He talks about "Truncated MRCV" (where candidates are eliminated until one gets a majority of first preferences) vs "Pure MRCV" (where candidates are eliminated until there is only one left).
But in Baldwin/Nanson, these are equivalent, right? A majority winner will also always be the person who doesn't get eliminated? Nanson talks about this on page 215 https://archive.org/details/transactionsproc1719roya/page/215/mode/1up
Also he talks about the exponent being adjustable in http://royminet.org/wp-content/uploads/2019/12/ElectionSimulationShedsLight.pdf:
"MRCV (pure) – Candidates are assigned an exponentially declining score based on the level at which the Voter has ranked them. An exponent of 0.7 was determined to work well."
Dowdall weighting is similar (though the system as a whole isn't runoff-based) in that the weights go down with each ranking, but is a harmonic progression rather than a geometric progression. — Psephomancy (talk) 15:30, 11 April 2020 (UTC)