STATCON 99: Why 580 kg May Be More Than You Think
Technical Field Note · Flight Planning & Decision-Making · 7-minute read
Beyond the Checklist — Aviation & Human Performance
Reviewed: 28 October 2026 · Editorial draft — not published
“Only 580 kilos?”
Imagine a long-haul preflight briefing. The operational flight plan shows a contingency fuel figure of 580 kg, marked CONT99.
The First Officer looks at the number and raises a perfectly reasonable concern:
“That seems very little for such a long flight. Shouldn't we take another three tonnes?”
It is a good question. But before we decide whether 580 kg is enough, we need to ask a better one:
What does CONT99 actually mean?
The answer exposes a surprisingly common confusion: treating a percentile as though it were a percentage.
A percentage measures fuel. A percentile measures a distribution.
A percentage is straightforward. If the planned trip fuel were 100,000 kg, then 5% of trip fuel would be 5,000 kg. The figure is calculated directly from the planned quantity.
A percentile answers an entirely different question. CONT99 refers to a statistical coverage level for the difference between planned and actual trip fuel. It does not mean 99% of trip fuel, nor that 99% of the aircraft's fuel has somehow been accounted for.
Under a correctly calibrated statistical method, the 99th-percentile threshold represents a fuel-consumption deviation that approximately 99% of comparable flights would not exceed. The remaining 1% can exceed it.
This is a statement about a distribution of past or modelled outcomes—not a promise about what any one flight will consume.
One B777 trip. One thousand comparable flights. Three different answers.
Consider a fictional set of 1,000 B777 flights, each with the same planned Trip Fuel of 100,000 kg (100 tonnes). That is a plausible long-haul planning magnitude, not an actual route-specific OFP. For clarity, the observations below are grouped into eight simplified values. This is a constructed teaching dataset, not operational flight data.
| Number of flights | Planned Trip Fuel | Actual Trip Fuel | Deviation | Cumulative flights |
|---|---|---|---|---|
| 100 | 100,000 kg | 99,400 kg | −600 kg | 100 |
| 200 | 100,000 kg | 99,850 kg | −150 kg | 300 |
| 300 | 100,000 kg | 100,090 kg | +90 kg | 600 |
| 200 | 100,000 kg | 100,220 kg | +220 kg | 800 |
| 100 | 100,000 kg | 100,320 kg | +320 kg | 900 |
| 50 | 100,000 kg | 100,400 kg | +400 kg | 950 |
| 40 | 100,000 kg | 100,580 kg | +580 kg | 990 |
| 10 | 100,000 kg | 100,900 kg | +900 kg | 1,000 |
First question — what was the average deviation? Add all the deviations, weighted by the number of flights, then divide by 1,000:
(100 × −600 + 200 × −150 + 300 × 90 + 200 × 220 + 100 × 320 + 50 × 400 + 40 × 580 + 10 × 900) ÷ 1,000 = +65.2 kg.
Expressed as a percentage of the 100,000 kg trip plan, that average deviation is +0.0652%, equivalent to +65.2 kg. Averaging historical deviations is a useful teaching comparison; it is not the regulatory fixed-percentage contingency method.
Second question — what threshold covers 90%, 95% or 99% of these deviations? Order the 1,000 observations by deviation and identify the selected rank (here using the nearest-rank convention):
- 90th percentile (rank 900): +320 kg. At least 900 of the 1,000 deviations are at or below this value.
- 95th percentile (rank 950): +400 kg. At least 950 are at or below it.
- 99th percentile (rank 990): +580 kg. At least 990 are at or below it; 10 are higher.
The +580 kg is therefore produced directly from the same 1,000 observations as the +65.2 kg mean. In this particular dataset, the 99th-percentile deviation is almost 8.9 times the average deviation. This illustrates why an average and a high percentile answer different questions; it does not mean that a percentile is always larger than any percentage.
Third question — what would a fixed percentage of Trip Fuel produce? A hypothetical 5% × 100,000 kg = 5,000 kg. That figure does not come from the observed deviations at all. It is a fixed fraction of the planned trip fuel, whereas the +580 kg represents the 99th-percentile deviation of this constructed dataset.
100 tonnes planned. What does 99 mean?
One entirely fictional B777 long-haul dataset, three genuinely different questions.
1,000 modelled observations
| Flights | Actual fuel | Delta vs plan | Cumulative |
|---|---|---|---|
| 100 | 99,400 | −600 kg | 100 |
| 200 | 99,850 | −150 kg | 300 |
| 300 | 100,090 | +90 kg | 600 |
| 200 | 100,220 | +220 kg | 800 |
| 100 | 100,320 | +320 kg | 900 |
| 50 | 100,400 | +400 kg | 950 |
| 40 | 100,580 | +580 kg | 990 |
| 10 | 100,900 | +900 kg | 1,000 |
Deviations are sorted from lowest to highest. Each row is a group of fictional flights assigned a simplified representative deviation.
Rank 900 / 1,000
Rank 950 / 1,000
Rank 990 / 1,000
Same data. Different result.
What about a fixed 5%?5% × 100,000 kg is a separate percentage-of-trip calculation. It does not use these 1,000 deviations.
5,000 kgConstructed data and a nearest-rank percentile illustration, not a real airline data sample or fuel-release method. Operational STATCON values depend on validated data, the approved fuel scheme and applicable minimums. The 99th-percentile number is not necessarily larger than a fixed percentage.
Same flights. Three different questions. The average describes the centre of this dataset, the percentile describes a selected upper-tail threshold, and a fixed percentage applies a predetermined fraction to the plan. In a real operation, the operator's approved scheme, calculation details and any applicable minimum fuel floor determine what actually appears on the OFP. Different percentile algorithms may also produce slightly different results on a small or grouped sample.
The number behind the number
For each relevant historical flight, the starting point is the difference:
Trip fuel deviation = Actual trip fuel − Planned trip fuel
Suppose a flight was planned to burn 100,000 kg but actually burned 100,350 kg. Its deviation is +350 kg. If the flight instead burned 99,200 kg, the deviation is −800 kg.
Now imagine an illustrative dataset of 1,000 suitably comparable flights, with corrected and validated fuel data. If their 99th-percentile deviation were +580 kg, approximately:
- 990 flights would have deviations at or below +580 kg.
- 10 flights would have deviations above +580 kg.
The 580 kg would not be the maximum overburn ever recorded. It would not describe the aircraft that burned the most total fuel. It would describe a threshold in the distribution of the difference between planned and actual trip fuel.
And the distribution does not have to be a perfect bell curve. Real operational data may be skewed or contain outliers. The result depends on the quality and representativeness of the data and on the statistical method used.
This is why data preparation matters: aircraft performance, actual versus planned mass, routes, winds, operational variations and other factors can influence comparisons. Regulations require a sufficiently meaningful dataset rather than a handful of loosely comparable flights.
A crucial distinction: the printed 580 kg may not be the raw percentile
There is another reason not to reverse-engineer a flight plan from one figure.
Under the EASA AMC6 CAT.OP.MPA.181 basic-fuel-scheme variation, statistical contingency is compared with a minimum amount equivalent to five minutes at holding speed, 1,500 ft above the destination aerodrome, in standard conditions. The applicable contingency figure is the higher amount.
So an OFP labelled CONT99 and showing 580 kg does not, by itself, prove that the raw 99th-percentile deviation was exactly 580 kg. The holding minimum may have determined the displayed amount, depending on the implementation and presentation of the flight plan.
That distinction matters. We should understand the method, but we should not claim to know how a particular OFP value was produced without its underlying calculation.
Why 90 and 99 are not interchangeable
The choice of statistical coverage is not simply a pilot selecting a preferred risk level.
For example, GM2 CAT.OP.MPA.181 in EASA material describes conditional scenarios: 99% statistical coverage for flights longer than two hours where an eligible fuel en-route alternate is available; 90% coverage only when additional conditions concerning the en-route alternate, two usable destination runways and weather criteria are met. In other circumstances, the guidance describes 99% coverage plus 3% of trip fuel.
You may encounter a 95th-percentile label in other contexts, but 95% is not one of the specific coverage examples given in these GM2 provisions. The applicable approved scheme must determine its use.
These examples illustrate a central principle: the statistical threshold belongs to a wider system of operational mitigations and fuel-planning requirements. A lower percentile is not inherently an unsafe shortcut, just as a higher percentile is not an unconditional guarantee.
Different approved fuel schemes and jurisdictions may have different requirements. The applicable rules and approved procedures always govern.
Does 580 kg mean we have only 580 kg available if something goes wrong?
No. Contingency fuel is just one element of the fuel plan. Depending on the operation, the plan also includes trip fuel, destination alternate fuel (or the applicable no-alternate provision), final reserve fuel, any required additional fuel, extra fuel for foreseeable constraints, and discretionary fuel.
These amounts have different purposes. They are not an unrestricted pool to spend casually, but neither should they be collapsed into the single number printed next to CONT99.
Equally important, statistical contingency deals with uncertainty represented by the model. A specific threat that is already known—for example, an anticipated delay or operational restriction—must be evaluated on its own merits and accounted for in accordance with the applicable fuel scheme. Statistics do not make current information disappear.
If the actual trip fuel deviation exceeds the statistical threshold, that is not automatically a fuel emergency. It is a reason to assess the evolving fuel situation against planned landing fuel, required reserves, available options and the approved in-flight fuel-management procedures.
Back in the briefing: would you add three tonnes?
The best answer is neither an automatic yes nor an automatic no.
A thoughtful crew would first establish what the 580 kg represents, whether the approved planning method and its assumptions apply to today's flight, what known threats have been accounted for, and how the complete fuel plan addresses the operation. The commander can then make a reasoned decision about any discretionary fuel.
The aim is not to persuade pilots to carry less fuel. It is to replace a reaction to an isolated number with a well-informed assessment of risk.
The takeaway
A percentage tells you how much. A percentile tells you where a threshold sits in a distribution.
CONT99 is not “99% of the trip fuel”. Nor does it mean the worst flight ever exceeded the plan by only the stated amount. It reflects a statistical approach to fuel-consumption deviations, subject to data quality, regulatory conditions and potentially a separate minimum fuel floor.
The next time a small contingency figure catches your eye, resist the temptation to judge it by its size alone.
Before asking whether the number is enough, ask what the number means.
What would you do?
Your OFP shows CONT99 — 580 kg. A colleague recommends adding 3,000 kg because the contingency “looks too small”.
What would you verify before agreeing or disagreeing?
Debrief: Explain the difference between the percentile and the fuel quantity; verify the actual contingency calculation and any minimum floor; review the complete fuel plan and known operational threats; apply the approved fuel scheme; and exercise informed commander judgement.
Sources and further reading
- EASA, Easy Access Rules for Air Operations, Revision 24 (March 2026): CAT.OP.MPA.181, AMC6 CAT.OP.MPA.181 and GM2 CAT.OP.MPA.181. https://www.easa.europa.eu/en/document-library/easy-access-rules/online-publications/easy-access-rules-air-operations?erules-id=ERULES-1963177438-12803
- M. Lingemann and D. Savchev, “Case Study: Optimizing Statistical Contingency Fuel at Lufthansa Cargo,” Aircraft IT Operations (2019), for background on historical data, normalization and percentile-based fuel planning. https://www.aircraftit.com/articles/case-study-optimizing-statistical-contingency-fuel-at-lufthansa-cargo/
Independent educational content for reflection and discussion. The numerical scenario and sample dataset are fictional. This article is not an operating procedure, a fuel-release recommendation or a substitute for current approved manuals, applicable regulations and operational judgement.
