Evidence · Fleet trials

Toto Kanko Bus — Japan

Four coaches were measured in April, May and June 2023 against a March baseline. The twelve measurements range from −3.4% to +44.0% and they do not point the same way: KB-11 falls from +10.3% to 0.0% to −3.4% across the three months, while KB-5 climbs from +12.0% to +44.0%. Two of the four buses were new. We publish this because the disagreement is the finding.

Operator record — primary data not held Coach operator

Twelve measurements on four buses in one depot, going in different directions. Any single one of them could be quoted as a result; together they are a caution.

−3.4% … +44.0%Range across all twelve measurements

+16.9%Median of the twelve measurements

2 of 4Buses that were new at the time of the trial

0.1 km/LPrecision of the recorded figures — about 4% at these values

Toto Kanko Bus trial — at a glance
Operator	Toto Kanko Bus, Tokyo — coach operator with more than 200 buses
Vehicles	4 coaches, of which 2 were new at the time of the trial
Baseline	March 2023, per bus
Measurements	April, May and June 2023 — twelve in total
Measured quantity	Fuel efficiency in km/L, recorded to one decimal place
Result	Range −3.4% to +44.0%, median +16.9%
Trend	None common to the fleet — one bus improves across the months, one declines

The result

Every measurement, not just the best one

Toto Kanko Bus — Every monthly measurement against each bus’s own March 2023 baseline. Higher km/L is better. Distance is shown because the monthly workload varied considerably.
Bus	Month	March baseline (km/L)	Measured (km/L)	Distance in month (km)	Change
KB-5 (new bus)	April	2.5	2.8	2588	+12.0%
KB-5 (new bus)	May	2.5	2.9	3765	+16.0%
KB-5 (new bus)	June	2.5	3.6	4210	+44.0%
KB-11	April	2.9	3.2	3142	+10.3%
KB-11	May	2.9	2.9	3907	+0.0%
KB-11	June	2.9	2.8	4386	−3.4%
KB-15	April	2.5	2.7	2930	+8.0%
KB-15	May	2.5	3.3	3913	+32.0%
KB-15	June	2.5	3.0	5756	+20.0%
KB-55 (new bus)	April	2.8	3.8	3926	+35.7%
KB-55 (new bus)	May	2.8	3.3	4198	+17.9%
KB-55 (new bus)	June	2.8	3.5	4388	+25.0%
Median across all twelve measurements					+16.9%

Fuel efficiency, km/L — higher is better · dashed line is each bus’s own March 2023 baseline · March 2023 baseline

Read the four panels against each other, not individually. KB-11 falls below its own baseline by June while KB-5 climbs away from it — on the same fleet, in the same months. Both new buses are among the strong risers, which is exactly what a running-in effect would look like. And note the axis: at these values one recorded decimal place is a step of about four per cent.

Toto Kanko Bus — how it was measured

Method before result

Each bus was measured monthly against its own March 2023 figure. The treatment was applied after the baseline month, and April, May and June were then compared against it.

Two of the four coaches — KB-5 and KB-55 — were new vehicles at the time of the trial. A new diesel engine improves its own consumption over the first months of service as it beds in, which is an effect that runs in the same direction as the one being measured and cannot be separated from it here. Those two buses also produced the two largest figures in the table.

The precision of the source data is a problem in itself, and it is the reason we lead with it in the summary. Consumption is recorded to one decimal place, at values around 2.5 to 3.6 km/L. At that magnitude a single step of 0.1 km/L is roughly four per cent, so a measurement of "2.5 → 2.8" carries a rounding uncertainty of several percentage points before anything else is considered.

We recalculated all twelve percentages from the recorded values. They reproduce the source, which confirms the arithmetic and nothing about the readings.

How we measure, and what makes a fuel trial credible →

Toto Kanko Bus — what this result does not prove

The limits of this trial

We publish these next to the result rather than in a footnote. If a limitation would change how you read the number, you should read it at the same time as the number.

Known limitations

Two of the four buses were new, and the running-in effect on a new engine cannot be separated from the treatment effect in this design. The largest figure in the table, +44.0%, is on one of those new buses.
The figures are recorded to 0.1 km/L, which at these consumption levels is a rounding step of about four per cent. Several of the smaller results in this table are within that uncertainty.
KB-11 shows the opposite of the expected pattern: +10.3% in April, no change in May, −3.4% in June. If the effect built over time as our other trials suggest, this bus should have improved. It did not, and we have no explanation.
Monthly distances vary by up to a factor of three between the baseline month and the measured months, so the comparison mixes changes in workload with changes in consumption.
There is no control group, and coach work — charter routes, passenger loads, motorway share — varies far more month to month than scheduled service work.
We hold a summarised record of this trial, not the operator’s primary data. The method, the periods and the per-vehicle figures are documented; the underlying fuel and distance logs are not in our possession, and we have asked for them.

Toto Kanko Bus — what it means

Reading this evidence in context

This trial was previously left off the site because its results were inconsistent and two of the buses were new. That was the wrong call. A dataset that disagrees with itself is exactly the kind of thing an operator needs to see before believing a number from anyone, and leaving it out made the rest of the evidence look tidier than it is.

The practical lesson is about trial design rather than about the product: do not run a fuel trial on new vehicles, record consumption to more than one decimal place, and keep monthly workloads comparable — otherwise the measurement uncertainty swallows the effect you are trying to find.

What this dataset supports is nothing more precise than "some buses improved and one did not". We would not quote a figure from it.

Toto Kanko Bus — FAQ

Questions operators ask about this trial

Why is +44% not on your front page?

Because it is a measurement on a new bus, recorded to one decimal place, against a baseline month in which that bus drove 1,465 km. Every one of those facts weakens it, and the same table contains a bus that got worse over the same period. Quoting the 44% would be the clearest possible example of the behaviour this evidence section exists to avoid.

Why does one bus get worse while another improves?

We do not know. KB-11 declined from +10.3% to −3.4% across three months while KB-5 rose sharply over the same period. Charter coach work varies enormously with route, passenger load and motorway share, and none of that is recorded here. With four vehicles and no control group, the honest answer is that this trial cannot tell us.

Does the running-in effect really matter that much?

On a new diesel engine, yes — consumption typically improves over the first months in service on its own. Two of these four buses were new, and both produced large positive figures. That does not mean the treatment did nothing; it means this trial cannot show what it did.

Toto Kanko Bus — where this data comes from

Provenance

Operator: Toto Kanko Bus, Tokyo
Fleet context: More than 200 coaches; 4 in the trial
Baseline: March 2023
Measurement period: April to June 2023
Source: Japanese field-study summary — derived record
Recalculation: All twelve percentages recomputed from the recorded km/L values
Open items: Primary records requested; delivery dates of the two new buses; monthly passenger and route data
Last checked: 15 August 2026

Prove it on your own fleet

The only baseline that matters is yours

Every trial on this site was run on someone else’s vehicles, duty cycle and baseline. A NanoEFX pilot measures yours — and we will help you design it with a control group so the result stands up.

Book a risk-free pilot All trial results → Calculate your ROI → Ask a question