Data Methodology
How NetbackBeacon collects inputs, how often they refresh, and how calculated values are produced. Figures that are not sourced from a connected provider are labelled as sample or indicative data in the interface.
Last updated:
Data collection
Connected public providers currently include FRED (macro and energy benchmark series), EIA (US inventories and product data), AISStream (vessel positions, where a key is configured) and GDELT (news and event signals).
Any module without a connected provider renders sample structures for illustration and is labelled accordingly.
Update frequency
FRED and EIA series follow the publisher's own release calendar (daily, weekly or monthly depending on the series). Scheduled background jobs refresh cached series and replay history.
AIS and news signals are refreshed on demand while a session is open, at the interval selected in the interface.
Data latency
Public statistical series are inherently lagged: EIA weekly stocks publish several days after the reference week, and several FRED series publish with a one- to two-day delay.
Displayed timestamps always refer to the observation date of the underlying series, not the moment the page was loaded.
Calculation methodology
Netback figures are computed as destination price less freight, less applicable handling, loss and time-value components, expressed in the product's native unit.
Scenario and replay outputs re-price the same engine with a modified market state, so results are comparable against a fixed baseline.
Estimated and indicative values
No interpolation is used. Where an input is missing, the last stored observation may be shown with its own observation date and a stale/delayed label — never as a new observation. Exception: the replay range mode carries the last FRED value forward without a maximum age; it shows the source observation date and that the value is carried. See the missing-data policy on the Data Sources page.
Risk scores that are not derived from a verifiable third-party index are presented as indicative risk scores.
Known limitations
Assessed physical differentials, term freight rates and proprietary LNG assessments are not licensed on the platform today.
AIS coverage depends on the terrestrial and satellite footprint of the upstream feed and is not exhaustive.
Details will be published before commercial launch.
Indicative signal agreement score
Measures how strongly, and in how consistent a direction, the drivers behind an opportunity point. It is computed in code as: score = clamp(35, 96, (mean |driver contribution| × 0.55 + directional agreement × 0.45) × 100), where directional agreement = |sum of driver signs| ÷ number of drivers.
Inputs: three drivers per opportunity, each a contribution between −1 and +1. Some drivers are fixed reference constants, not observed data. Period: current sample inputs only. Missing-data effect: a driver with no data contributes 0, which lowers the score. The 35 floor and 96 cap are display bounds, so 96/100 means 'maximum displayed agreement', not certainty.
- Not a probability of success — the score is not calibrated as a probability.
- Not historical forecast accuracy — it has not been back-tested.
- Not a data-quality score — input freshness is shown separately by data-status labels.
- Not model confidence in a price forecast — no price is forecast.
Opportunity status thresholds
Each opportunity's strength score (0–100) = 50 + 30 × (positive contributions − 0.7 × negative contributions), clamped. Open ≥ 75, Watch 50–74, Closed < 50. These are rule thresholds on sample signals, not a netback opening threshold.
Netback, arbitrage margin and route difference
- Netback
- Destination sales value − modelled delivery costs, per tonne. Source product cost is not deducted.
- Arbitrage margin
- Destination netback − origin (source) price. Requires a source price; where none is modelled the margin is shown as Not modelled.
- Difference vs next-best route
- Best route netback − second-best route netback. Says nothing about absolute profitability.
Crack spread: indicator only
In the current code the crack spread is an explanatory market indicator (a driver in opportunity status). It is not added to netback: netback uses the product sales price directly, so adding the crack would count refining margin twice.
LNG: indicative netback comparison
LNG rankings are an indicative netback comparison. Deliverability, terminal/regas slots and contractual rights must be verified separately; a ranking is not a guarantee that a trade can be executed.
- TTF is quoted in EUR/MWh; JKM in USD/MMBtu. Convert TTF: 35 EUR/MWh × 1.08 USD/EUR ÷ 3.412 MMBtu/MWh ≈ 11.08 USD/MMBtu.
- Compare like with like: JKM sample 12.40 USD/MMBtu for the same delivery month → gross spread ≈ 1.32 USD/MMBtu.
- Basis still differs: JKM is DES North-East Asia, TTF is a Dutch virtual hub price — regas, slot and shipping costs must be deducted before any netback conclusion.
- All numbers are illustrative sample values, not market quotations.