Things we found in the data.

This chart shows 30 Delaware operators, names removed. Each line is the median half-cycle IRR of the wells an operator would drill each quarter over the next decade, at flat $75 oil, one shared cost model, and each operator drilling its best remaining rock first at its own pace. Because costs and prices never change, every decline you see is rock depletion.

The dashed line is the whole basin. Its rebound is survivorship. Operators with thin inventory either fall below a 10% hurdle or exhaust their locations, and the median shifts to the few companies that still have deep, high-quality rock.

For many portfolios, the "10-year runway" is really a 3 to 5 year runway followed by a cliff.

The Geo Quality view maps predicted 12-month oil per lateral foot across the Midland Basin, from P10 7.4 to P90 19.0 bbl/ft, and fades where fewer wells support the estimate. Draw an east-west or north-south line and the panel beside it stacks the geology contribution to production along that line, attribute by attribute.

Going west to east across Martin and Howard the geology effect swings from plus 25 percent to minus 10 percent in under 90 miles. That is the kind of number an acreage conversation needs before anyone talks about completion design.

Every section in the Permian, colored by how many undrilled locations are left once the developed laterals, the spacing assumption and the zone extents are accounted for. Grey sections are fully developed for their quality tier. The map is built from the laterals themselves, not from an operator slide.

Inventory is a function of the density decision, so the map changes when the spacing assumption does. That sensitivity is the point: it shows where the remaining locations are robust and where they only exist on paper.

Measured against the most isolated wells in each basin, the most crowded wells give up 26.6 percent of 12-month oil in the Midland, 25.5 percent in the Delaware, 38.2 percent in the Bakken, 39.8 percent in the Denver-Julesburg and 35.3 percent in the Anadarko. The curves differ in shape as well as depth, which matters for where the next well goes.

The second panel tracks the median drainage score of each year's new wells. Permian wells have been getting more crowded since 2023 while Bakken and DJ wells have been getting more isolated, so the degradation curves are being walked in opposite directions.

Each lateral is colored by the angle between its azimuth and the local maximum horizontal stress. Most Midland wells sit in the 80 to 90 degree band; the Delaware is far more mixed, with thousands of wells drilled within 60 degrees of SHmax.

The inset plots the performance residual against that angle. The relationship is flat until roughly 60 degrees and then climbs steeply, which is the behavior you would expect from transverse fractures and a useful check on any acreage where the azimuth was set by the lease line.

The short video starts on the basin picker, opens a Delaware project and colors 42,546 wells by thermal maturity in a few clicks. Pressure gradient, reservoir temperature, organic content and every other attribute sit in the same dropdown, already joined to the laterals.

The same allocated production, surveyed laterals, standardized landing zones and spacing measurements sit under every basin, so a type curve in the Uinta is built the same way as one in the Delaware, and an analog search does not stop at a basin boundary.

For a planned pad in Loving County the engine selected 25 Wolfcamp A analogs with 36 months of cohort production, bolded them on the map in their operator or formation color and left every other Wolfcamp A well in thin grey. Click a dashed neighbor to add it, click a cohort well to remove it, and the type curve rebuilds.

Analog selection is usually the least documented step in a forecast. Putting it on a map, with the feature filters and cohort size exposed, makes it something a reviewer can argue with.

Among 42,547 Delaware wells the survey-based laterals pick out the horseshoe wells immediately, highlighted in red against the straight laterals around them. Filtering by column attributes and lasso narrows the set to the 16,313 wells that matter for the comparison.

The short video shows the map doing the work. U-laterals are now common enough that a spacing or inventory study that treats them as straight sticks is wrong on both counts.

The platform walk-through follows a single case from the data, through model training and driver ranking, insights, pad design and forecast economics, to a decision. Each step reads from the same dataset, so nothing is re-exported, re-joined or re-explained on the way.

The Basin Pulse screen adds the market layer: Permian production, rigs, DUC inventory, WTI and Henry Hub from live EIA series, next to curated operator activity from the latest quarterly disclosures, with each operator flagged green, yellow or undisclosed.

Select six tracts in the Midland Basin and the deal flow screen returns 2,351 gross acres at a blended 66,255 dollars per acre, split between remaining locations and PDP, with lateral footage, the Texas and New Mexico mix and a per-tract value on the map.

Below it, the 286 wells within three miles summarize the neighborhood: median oil per foot, lateral length, proppant and fluid loading, landing zone mix and drilling vintage. The Valuation tab then runs the productivity methods on that acreage to make the number defensible.

The Pad Planner places planned wells on the map and in a cross section at once. Each well gets a TVD, a cross-offset and a lateral length; the drainage view below shows the planned wells against the existing producers in each bench, from the Upper Spraberry down to Wolfcamp C.

Spacing is a slider. Move it and the drainage halos, the parent-child count per well and the checks at the top of the screen update, so the pad is designed against the wells that are actually there.

Tessera Quest indexes every well drilled in the United States, vertical, horizontal and deviated, with full production history: 3.9 million wells, 263 thousand of them horizontal, 149 million production records, 17 states and more than 25 basins at the time of the beta.

You ask in plain language and Quest answers with the wells, the chart and the number. It reports aggregates only, which is what makes it safe to put in front of a whole team.

The spacing module partitions drainage between wells with a parent-weighted, time-aware Voronoi diagram. For 3,534 Midland wells the P50 drainage area starts at 33 acres and falls to 29 acres by 36 months as neighbors arrive; the P10 stays at 7 acres throughout.

The same screen shows the unbound drainage ellipse, a five-well staggered gunbarrel and the lateral overlap along the wellbore, so the single number has a picture attached to it.

The Permian well explorer puts 75.5 thousand horizontal wells from 41 operators, 2.4 thousand fields and 53 counties on a single map, colored by operator. Beside it, daily oil stacks by county and by first-production vintage, and the average rate per active lease runs underneath.

The vintage chart is the one to watch: each year's wells arrive on top of a base that is declining, and the shape of that stack is the Permian treadmill in one picture.