Three questions people argue about, answered with arithmetic instead of opinion. Everything below is computed from public reference data: HMRC, the Bank of England, legislation.gov.uk and fund managers' own published reports. Nothing here comes from anybody's uploaded transactions.
HMRC's manual says the rate is the one in force on the day, then says HMRC does not prescribe which reference point to use. Eleven years of HMRC monthly rates against Bank of England daily spot rates, and what the gap is worth in pounds.
Offshore fund income nobody paid you, counted in aggregate: how often the published figure is nil, how often a nil fund stops being nil, and why nine records in ten need converting before they can go on a return.
Six months after the period end, says reg 94(4). Across every published record we hold, that crosses the 5 April line more often than not, and the fund has until that same day to publish the number.
Every number on these pages is computed from the source data rather than typed into the prose, and every page has a section saying what it cannot tell you. Where a figure could not be sourced it is named as unavailable, with the reason. There are no fund prices in this project and no market feed is used for anything, so no figure here is a percentage of a price.
These are analyses rather than a dataset, so there is nothing to download: aggregates, a handful of named examples, and the method, so you can redo the work from the same public sources. Not one figure comes from a calculation anybody ran here. Uploaded transactions are processed in memory and discarded, signed in or not.
More on the rules themselves in the guide, fund-by-fund figures in the ERI lookup, and the tools in the calculators.