· Correction to the notes of 12 August
PHAROS, the headline figure. The first note speaks of the barrier's weak crossing. The figure the engine reports as its headline is the fifth percentile of detection probability over a sampled set of crossings. The single worst sampled crossing is reported next to it. They are two different numbers.
PHAROS, the comparison. The note says that most planning tools take the union of the radar and the sonar footprints. We have no survey to support a statement about most tools, and withdraw it. Scoring on the intersection is the criterion of our model, chosen for its operational objective.
PHAROS, maintenance. Routing a boat before a gap opens is a direction of development. Availability objectives and the barrier over a sea-state forecast are implemented in the engine.
TerraScan, the example. The score of 71, the 44 points and the pass eleven days earlier are a synthetic example written for the note. They are not taken from a real territory.
The three labels. The note says that every operational figure in our products is labelled measured, computed or estimated. That is implemented in KraftSync Transit. For the other products, what is labelled and how is stated on each product page. The figures of 46 and 12 per cent are an illustration.
All five points are now corrected in the notes themselves: terrain masking, composite scores, value lineage.