Finding a similar well is the start of an investigation. The useful question is why that well is relevant to the planning decision in front of you.
Depth and location are accessible descriptors. Geological sequence, pressure, rig capability and trajectory may also matter. A practical retrieval system needs to show which dimensions it uses and which important differences it cannot represent.
Make selection inspectable
An analogue list should come with enough context to understand the selection. Which filters were applied? Which features drove the ranking? How much support is available? Can the user inspect the underlying records?
Saving that selection alongside an estimate makes later investigation possible. Without it, a model change or new data import can silently change the apparent explanation of a historical forecast.
What Strata currently implements
Strata is an independent research prototype. It supports map-based well exploration and analogue retrieval using depth and geographic distance, with fixed weighting and additional regime and campaign filters.
The prototype produces drilling-duration quartiles and metres-per-day estimates. Saved forecasts retain the selected analogue wells, and diagnostics expose error, interval coverage and sample support.
This implementation does not establish calibrated cost distributions, formation-specific geological modelling or campaign uncertainty aggregation. Support-based confidence labels are not validated probabilities.
Design questions for further work
A richer representation could include geological and operational dimensions where reliable data is available. Expert annotations might help document why an analogue is useful or misleading. Adjustable weighting could support exploration, provided it does not become a way to choose a preferred answer after seeing the result.
These are research directions, not descriptions of delivered Strata features. Each needs a defined question, usable data and evaluation against a simpler baseline.
Evaluate the question that will actually be asked
A system intended for planned wells must be evaluated with information that would be known at planning time. Outcome-derived features need particular scrutiny. Leaving a well out of the retrieval pool is insufficient if other inputs still reveal its realised performance.
That evaluation work is part of the boundary between a useful prototype and an operational tool. Strata has not been used in a client engagement, and no measured client outcomes are claimed.
Read the current prototype scope or arrange a technical demonstration.