Finding 1, from the archive

Does inspection history predict casualties?

Across 101,374 vessel-years, 2004–2021, the answer recorded here is no. A barge's recorded deficiencies before a year say nothing measurable about whether it appears in a casualty record in that year.

Deficiencies recorded before the year, against casualties in it

Prior deficiencies	Vessel-years	With a casualty	Rate	Lift
None	62,460	17,421	27.9%	1.00
1	15,833	4,456	28.1%	1.01
2-3	11,571	3,156	27.3%	0.98
4-9	8,344	2,266	27.2%	0.98
10 or more	3,166	875	27.6%	0.99

Lift is the band's rate divided by the rate across all 101,374 vessel-years (27.8%); 1.00 means the band is indistinguishable from the population. The spread across every band, from no recorded deficiencies to ten or more, is 0.04. There is no trend to read.

The same method, on a signal that is there

Prior casualties	Vessel-years	With a casualty	Rate	Lift
None	71,994	14,964	20.8%	0.75
One or more	29,380	13,210	45.0%	1.62

A factor of 2.16 on the same data. The method can separate groups. It does not separate them by deficiency history.

The biggest caveat

The cohort is casualty-enriched. 8,608 of 17,663 hulls (49%) have at least one recorded casualty, so the 27.8% base rate is a rate within this cohort and is not a casualty rate for the US barge fleet. The comparison between bands is unaffected. Both sides are what the Coast Guard wrote down, and this is a correlation over a window, not a causal claim. It says nothing about any individual vessel: Maury does not score, rate, or grade vessels.

Cohort definition. One row per vessel-year; prior deficiencies counts deficiencies recorded against the hull before that year began, and casualty means the hull appears in a Coast Guard incident investigation for that year.

Citing this

Maury (Thalor Technologies Inc). “Does inspection history predict casualties?” 101,374 vessel-years, 2004–2021. Computed from USCG PSIX and IIR records retrieved 5 September 2026. https://mauryrecords.com/signal.html

The full method — every limit, how to check it, and what it does not say
Finding 2: does a change of operator go with a casualty?
Populations: cohorts by type, build decade and length · operators of record

The population, not the hull

Does this match what you see?