All trial results
Twelve documented trials across road transport, buses, taxis, pharmaceutical logistics, stationary boilers and telecom generators, in Germany, Denmark, Japan and Nigeria. Results range from a median of +1.7% on telecom generators near idle — too small to establish a clear effect — to 22.1% less fuel on one taxi (28.3% better km/L) — and the method behind each one explains most of that spread. Below them sit 8 operator records from our Japanese master dataset: real named fleets, 43 vehicles, but no trial protocol behind them — a weaker class of evidence that we keep separate rather than blend in.
| Trial | Where | Sample | Method | Measure | Result |
|---|---|---|---|---|---|
| Waste Services | Germany · refuse trucks | 5 vehicles · 296 vehicle-days | Own baseline · blinded · utilisation fell 8.7% · 16 weeks | Consumption, l/100 km | −17.7% |
| Municipality of Frederikshavn | Denmark · municipal truck | 1 vehicle · 1,484 km | Mode-corrected model, with Scania | Consumption, l/100 km | ≈5% (modelled) |
| Alfresa (A・L Plus) | Japan · pharma logistics | 2 vehicles · 4 measurements | Year-on-year, same calendar months | Efficiency, km/L | +8.4% … +20.0% |
| Sanden Kotsu | Japan · service buses | 2 buses · 3 months | Against untreated control group | Efficiency, km/L | +7.8% |
| Toyama Transport | Japan · logistics | 1 truck · 20,422 km | Own baseline · 3 applications | Efficiency, km/L | +13.1% |
| Chuo Taxi | Japan · taxi | 1 hybrid · 13,257 km | Own baseline · 2 applications | Efficiency, km/L | +28.3% |
| Hot-spring hotel boilers | Japan · heavy-oil boilers | 2 boilers · 6 months | Year-on-year, same calendar months | Consumption, litres/month | −10.7% … −16.0% |
| IHS Towers | Nigeria · tower generators | 11 sites · 5 months | Own baseline | Saving vs own baseline | +1.7% median — no clear effect established |
| Maruun Transport | Japan · mixed haulage | 9 vehicles · 2022 → 2023 | Year-on-year, same calendar month | Efficiency, km/L | +2.9% median · 3 of 9 negative |
| Toto Kanko Bus | Japan · coaches | 4 coaches · 12 measurements | Own baseline · 2 buses new | Efficiency, km/L | −3.4% … +44.0% · no common trend |
| Nakamoto Transport | Japan · distribution | 3 trucks · 3 months | Own baseline · single application | Efficiency, km/L | +7.2% median |
| Nishitetsu | Japan · route buses | 3 buses · Feb–Jun 2020 | 12-month rolling baseline — pandemic window | Efficiency, km/L | +7.8% median — not claimed |
Each one with its own page
Waste Services
−17.7% fuel consumptionDrivers blinded, 16 weeks of daily data — but the trucks also ran with 8.7% lower utilisation, in a warmer season than the baseline.
Read the trial →Municipality of Frederikshavn
≈5% fuel consumptionCarried out with Scania. The raw figures are not comparable — the result comes from a drive-mode correction.
Read the trial →Alfresa (A・L Plus)
+8.4% … +20.0% km/LYear-on-year comparison of the same months. The same figures appear in three independent records.
Read the trial →Sanden Kotsu
+7.8% km/LThe only trial measured against untreated control vehicles running in the same period.
Read the trial →Toyama Transport
+13.1% km/LThree applications over 66 days, with each recorded value higher than the last — 7.7%, 9.7%, 19.0%; the trial cannot say why.
Read the trial →Chuo Taxi
+28.3% km/LOne hybrid taxi over 106 days. A large result on a single vehicle — shown with its aggregate.
Read the trial →Hot-spring hotel boilers
−10.7% … −16.0% fuel consumptionSix monthly measurements, each against the same month a year earlier. Evidence beyond vehicle engines.
Read the trial →IHS Towers
+1.7% median · no clear effect saving vs own baselineTelecom generators running continuously at near-zero load: a median too small to establish a clear effect for that duty cycle. Published in full because it shows where the boundary is.
Read the trial →Maruun Transport
+2.9% median km/LThree of nine vehicles lost fuel efficiency after treatment. The widest spread in our Japanese records.
Read the trial →Toto Kanko Bus
−3.4% … +44.0% km/LTwelve measurements that do not agree with each other — and two buses that were new at the time.
Read the trial →Nakamoto Transport
+7.2% median km/LThe newest truck recorded the largest change — the opposite of how the mechanism is usually explained.
Read the trial →Nishitetsu
+7.8% — not claimed km/LThe best baseline design on this site, ruined by the months it happened to run in.
Read the trial →Operator records from the master dataset
These 8 operators come from our Japanese master test dataset: 43 vehicles at named fleets, with before-and-after fuel figures we can recalculate but no trial report, no measurement period and no control group behind them. They are not trials, and we do not present them as trials. They are published because we hold them, and because an operator judging us is entitled to see the ordinary records as well as the twelve we chose to write up.
| Operator | Where | Sample | Method as recorded | Result |
|---|---|---|---|---|
| ALSOK (Kanagawa) | Japan · security transport fleet | 14 vehicles | Year-on-year comparison + Before/after, no seasonal control | +8.8% … +30.2% |
| Daiwa Logistics (Misato branch) | Japan · regional freight, 4–6 t trucks | 6 vehicles | Against a reference figure | +9.3% |
| Bridgestone Cockpit store | Japan · tyre retail — customer vehicles | 5 vehicles | Not recorded | +14.1% |
| N Kotsu | Japan · lpg taxi fleet | 4 vehicles | Against a reference figure | −6.6% … +10.1% |
| Sanwa Kotsu | Japan · lpg taxi fleet | 4 vehicles | Against a reference figure | +2.5% … +8.6% |
| Inabe Logistics Service | Japan · tipper trucks and heavy haulage | 4 vehicles | Before/after, no seasonal control | +6.3% |
| Iwaida Taxi | Japan · lpg taxi fleet | 3 vehicles | Against a reference figure | +13.2% |
| Tamagoya | Japan · delivery vans | 3 vehicles | Not recorded | +19.1% |
ALSOK (Kanagawa)
+8.8% … +30.2% km/LThe largest operator group in our dataset — and the clearest illustration that the measurement method, not the vehicle, decides the size of the number.
Read the report →Daiwa Logistics (Misato branch)
+9.3% km/LSix diesel trucks measured against a reference fuel-efficiency figure. The tightest group in the dataset — every vehicle between +4.2% and +17.6%.
Read the report →Bridgestone Cockpit store
+14.1% km/LFive unrelated private cars treated at a tyre retail store. No fleet, no duty cycle, no recorded method — the weakest group we publish.
Read the report →N Kotsu
−6.6% … +10.1% km/LContains the only negative measurement in the entire 386-record dataset — a taxi whose fuel efficiency fell by 6.6% after treatment.
Read the report →Sanwa Kotsu
+2.5% … +8.6% km/LHolds the highest figure in the whole dataset, +75.7% — which we traced to an implausible baseline and do not use.
Read the report →Inabe Logistics Service
+6.3% km/LThe narrowest result in the whole dataset: four heavy vehicles inside a 2.2-point band, and the lowest median of any operator here.
Read the report →Iwaida Taxi
+13.2% km/LTwo conventional taxis and one hybrid. The hybrid returns the largest figure — as it does in every hybrid record we hold.
Read the report →Tamagoya
+19.1% km/LThree vans of the same model and near-identical age — and baselines that differ by a factor of two, which is the part worth reading.
Read the report →All 386 individual measurements, including the 300 we cannot verify →
Why the spread is so wide
A range from +1.7% on telecom generators near idle, too small to establish a clear effect, to 22.1% less fuel on one taxi looks like noise until you sort the trials by how they were measured. What emerges is not that one method gives bigger numbers than another — it is that the trials measured against their own past scatter across the entire range, while the better-controlled ones land in a narrow band.
Method decides the spread, not the level
Of the trials plotted here, those measured against their own history run from 1.7% (telecom generators near idle, too small to establish a clear effect) to 22.1% — the whole range. The two compared with the same months a year earlier sit at 11.6% and 13.0%, and the one with a parallel control group at 7.2%. Better control narrows the spread; it does not simply shrink the number.
Sample size matters more than percentage
A 28% result on one taxi is weaker evidence than an 8% result across a controlled group. We show sample size in the same table as the result so the two are never separated.
Results differ by application
Refuse collection, urban taxi work and boiler operation showed the larger results. Telecom generators running continuously at near-zero load showed a median of +1.7% — too small to establish a clear effect for that duty cycle. These are observations across different trials, not a measured cause.
Later readings were higher
Where trials ran several applications, later measurements were consistently higher than early ones. Route, load and season can explain part of that; a single early measurement may understate a result.
Run the trial that actually applies to you
None of these trials is a promise about your vehicles. A pilot on your own fleet, designed with a control group, is the only figure worth budgeting against.