Second Pass and Refinement

What is known so far

As with DSP First Pass, much of what this section would cover is documented within the technical processes, though never gathered into a second-pass account.

Optioneering of selected options or pathways

Detailed Option Assessment describes the iterative appraisal stage directly. On a second pass, additional information is incorporated:

• Alternative configurations or variants, developed or modified for optimisation
• Option specification detail: material requirements, construction method, operational and maintenance requirements, service life, policy details

It also states the rule that governs repeat rounds: the assessment criteria, indicators and weightings remain unchanged between rounds, so that results stay comparable. See keep the framework fixed.

The stage is described as typically supporting development of a Strategic Business Case, or further optimisation and sensitivity analysis on option configurations.

Optimisation for specific objectives or decision dimensions

Portfolio Optimisation documents five distinct iteration triggers, including objective and constraint iteration, which arises when results or stakeholder input indicate that the defined objectives do not adequately represent the programme decision.

Its handling of competing objectives, through a Pareto frontier and an epsilon-constraint procedure, is the handbook's most concrete treatment of optimising against more than one dimension.

Additional modelling, evidence gathering and iteration

Integration with IMP sets out what the modelling platform contributes at each stage, and the four-step feedback loop between evidence and decision.

Asset / Network Evaluation lists six iterative activities and the proportionality test for deciding whether further data collection is worth undertaking.

What this page should cover

The distinction between a first and second pass is the thing genuinely missing. The technical processes describe iteration within a stage; this section would describe iteration of the whole cycle, and would need to answer:

• What triggers a second pass, as opposed to further iteration within a stage
• What is deliberately left until the second pass
• How much of the first pass is redone, and how much is carried forward unchanged
• Where the second pass ends

The upstream dependency is Outcome Assessment, whose promised output schema for reporting, learning, or second-pass refinement would define what this section consumes. That schema is also unwritten, so the two gaps compound.