Critique of Lindeman, Stark, Yates’ 'BRAVO: Ballot polling Risk-limiting Audits to Verify Outcomes' and of Ottoboni, Bernhard, Halderman, Rivest & Stark’s 'Bernoulli Ballot Polling: A Manifest Improvement for Risk-Limiting Audits'
5 Pages Posted: 23 Jun 2019
Date Written: June 8, 2019
Lindeman, Stark & Yates’ BRAVO (ballot polling audits to verify outcomes) method and Ottoboni, Bernhard, Halderman, Rivest & Stark’s BBP (Bernoulli ballot polling) method claim to be effective ways to audit election results that limit the risk of certifying incorrect election outcomes to a desired low rate alpha, α Both methods are a takeoff of Wald’s Sequential Probability Ratio Test (SPRT). Unfortunately, neither of these methods of conducting post-election audits limit the risk of certifying incorrect election outcomes to their stated limit.
Keywords: BRAVO ballot polling, Bernoulli ballot polling, Lindeman, Stark, election audit, risk limiting
Suggested Citation: Suggested Citation