figure is labelled with which it is, so you can verify our work.

Data labels

Four provenance badges used on every figure

Official statistic

Every figure on this site carries one of four provenance badges. This is our commitment to you: you will always know where a number comes from and whether it is an official government figure or our own calculation.

Official statistic
Published directly by an Australian or international government body
Government API
Fetched from an official government or commercial data feed
Calculated from official data
Composite metric calculated from multiple verified government and industry sources
See Supply health index and Data sources below for the approach.
Editorial assessment
Our team's assessment based on current reporting and expert analysis

Where our data comes from

Named sources, update cadence, and the live-pipeline principle

Official statistic

Our data-sourcing principle

Every number on this site traces back to a live cron-to-Neon pipeline. We do not display static, sample, or placeholder data anywhere. If an upstream feed is temporarily unreachable, the relevant display renders a data unavailable panel rather than a stub value, and the panel explains which feed is missing and when the next automated refresh is due. Derived figures (like Singapore Mogas benchmarks calculated from live wholesale and live AUD/USD) are acceptable only when every upstream leg is live-pipelined, and they are always marked with the Calculated from official data badge so you can see the derivation.

Fuel prices (all 8 states and territories)

State and territory retail fuel prices come from official government mandatory reporting schemes. Seven jurisdictions run mandatory reporting schemes where service stations must report pump prices by law. Tasmanian prices come from FuelCheck TAS, which runs on the NSW FuelCheck platform. SA data comes from the SA Fuel Pricing Information mandatory reporting scheme (the same platform as QLD). NT retail data comes from MyFuel NT, the official Northern Territory Government consumer fuel price tracker, covering about 210 outlets (as at mid-2026) across all 9 NT regions.

Update frequency: Daily, seven days a week. VIC carries a mandatory 24-hour delay imposed by the Victorian Government.

Primary sources: NSW/ACT/TAS FuelCheck (CC BY-SA), Service Victoria (VIC fuel prices) (CC BY 4.0), WA FuelWatch (CC BY 4.0), QLD Fuel Prices (CC BY 4.0), SA Fuel Pricing Information, MyFuel NT

Fuel reserves (DCCEEW MSO Dashboard)

National fuel reserve levels (petrol, diesel, jet fuel) come from the Australian Government Department of Climate Change, Energy, the Environment and Water (DCCEEW) Minimum Stockholding Obligation Dashboard. The dashboard publishes weekly data including days of cover, stock volumes, surplus percentages, and MSO baselines for each fuel type. All reserve figures shown on the site are official government data.

Update frequency: Weekly on Saturdays from DCCEEW.

Primary sources: DCCEEW MSO statistics

Crude oil and exchange rates

Brent crude oil prices come from the US Energy Information Administration (EIA) series EPCBRENT (Europe Brent Crude Oil Spot Price, FOB), with a FRED fallback (series DCOILBRENTEU) for continuity. This is the spot/physical Brent assessment, a public proxy for the Platts Dated Brent benchmark that prices real North Sea cargoes (BFOE). It is not the front-month ICE Brent futures contract (paper oil traded on exchange), which can diverge from physical cargoes during supply disruptions. We use spot because it reflects what refiners actually pay to lift crude, which is the relevant signal for downstream Australian fuel pricing. AUD/USD exchange rates come from the European Central Bank via the Frankfurter API. Both are official, publicly funded data sources used by governments and financial institutions worldwide. Charts on /oil-prices render a rolling 90-day window when the market-data cron has populated the history table.

Update frequency: Collected daily. Exchange rates publish on business days; Brent crude is a daily series released weekly.

Primary sources: EIA Brent spot, FRED Brent (DCOILBRENTEU), Frankfurter (ECB reference rates)

Singapore benchmarks (Mogas 92, 95, Gasoil)

Singapore refined-product benchmarks are Calculated from official data figures rather than direct feeds. Derivation chain: the live AIP terminal gate price for ULP (c/L AUD; the national figure is a population-weighted mean of the seven capitals AIP publishes daily, which reproduces the national average AIP itself used to publish to within 0.1 c/L) is first stripped of Australian taxes to recover the pre-tax wholesale equivalent (AIP TGPs are GST-inclusive and excise-inclusive per AIP, ACCC, and ACAPMA/OPIS documentation: we divide by 1.10 to remove GST, then subtract the current effective excise, which steps through the two-phase relief schedule: 20.6c/L from 1 April to 30 June 2026 (the deep cut, 61.6 per cent off the full rate), 36.6c/L from 1 July to 2 August 2026 (the extended, reduced cut announced 20 June 2026, 31.8 per cent off), and 53.7c/L when the full rate resumes from 3 August 2026; the current rate, its ATO derivation and every rate since the relief began). The pre-tax wholesale equivalent minus the effective Mogas spread (the editorial baseline constant mogasSpreadCentsPerLitre multiplied by mogasCrisisMultiplier) yields the Mogas 95 equivalent in c/L; converted to USD per barrel via live ECB AUD/USD. The crisis multiplier captures Hormuz-era freight and insurance widening. It was stepped down to 1.5 on 3 July 2026 during the mid-June reopening, then back up to 2.0 on 17 July 2026 (review range 1.8 to 2.2 while the strait is closed) after the strait was declared closed on 12 July, producing an effective Mogas spread of approximately 20c/L AUD, with Gulf freight and war-risk insurance now several times pre-crisis and hull cover back above the March peak. The multiplier steps back to 1.0 once the crisis resolves; until then, it carries a weekly editorial review cycle while the closure persists. Mogas 92 is Mogas 95 minus a quality (octane) discount of US$2.50 per barrel (constant mogas92QualityDiscountUsdPerBbl, six-month review cycle, unchanged by the crisis). Gasoil 10ppm anchors on Mogas 95 USD and scales the historical Mogas-to-Gasoil differential by gasoilCrisisMultiplier (same 1.3 to 1.7 review range, currently 1.5), capturing diesel demand inelasticity stretching the crack spread during supply stress. A widening Mogas-to-Brent crack spread signals that refining capacity, not crude supply, is the binding constraint on pump prices. Any upstream leg missing (wholesale or AUD/USD) causes the relevant card to render data unavailable rather than a stub value.

Update frequency: Derived daily from live wholesale and FX feeds.

Primary sources: AIP terminal gate prices

Supply chain and geopolitical context

Supply chain vulnerability indicators draw on data from the IEA, DCCEEW, ABS trade statistics, and published industry reports. Chokepoint status (Hormuz, Malacca, Red Sea, Indonesian straits) is our own Editorial assessment assessment reviewed weekly and clearly labelled as such. Australia's refinery capacity and import dependency figures come from DCCEEW and company annual reports. The full Strait of Hormuz explainer is in section 6 below.

Update frequency: Varies by source. Editorial assessments are reviewed weekly.

Primary sources: National Fuel Security Plan, EIA chokepoint analysis, IEA on the Strait of Hormuz

A note on what we publish here: we name our sources so you can verify our data independently, and where a source publishes its data under an open licence we name that licence and link it. Two do not appear with one: South Australian data comes through an authorised data-publisher scheme rather than a public open-data licence, and the Northern Territory publishes no licence terms for MyFuel NT. We would rather show nothing there than guess. We do not publish API endpoints, authentication details, or technical integration specifics. If you are a journalist, researcher, or government official and need to verify a specific figure, please contact us at info@fuelcrisis.com.au.

Data freshness and known delays

What updates daily, what is weekly, and where government-imposed delays apply

Official statistic

Daily (government databases)

NSW, ACT, QLD, SA, TAS and WA fuel prices are published every day, weekends included, by state government reporting schemes. We collect them twice daily, at 06:30 and 10:30 AEST. Exchange rates follow the European Central Bank, which publishes on business days only. Brent crude is a daily series that the EIA releases weekly, so the newest observation can be several days old through the middle of a cycle.

24-hour delay (VIC prices)

Victorian fuel prices carry a mandatory 24-hour delay imposed by the Victorian Government on all third-party access to the Servo Saver / Fair Fuel public API. Retailers report prices to the Victorian Government in real time, but the government withholds data from third-party consumers (including FuelCrisis) for 24 hours for commercial sensitivity reasons. This delay is applied at the source by the Victorian Government, not by FuelCrisis. As a result, VIC prices shown on this site are always one day behind. We note this delay where VIC data appears so you always know the data is not current-day.Dating change, July 2026: we now date the Victorian figures for the day those prices were actually charged, rather than the day we collected them. Victoria therefore shows one day earlier than the other states on the state table, and each state carries its own as-at date so you can see which day you are reading. Two consequences are worth stating plainly. First, the national daily average now ends a day earlier than it used to: Victoria is about a quarter of the country by population, and a day without it does not meet the coverage floor we require before publishing a national figure, so today's point appears tomorrow. Second, figures published before this change carried the collection date, so a Victorian series read straight across July 2026 has a one-day step in it. We have not rewritten the older rows, because they are a faithful record of what we published at the time.For real-time VIC prices, use the official Servo Saver app provided by Service Victoria.

How national averages are calculated (population-weighted)

The primary national average shown on this site is population-weighted using Australian Bureau of Statistics Estimated Resident Population (ERP) figures by state and territory (December 2024 quarter, Catalogue 3101.0). Each state's retail figure is weighted by its share of the national population, so NSW (32%) and VIC (26%) contribute proportionally more than TAS (2%) or the NT (1%). States with no available price data are excluded and the remaining weights renormalise automatically. The population weights are reviewed annually when ABS publishes the next ERP release (typically December each year); inter-censal drift is well under one percentage point per jurisdiction per year, so a 12-month refresh cycle is sufficient. Our figures are median station prices from the mandatory government feeds, blended across metro and regional areas by ABS population. This applies site-wide: the per-state table figures, the capital-versus-regional figures, the national average and its week-ago and month-ago comparisons all come from the same median-based series. We use the median because our feeds include every station, including remote outlets charging far above city prices, and a simple average would overstate what a typical driver pays. Official averages (WA FuelWatch, AIP, ACCC) use arithmetic means over different station panels, so our figures can sit 1 to 2 c/L away from theirs. That difference is methodological, not an error. Methodology change, 14 July 2026: we unified every figure on the site onto this single median-based series. Before that date, the per-state table figures were simple means of all station prices in the state, and the week-ago and month-ago movement figures were an unweighted average of those state means, while the headline national average was already a population-weighted median blend. Because a mean and a median can sit 1 to 2 c/L apart on our feeds, movement figures shown before this change, and comparisons whose look-back window still reaches into the older series (up to 7 days for week-ago, up to 30 days for month-ago figures after the change), could carry a methodological offset of roughly 1 to 2 c/L that was not real price movement. One comparison never grows out of this: the pre-crisis baseline (the AIP national average for the week ending 22 February 2026) is a fixed mean-based figure, so comparisons between today's median-based prices and that baseline are permanently cross-methodology, and a gap of up to 1 to 2 c/L against it can be methodological rather than a real price difference. This was a methodology improvement, not a factual error, so it is recorded here rather than on the corrections page. Because VIC data carries a mandatory 24-hour government-imposed delay, we date Victorian figures for the day the prices were charged rather than the day we collected them (see the VIC delay note above). Every day in the national daily series is therefore a like-for-like blend: all eight jurisdictions on the same trading day, none of them carrying yesterday's prices under today's label. The price of that is a day of currency, not accuracy. Victoria's figures for a given day do not reach us until the following morning, and a day missing Victoria falls below the population-coverage floor we require before publishing a national figure, so the daily series ends one day back rather than at today. We took that trade deliberately. Under population weighting VIC's roughly 26 per cent share meant that publishing a mixed day put about a quarter of the day's price movement into the national figure as pure lag, unlabelled: a 10 c/L national swing would have carried roughly 2.5 c/L of it. In practice the error has been small (Victorian day-to-day moves have run about 1.5 c/L through this crisis, so about 0.4 c/L on the national number), but it grows silently whenever volatility returns, and a figure that is a day old and says so is worth more than one that is current in name only. Excluding Victoria entirely from a "national" figure would be the worse answer again.

Which fuels we publish, and which get a national figure

The state price tables carry seven fuels: ULP 91, diesel, Premium 95, Premium 98, E10, LPG (autogas) and premium diesel. Not all seven get a national figure, and the difference is a coverage rule rather than an editorial preference.

A fuel earns a national average only when enough states and territories report it for a population-weighted figure to mean anything. ULP 91 and diesel clear that comfortably, and so do Premium 95, Premium 98 and LPG. E10 and premium diesel are published per state only. For E10 that is deliberate: it is not sold everywhere, and a national number would average across markets where the fuel barely exists. Premium diesel would now clear the floor on coverage alone, but promoting a fuel to a national headline is an editorial decision we have not yet taken, so it stays per state and is labelled that way.

E85 is no longer published at all. It was withdrawn from the price tables on 25 July 2026 for a reason worth stating: it has all but disappeared from Australian forecourts, a state median rested on a handful of sites, and in one state the entire month of data sat at a single unchanging price that could not be a real market reading. Collection continues so the series keeps accruing, but we would rather show nothing than a figure that looks like a market price and is not.

Retail prices (NT)

NT fuel prices are sourced from MyFuel NT, the official Northern Territory Government consumer fuel price tracker operated by NT Consumer Affairs. Updated daily, covering about 210 outlets (as at mid-2026) across all 9 NT regions (Barkly, Central Australia, Darwin, East Arnhem, Katherine, Litchfield, Palmerston, Tiwi Island, Top End Rural). This completes 8/8 state and territory retail coverage.

SA data attribution

South Australian fuel price data is based on or contains data provided by the State of South Australia (Office of Consumer and Business Services). SA prices are sourced from the SA Fuel Pricing Information mandatory reporting scheme, where retailers must report price changes within 30 minutes.

Fuel reserves (weekly MSO, monthly APS)

Reserve figures come from two official DCCEEW sources: the Minimum Stockholding Obligation dashboard, published weekly on Saturdays, and the monthly Australian Petroleum Statistics. The days-of-cover figure is labelled with whichever of the two reported most recently, and with that source's publication cadence (weekly or monthly). Every figure shown is official government data; where an official cover figure is unavailable for a fuel, that slot reads "data unavailable" rather than an estimate.

What happens when a feed fails

Per our data-sourcing principle above, any display element whose upstream leg is missing renders a data unavailable panel, never a stub value. The panel names the missing feed and when the next automated refresh is due. Cron routes publish run health to /admin/health for operator visibility.

How we call the price direction

The wholesale window, the threshold, the freshness gate, and what is deliberately not an input

Calculated from official data

The homepage verdict (rising, easing or steady) is a Calculated from official data reading of the wholesale market, because wholesale petrol leads pump prices by about one to two weeks. The verdict compares the mean of the latest 5 trading-day AIP terminal gate prices for petrol with the mean of the 5 before them. A move of more than 2 per cent between the two means is rising or falling; anything smaller is steady. It is called only when the newest reading is no more than 4 days old and the 10 readings span two consecutive trading weeks; otherwise no direction is given. Brent crude is not an input.

The input.AIP publishes a terminal gate price for petrol for each of seven capitals every trading day. The national figure the verdict uses is a population-weighted mean of those seven, using the same ABS Estimated Resident Population weights as the site's retail averages; on every date where AIP's own published national average is on file (30 July to 14 August 2026) that mean reproduces it to within 0.1 c/L, so the derived series is continuous with the published one. Before comparing the two windows the series is excise-normalised: the 1 July and 3 August 2026 tax steps are subtracted out so a tax change cannot read as a market move.

The same rule, per capital. Each state page carries a wholesale line for its capital: the latest terminal gate price AIP publishes for that city, its move on the week, and a direction from the same rule applied to that city's own series rather than to the national figure. The windows, the threshold and the freshness gate are identical, so a capital's direction is withheld on the same days the national one is, and a capital can read differently from the national figure only because its own prices moved differently. Canberra has no terminal gate price and carries no line. The seven capital readings, with their weekly moves, are also in the weekly numbers table on the journalists' page.

The gates, in numbers. Two windows of 5 trading days each; a direction needs a move of more than 2 per cent between their means; the newest reading must be no more than 4 days old; and the 10 readings must sit inside two consecutive trading weeks. If any gate fails the homepage says so and gives no direction. A withheld verdict is the design working, not a fault.

What is not an input.Brent crude. Until 11 September 2026 the verdict blended Brent with wholesale and fell back to Brent alone when the wholesale feed was stale. That produced a published "wholesale prices are easing" on 10 September from a Brent window that ended on 1 September, while terminal gate petrol had risen about 9 c/L in five trading days and Brent itself was up 18 per cent on the month. Crude is a different product on a different release cycle (the series the site holds is published weekly) and is not a substitute for a wholesale petrol price on a two-week horizon, so it was removed rather than re-weighted.

Series note: no wholesale readings exist for 15 August to 3 September 2026. AIP withdrew the weekly spreadsheet the site read and moved to a daily on-page table; the fetcher failed silently until 11 September and no archive of the page exists for the gap. The verdict was withheld until two consecutive trading weeks of the new series were on file.

How we calculate days remaining

DCCEEW weekly figure, MSO baselines, and what the status colours mean

Official statistic

The reserve bars on the dashboard show how many days of fuel cover Australia holds. Here is how those numbers work.

Official figure

Taken directly from DCCEEW. The Minimum Stockholding Obligation dashboard publishes weekly on Saturdays: fuel entities (importers, refiners) report stock held as of Tuesday, submitted by Friday, and the department publishes the aggregated national figures on Saturday. The monthly Australian Petroleum Statistics is the other official source. The days-of-cover figure carries whichever of the two reported most recently, labelled Official statistic with its cadence.

What the dashboard shows

All reserve figures shown are official DCCEEW data, from the weekly MSO dashboard or the monthly Australian Petroleum Statistics. We display days of cover, stock volumes, the margin to MSO baselines, and period-on-period trends. Where an official cover figure is unavailable for a fuel, that slot reads as unavailable; we do not substitute an interpolated or estimated figure.

Key data points

Days of cover: the number of days Australia's current fuel stocks would last at the current rate of consumption, as reported by fuel entities to the DCCEEW.
MSO baseline: the Minimum Stockholding Obligation is the legally mandated minimum number of days each regulated fuel entity must hold of its own throughput. It is a per-entity obligation, not a national days-of-cover floor, so the reserve bars use it as a reference line rather than a statutory national target. The obligation is currently under temporary relief to 30 September 2026.
Week-on-week trend: compares the latest weekly days of cover to the prior week. A change of less than 2% is treated as stable.
IEA 90-day target: the International Energy Agency recommends member countries hold at least 90 days of net import cover. Australia has consistently fallen short of this benchmark.

Data currency note

Reserve figures are updated weekly when the DCCEEW publishes new MSO data. Between publications, the most recent official figure is displayed. Actual stock levels may change between reporting periods due to consumption, imports, and refinery output.

Editorial policy

Fact versus assessment, corrections, accessibility, and not financial advice

Editorial assessment

How we handle the line between data and opinion, corrections, and accessibility.

Fact versus assessment

Official government statistics are presented as-is, without editorial interpretation. Where we add context, analysis, or our own calculations, we use the Editorial assessment or Calculated from official data badges so you always know what is a government figure and what is our assessment. The colour of the badge is consistent across every page on the site.

Corrections

If we get something wrong, we correct it and note the correction in the CHANGELOG. All versions of the site are archived in Git. We do not silently edit published content. Structural route changes (see section 7 above) are also logged.

Accessibility and WCAG

We target WCAG 2.1 AA compliance. Recent passes in April 2026 hardened the site in four specific ways: (1) status badge text on tinted same-hue backgrounds was migrated from saturated colour tokens (e.g. text-fc-red) to lighter text variants (e.g. text-fc-red-text) for AAA text contrast (WCAG 1.4.3); (2) freshness and data-provenance rows were upgraded from role="note" to role="status" aria-live="polite" so screen readers announce feed health changes; (3) SVG charts and visualisations now wrap in role="group" with coherent aria-label (WCAG 1.3.1 and 4.1.2); (4) small text links and breadcrumbs are wrapped to meet the WCAG 2.5.8 AA 24x24 CSS pixel pointer target floor. Keyboard focus rings use :focus-visible so mouse clicks do not leave visual rings behind.

Not financial advice

Nothing on this site constitutes financial, investment, or emergency management advice. We present data and analysis to help Australians understand the fuel situation. Decisions about purchasing, stockpiling, or financial planning should be made in consultation with appropriate professionals.

Citing and reusing our data

Attribution, licence, stable download URLs

Editorial assessment

Our published datasets carry an open licence, an attribution line and permanent download URLs, and our charts may be republished with attribution. The data and citation page has the licence terms, the citation format, the CSV and chart URLs, and a contact route for verification questions.

Questions about our methodology?

Stay updated on how we build this

We will notify you when we add new government data sources, refine our calculations, or issue corrections. Full transparency, no surprises. For specific verification requests, email info@fuelcrisis.com.au.

Email sign-up is not open yet. Until it is, the RSS feed carries every update.

Our methodology