Case Studies

Up to 99.99% lower residual squared error

Gaussian benchmarked physics-based DCA against Arps on two Wolfcamp wells, reducing residual squared error by 99.99% and 98.35%, respectively. It generated 40-year oil EURs of 201 and 340 Mbbl and used field-calibrated diffusivity distributions to quantify reserve uncertainty.

Exposed $7.50B of peak cash-flow risk

Instead of relying on one deterministic valuation, Gaussian tested 10,000 probabilistic development outcomes and showed a P50 NPV10 of $47.21B versus a prior $15.97B benchmark, while also exposing $7.50B of peak cash-flow risk.

P90 misses 4.5 Mt CO₂ target

For Nini West, GPT did more than confirm storage potential — it exposed the risk range behind the 4.5 Mt CO₂ target. While the deterministic forecast reached 8.38 Mt and the P50 case reached 7.73 Mt, the P90 case was only 3.25 Mt, showing that the target is technically achievable but vulnerable to injectivity, completion, pressure-management, and data-quality risks.

Tighter spacing demonstrated 70,000-barrel gain

Gaussian showed that completion design choices in shales matter: 30-ft fracture spacing delivered ~70,000 bbl more than 100-ft spacing over 3,000 days, while higher-conductivity proppant increased cumulative production from ~170,000 to ~230,000 bbl.

Fracture half-lengths matched within 9.5%

Gaussian converted routine post-frac reports of Eagle Ford 22-stage well into stage-level fracture dimensions and validated the result against independent production-history matches using CMG-IMEX, with only ~9.5% difference from GPT.

Up to 30% modeled production uplift

In a mature Gulf of Mexico reservoir with >100 million bbl produced, Gaussian corrected a misleading depletion interpretation and showed that pore-fluid expulsion could add up to ~30% relative to Darcy-only forecasting.