TL;DR: We measured every Brisbane ULP peak from June 2024 to July 2026 — 13 peaks, 14 troughs. The modern cycle (six laps, peaks from May 2025 on) runs a median 38 days peak to peak, mean 39.2. It hasn’t got longer or more random: the early laps ran long and scattered — 77, 69 and 89-day cycles; since May 2025 they’ve locked into a tight band, mean 39.2 with a stdev of about 4.6 days. That’s close to clockwork. Today (27 Jul) Brisbane ULP sits at 189.9 c/L, still climbing 28 days off the 29-June trough — mid-lap, not peaked.
There’s a line you hear from every driver who’s given up on timing their fills, and a bloke on Whirlpool put it flat:
“There is simply no real cycle anymore, and it’s guess-work and gamble.” — ‘Quiet Rider’, Whirlpool
That’s the claim worth testing. Not “does the cycle exist” — everyone can feel it — but whether it’s still regular enough to plan around, or whether it’s drifted into noise the way people say it has.
What we checked
Every price here is the Queensland Government feed — the actual price each station reported charging, not a driver’s guess at the bowser.
We pulled every ULP (regular 91) price-change report from the Brisbane-metro stations inside 30 km of the CBD, going back to June 2024. On any given day 424 to 440 of the 473 in-range stations are active, so we carried each station’s last known price forward and took the median across the panel, one number per day, for 25 months. Then a turning-point pass walked the series and marked every peak and trough. That found 13 peaks and 14 troughs. We ran the detector at three sensitivities (15, 20 and 25 c/L reversal thresholds) and every one returned the same 13 peaks — so the count isn’t an artefact of where we drew the line.
Then we measured the gap, in days, between one peak and the next. Thirteen peaks give 12 peak-to-peak intervals.
One honest exclusion up front. There’s a known data gap from early March to late April this year where reports went quiet, so the carry-forward invents a fake ramp to 259.9 c/L. The one interval that crosses that stretch — the Jan-to-March gap — is dropped from the maths, because it’s a data hole, not a 90-cent cycle. That leaves 11 clean intervals to measure.
What the feed shows
The cycle hasn’t got longer. It’s got shorter, and far more regular.
Split the clean intervals by era and the trend is plain. (The spread figures here come off the raw daily series, not the rounded bar labels in the chart below — recompute from the bars and you’ll land a touch higher, which is the rounding, not a disagreement.) Through the first five laps (mid-2024 into early 2025), cycles averaged 68 days and scattered badly — a standard deviation of 15 days, so one lap ran twice the length of another. Two of them genuinely dragged: 77 days across mid-2024, 89 across early 2025. Those are real single long cycles, not skipped peaks — we checked inside them and the price just plateaued and drifted rather than turning. They drag the full 25-month mean up to 52 days.
But from May 2025 on, the laps snap into line. The modern regime runs a median 38 days peak to peak (mean 39.2 across six laps), with a standard deviation of about 4.6 days. A regression across the eleven clean intervals slopes down at about 4.2 days per cycle — steadily tightening, not loosening.
Here are the clean intervals in order, so you can watch them converge:
The scatter is the whole story. A coefficient of variation of 0.12 in the modern regime isn’t noise — it’s close to a metronome. Whatever “no real cycle anymore” feels like from the driver’s seat, the feed says the opposite: the cycle is more predictable now than it was two years ago, not less. (If you want the mechanics of why the cycle exists at all — refiners raising in lockstep, then undercutting each other down — we pulled that apart in how the Brisbane price cycle works. That piece called the lap “about 37 days” off the five most recent cycles; over the full six-lap modern run measured here the median lands at 38. Same cycle, one day apart — which is the point: it’s that steady.)
For scale, the swings themselves are steady too. We get one more clean swing than we do clean lap-length: measuring peak-to-trough amplitude keeps a swing that measuring peak-to-peak length has to drop, because that lap’s far peak falls on the wrong side of the March–April gap — you can read its depth but not its full length. So amplitude runs on 12 swings where length ran on 11. The median peak-to-trough drop across those 12 clean swings is 46.5 c/L (mean 44.8, range 27 to 59) — which sits right on the canonical “about 45 c/L” every amplitude read lands on. The median runs a touch above the mean here, which reads backwards until you look at the two shallow swings — a 27 and a 30-cent lap in late 2024 and early 2025 — that pull the average down below the middle value. So the depth of a lap and the length of a lap are both stable. That’s the definition of a cycle you can plan around.
Where we are now
The last trough was 29 June at 151.9 c/L. Since then the median has climbed +38 c/L to 189.9 today (27 Jul), day 28 of this lap. (Our diesel piece the same day clocks the petrol climb at +36, measured off the 30-June panel value of 153.9 rather than the 29-June trough floor — same 189.9 top, a start point two cents and one day apart. A cycle-length read wants the true trough, so we anchor on the 29th.) That’s the top of the range so far — but “so far” is doing work, because this lap is still on its way up. It hasn’t peaked. On the ~38-day cadence, a typical peak lands about 46 c/L above the trough, which would put the top near 198 c/L. We’re mid-climb, not at the crest.
One thing to keep separate. The level this July is high partly for reasons that have nothing to do with the cycle. The fuel-excise relief stepped down on 1 July to a 16 c/L cut — a clean step up in wholesale that day (Tom clocked it at +15.2 c/L in the diesel TGP) — and that remaining 16 c/L cut fully reverts on 1 August, with crude climbing on top of both. That’s a floor-raiser sitting under every station — Tom’s beat, not this one, and he lays out the excise arithmetic in why Brisbane diesel is up 52 cents. The cycle shape — 38 days, regular — holds regardless of where the floor is.
The verdict
The “no cycle anymore” claim is wrong for Brisbane, and the feed says so cold. Thirteen peaks in 25 months. The modern lap is 38 days, give or take four or five. It has tightened, not frayed. If anything the cycle is easier to time now than when Quiet Rider gave up on it.
So what
If you can wait, day 28 of an ascending lap is not your week — you’re near the top of the climb with room still to run. On a 38-day cadence the next honest fill window is the glide-down after this peak turns, and that’s still a couple of weeks out. You don’t have to guess the day. You can watch the metro median and your own regular servos on the map at fill.pinly.com.au and decide from the number, not the folklore.
The honest limit
This is a median across roughly 435 stations — a market level, not a promise about your corner, which can lead or lag by days. The 38-day figure rests on the six modern laps; a wider clean run would firm it up, and the March–April data gap cost us one interval. This is ULP/91, Brisbane metro only — diesel runs its own flatter, less cyclical pattern and none of these numbers carry to it, which is exactly why it’s worth its own read: if you run a diesel, Tom’s breakdown linked above is the one for you. And today’s “top of range so far” is an informational read on an unfinished lap, not a call: the peak hasn’t printed yet.