5-year trend — revenue, cash flow & FCF ($B)
How this company converts strategy into free cash flow
Each panel shows FY2021→FY2025. Current value labelled; click a company chip above to switch.
Year-end balance-sheet line items as reported (standardized), $B. The 3-statement model is seeded from these exact figures.
Balance sheet by year ($B)
Net debt & leverage
Free cash flow trend, FY2021→FY2025 ($B)
Click any company (chip, bar, or row) to load it into the model. Table reflects the selected year; trend chart shows all five years.
Integrated three-statement model — seeded from exact FY balance sheet, fully editable
Revenue & EBITDA
Cash & free cash flow
Drivers
Income statement ($B)
Balance sheet ($B)
Cash flow statement ($B)
Working capital & rates
Scenario analysis
EBITDA by scenario
Ending cash by scenario
Final-year comparison
Two-variable sensitivity
Heatmap
Rows/columns step ±2 around base; center cell (white border) = base case.
Discounted cash flow valuation
Unlevered free cash flow ($B)
Enterprise value bridge
Seed the model from a past year's actuals, forecast forward with steady drivers, and compare to what really happened. The gap is your error — watch where commodity swings break a naive model.
Projected vs actual ($B)
Forecast error by year
A simplified linear price model: EBITDA/FCF move with Brent at the sensitivity below (default from this company's history; adjustable), anchored to the selected year's actuals. The breakeven is the oil price at which free cash flow just covers each payout.
Free cash flow vs. payout across Brent ($B)
Feed the market price in and back out what it implies. If the implied growth looks heroic or trivial versus your base case, you've found a possible mispricing — or a flaw in your assumptions.
DCF value/share vs. revenue growth — where the market sits
1,500 trials drawing revenue growth, gross margin, capex intensity and WACC from distributions around your base case. Read the spread, not the point. Results are seeded (reproducible); Re-run advances the seed.
Unlevered FCF fan — P10 / P50 / P90 ($B)
DCF value / share distribution
Forecast risk — driver impact & sanity checks
The tornado ranks which assumptions move value most (so you know where to spend your research). Guardrails flag the classic forecasting mistakes.
Tornado — swing in DCF value/share
Guardrails
Variance: actual vs. plan — plan = model's first forecast year
Type actuals to see the bridge. Favorable = higher revenue/profit or lower cost.
Variance by line ($B)
Governance & disclosures — model card, data lineage, limitations
This tool is documented against IBM's AI governance framework: the Principles for Trust & Transparency, the five Pillars of Trustworthy AI, and Everyday Ethics for AI. It is a deterministic, rules-based financial model (not a machine-learning system); the governance controls below apply the same standard of transparency, explainability and accountability.
1 · Intended use & "augment, not replace"
Intended use: internal financial education, scenario exploration and decision-support for analysts who will apply their own judgment and verify against primary filings. It surfaces relationships (price → cash flow, drivers → value) to augment human reasoning.
Out of scope: investment advice, buy/sell/hold recommendations, automated trading, regulatory filings, audited valuations, or any use where a person relies on an output without independent review. No output is a recommendation. Human oversight is required.
2 · How this maps to IBM's pillars
| Pillar / principle | How it is addressed here |
|---|---|
| Transparency | Persistent AI-assisted / not-advice banner; this model card; version, owner and as-of dating in the footer; every estimated input is tagged. |
| Explainability | The full method is documented below and visible in-product (driver inputs, balance-sheet seeds, DCF build, price sensitivity). No black box — every number is reproducible from inputs. |
| Robustness | Monte Carlo uses a seeded pseudo-random generator, so results are reproducible and auditable; the seed is shown and only changes when you deliberately re-run. Guardrails flag invalid assumptions. |
| Fairness / Value alignment | Company narratives are editorial summaries of public strategy and are labelled as such; they are not endorsements. Coverage and driver estimates reflect disclosed analyst judgment, not a neutral oracle. |
| Privacy / User data rights | No personal data is collected. The only stored data is your last-selected company/year/segment in this browser's localStorage; you can clear it below. |
| Accountability | Owner, version and review status are recorded in the footer; limitations and known biases are disclosed; the tool defers decisions to the user. |
| Data belongs to its creator | Third-party financial data is attributed to its providers and is subject to their terms; users must verify against primary SEC/company filings before relying on any figure. |
3 · Methodology (explainability)
Three-statement engine: a driver-based model links the income statement, balance sheet and cash flow; the balance sheet is forced to tie every year (checked live). Forecasts are seeded from the selected company-year's reported revenue, EBITDA margin, and exact balance-sheet line items (cash, net PP&E, debt).
Free cash flow is defined uniformly as operating cash flow − capex (organic), which can differ from a company's own reported "free cash flow." DCF discounts unlevered FCF at WACC with a Gordon terminal value; outputs are illustrative and highly sensitive to WACC and terminal growth.
Oil-price sensitivity is a deliberately simplified linear approximation (ΔEBITDA per $1/bbl Brent), estimated from each company's history. It is an ESTIMATE, not a fitted statistical model, and should not be read as a precise elasticity.
Monte Carlo draws revenue growth, gross margin, capex intensity and WACC from normal distributions around the base case using a seeded generator. Probabilities are model artifacts, not real-world frequencies.
4 · Data provenance & lineage
| Item | Origin | Status |
|---|---|---|
| Revenue, net income, EBITDA, balance sheet (FY2021–25) | Standardized financials, Fiscal.ai / S&P Global Market Intelligence via stockanalysis.com (≈ May–Jun 2026) | SOURCED |
| Operating cash flow & capex split | Company results / 10-Ks; FCF computed as CFO − capex | SOURCED |
| Production (mboe/d) | Company reports; some early-year figures approximated | SOURCED / EST |
| Share prices | Recent quote via stockanalysis.com (CTRA = last trade pre-delisting) | SOURCED |
| Oil-price EBITDA sensitivities, Brent reference | Internal estimate from historical EBITDA vs Brent | ESTIMATE |
| Forecasts, DCF value, reverse-DCF, breakeven, Monte Carlo | Model outputs from user-editable assumptions | ILLUSTRATIVE |
| Strategy narratives (GTM / product / treasury) | Editorial summaries of public disclosures | EDITORIAL |
"Standardized" third-party figures may differ from as-filed statements. BP's net income is shown as underlying replacement-cost profit (GAAP was distorted by impairments). Verify any figure against primary SEC/company filings before use.
5 · Limitations & known biases
Simplified single-segment model (integrated downstream/chemicals not modelled separately); linear price sensitivity ignores hedging, mix and non-linear effects; DCF is point-in-time and assumption-driven; per-share figures imply false precision and should be read as ranges; coverage is limited to 12 large-cap names and excludes smaller/private operators; the 2022 commodity spike makes naïve trend forecasts unreliable (see Backtest). Narratives may carry optimistic framing from company disclosures.
6 · Privacy & your data
No personal data, analytics or network calls leave your browser at runtime (data is embedded; charts load from a CDN). Your last selection is stored locally only.