The Lower-Tertiary Drilling Learning Curve

Across 12 Paleogene fields, 184 development wellbores burned 11,124 rig-days (30.5 rig-years) of drilling — a median of 46.5 days per wellbore (IQR 19–82). Each extra 1,000 ft of true vertical depth adds only about 3.3 drilling days, and TVD explains just 11% of the spread (r²=0.11): in this play field execution and hole trouble — not depth — drive the learning curve.

Wellbores (drill-days derivable)

184 / 217

Total rig-days

11,124 d

Median drill-days

46.5 (mean 60.5)

Median completion-days

33.0 (n=186)

Days per 1,000 ft TVD

3.3 r²=0.11

Median mud weight

16.8 ppg (max 20.0)

Median water depth

6,820 ft (max 9,525)

Deepest wellbore TVD

35,115 ft

Learning curve

Drilling days vs true vertical depth

Each dot is one of the 184 wellbores with both a recorded drilling duration and a TVD. The dashed line is the least-squares fit (slope 3.3 days per 1,000 ft, r²=0.11); the highlighted point is the deepest hole, North Platte 004 at 35,115 ft TVD, drilled in only 50 days — deeper is not automatically slower.

Drilling days vs max wellbore TVD, 184 GoM Lower-Tertiary wellbores. The weak fit (r²=0.11) is the point: depth sets a floor, execution sets the spread.

By field

Median drilling days by field

Field medians range from 28.0 days (Chinook) to 107.0 days (Tiber) — a 3.8× spread across the same play, at broadly similar depths.

Median drilling days per field (n = wellbores with a derivable duration). St Malo's low median reflects many short re-entry / sidetrack intervals in a 65-wellbore program.

Repetition vs step-out

What actually drives drilling days — repetition vs step-out

Ordering each field's wells by spud date and splitting them in half, the honest signal is depth-normalized: rig-days per 1,000 ft TVD (dpk), first-half median vs last-half median. That controls for a field simply moving to shallower targets. Of the 11 fields with n ≥ 4 datable wells, 5 learned (dpk fell >10%), 4 stepped out (dpk rose >10%), and 2 held flat.

Percent change in depth-normalized drilling intensity (rig-days per 1,000 ft TVD), first half → last half of each field's spud-ordered wells. Green = repetition compounding; red = step-outs into new fault blocks resetting the clock.

St Malo spotlight — why depth-normalizing matters

Across St Malo's 57 wells, raw drilling time looks like a clean win — first-half median 42 days falling to 19 days (-54% raw). But the last-half wells are ~5,078 ft shallower (TVD drift -5,078 ft), so a good chunk of that speed-up is just easier holes. Depth-normalized, the real, defensible learning is -32% — 1.6 → 1.1 rig-days per 1,000 ft.

Repetition compounds: 5 fields cut depth-normalized drilling intensity 26–67% (Chinook -67%, Shenandoah -57%, St Malo -32%, Stones -27%, North Platte -26%). Step-outs reset the clock: 4 fields worsened by +29–266% as they pushed into new fault blocks and reservoir compartments (Cascade +266%, Anchor +119%, Julia +45%, Jack +29%).

Honest note: this is depth-normalized and limited to fields with n ≥ 4 datable wells. Pooled across every well-in-sequence, the learning trend is weak (r ≈ -0.28 over 182 wells, -1.1 days per additional well) — so the thesis is not “wells always get faster,” but rather that repetition within a known block compounds while step-outs pay the first-well tax again.

Mud program

Mud weight vs depth

Maximum drilling-fluid weight rises with depth as pore-pressure ramps in the Paleogene section — median 16.8 ppg, topping out at 20.0 ppg. 5 readings outside the physical 9–22 ppg band (including one 80 ppg keying error) are excluded as data errors.

Max mud weight (ppg) vs max wellbore TVD, 175 wellbores with a valid reading.

Fastest

Quickest wellbores

Field	Well	Drill d	TVD ft	WD ft
St Malo	PS004	1	23,924	6,820
Stones	008	1	11,226	9,525
Stones	009	1	11,230	9,525
Big Foot	001	2	8,150	5,190
Big Foot	002	2	22,137	5,190

Sub-3-day entries at deep TVD are re-entry / completion-only intervals, not full wells — the duration is real but the scope is partial.

Slowest

Longest wellbores

Field	Well	Drill d	TVD ft	WD ft
Kaskida	002	364	31,375	5,860
Jack	003	226	28,238	7,240
Cascade	CA006	222	26,969	8,200
Cascade	CA003	209	26,804	8,200
St Malo	PS002	200	28,498	6,820

The slowest hole, Kaskida 002, took 364 days — 7.8× the play median.

Field table

Per-field drilling summary

Field	Wellbores	Median d	Total rig-d	Median TVD ft
Chinook	5	28.0	317	26,222
St Malo	57	30.0	2,297	28,049
Big Foot	24	36.5	1,207	22,338
Kaskida	7	37.0	556	31,375
North Platte	12	46.5	675	33,460
Anchor	13	51.0	821	33,367
Jack	8	58.0	652	28,278
Shenandoah	17	69.0	1,238	31,025
Stones	21	71.0	1,457	27,488
Cascade	9	71.0	888	26,685
Julia	9	96.0	802	30,334
Tiber	2	107.0	214	32,944