The Verdict isn't a guess.

If you've been filling up in Queensland for a few years, you've probably noticed it: petrol is cheap one week, expensive the next, then cheap again.

That's not random. Brisbane runs on a price cycle — typically 7 to 10 days from bottom to top and back. Stations drop prices to compete for customers, then gradually lift them back up over the following days.

FuelVerdict is built to compare unleaded prices from the historical files with the cycle in those files. That comparison is not switched on. The model has to clear a 70% back-test on held-out Queensland history before it is allowed to say FILL NOW or HOLD OFF. Until then the verdict card says CALIBRATING. The prices are delayed. They are not live bowser prices.

When the cycle model is allowed to run, the verdict is designed to be one of three words. That model is off. The price page says CALIBRATING.

FILL NOW — prices are at or near a cycle low. Filling up today means you're buying at the bottom, not the top.

YOUR CALL — prices are mid-cycle. Not a great time, not a bad time. If you're low, fill up; if you can wait a few days, it might be worth holding.

HOLD OFF — prices are near a cycle peak. If you can wait, wait. We'll tell you how many days until the next expected low.

We never soften a HOLD OFF to protect anyone's feelings — or anyone's revenue. If the data says hold, we say hold.

The centre of the bullseye is the cycle low — the cheapest point. The outer ring is the cycle peak. While the model is calibrating, the marker stays hollow. That is what the price page shows. A filled marker, once the model is allowed to run, sits near the centre for FILL NOW and in the outer ring for HOLD OFF.

Once you know whether to fill up, FuelVerdict shows you where.

Best nearby — the lowest unleaded price within 5 km of the suburb centre. The page does not read your location.

Lowest within 50 km — the lowest unleaded price within 50 km of the suburb centre. That is not a true city-wide search yet.

Both are ranked on price only. No station has ever paid to appear. No station ever will.

A station 12km away might have petrol that's 8¢/L cheaper. But if the detour costs you more in fuel than you'd save, it's not actually cheaper. FuelVerdict does that maths for you: price difference × your tank size, minus the cost of the extra fuel for the detour. The result is a dollar figure:

“Historical gap of about $4.20 after fuel cost. The bowser price now may differ.”

“Historical gap the other way: about $2.10 more for the extra distance. The bowser price now may differ.”

You get a real answer, not a misleading cents-per-litre comparison.

Cycle prediction isn't perfect. Brisbane prices generally follow the pattern, but refinery disruptions, major holidays, and global oil movements can throw the timing off.

FuelVerdict back-tests the algorithm against historical data before any change is allowed to drive a verdict. The Stage 1 bar is 70% or better on held-out real history — a correct FILL NOW or HOLD OFF at least 7 times out of 10. A clean result on made-up data does not clear that bar. Accuracy will be published once it has been measured. Misses will not be hidden.

When the signal is noisy, we say YOUR CALL instead of pretending we know. The honest uncertain answer is more useful than a confident wrong one.

The prices are the Queensland Government's monthly fuel-price CSVs on data.qld.gov.au, published under Creative Commons Attribution 4.0. Data delayed. A stored price is shown with the time it was recorded. It is a historical record, not a live bowser price. The Fuel Prices Queensland live feed is a separate licence and is not connected. No application has been submitted. The full credit, licence link, and file list are on the data sources page. FuelVerdict does not accept price submissions from stations, fuel brands, or anyone else.