Two-child limit full roll-out page says the exceptions are unmodelled and approximates the FYE 2011 line #2708

docs/book/validation/two-child-limit-full-rollout.md (reproducing fully rolled-out two-child limit estimates, such as the IFS Green Budget 2025 comment) no longer matches the model on main. Both points predate #2685; an independent review of that PR found them out of scope (findings 3 and 4).

1. Exceptions. Lines 54–58 and 170 say the model does not apply the statutory exceptions to the limit (multiple births, adoption, non-parental care, non-consensual conception). From 2.161.0 it does, for Universal Credit and Child Tax Credit (uc_individual_child_element, uc_child_limit_exception, uc_child_limit_disregarded, is_CTC_child_limit_exception), as person-level inputs that default to false. The page should distinguish what the model supports from what the dataset supplies.
2. FYE 2011 line. Lines 46–52 and the snippet (lines 77–86) approximate the FYE 2011 absolute line by overriding household.poverty.absolute_poverty_threshold_bhc/_ahc with £305.70/£261.92, described as the FYE 2021 values in the parameter file. The parameter file on main has £301.34 and £259.01 there (absolute_poverty_threshold_bhc.yaml:26, _ahc.yaml:26), and 2.155.0 added in_absolute_poverty_fye2011_bhc/_ahc (and poverty_threshold_fye2011_*), so the approximation is no longer needed.

The "Results, 2025-26" section was produced by the old snippet on 2.147.2. It should be re-run on a version that has both changes, or labelled explicitly as 2.147.2 output.