The follow-up explains that the original question was read as covering all available pre-dreadnoughts and dreadnoughts, including ships still owned by the Royal Navy outside active commission. The count runs through older classes as well as newer battleships and battlecruisers. Timing matters: Bulwark is included if the scenario starts before her loss in November 1914. Ships under construction and orders cancelled around August also affect the total. The running arithmetic changes during the discussion, and roughly ninety is repeatedly identified as a rough estimate. It should therefore be understood as the scale assumed for the counterfactual, rather than a verified order of battle.
Notable Point · 4:22:45-4:30:25
Why the Iowa counterfactual uses an approximate British fleet count
The estimate of roughly ninety replacement Iowas depends on counting older retained ships and ships under construction at the chosen point in 1914.