Where a cargo earns the most, and what feed gas will cost.
A U.S. LNG seller compares what each market pays after shipping, boil-off and regasification. This page models those netbacks, checks them against where Venture Global's cargoes actually went and what it reported earning, and tests how well the Henry Hub feed-gas price can be forecast.
Which market paid more after shipping costs?
Method. FOB netback at Calcasieu Pass = delivered price × (1 − boil-off) − charter and canal costs − regasification, for a 3.6 Bcf cargo to Rotterdam or Tokyo. Sea distances come from the open-source searoute network. The probability panel repeats the calculation 4,000 times per month with every assumption drawn from its low–base–high range.
Did cargoes go where the netback was better?
Limits. Only Europe-versus-Asia volume is compared. Long-term contracts, buyer destination rights and the Asia price proxy all weaken the link, so a moderate correlation is the expected ceiling, not a disappointing result.
How do reported prices compare with the model?
Source. Weighted average price of LNG sold per project and quarter, from Venture Global's 10-Q and 10-K filings (Q4 derived from annual totals). Modeled netback is volume-weighted over each quarter's actual Europe and Asia cargoes. Realized prices include intercompany sales to VG Commodities and mix of FOB and delivered sales.
Can the Henry Hub feed-gas price be forecast weeks ahead?
Test error, January 2022 onward
Outlook from the latest week
Point forecasts only. With no model reliably beating "no change", treat these as a sanity check, not a trading signal.
Method. Weekly averages of the EIA daily price; log-price changes 1, 4 and 8 weeks ahead. Each test year is predicted by models refit only on data whose targets were known before 1 January. The Conv1D + LSTM reads the last 52 weekly changes plus a seasonal cycle, is an average of 5 seeds, and uses early stopping on the latest 15% of its training period. The Diebold–Mariano test compares each model's absolute errors with the naive forecast.