{% extends "base.html" %} {% block title %}{{ contest }}: Audit {% if contest.election.random_seed %}Selection {% endif %}Report | {{ block.super }}{% endblock %} {% block content %} {% with contest.election.random_seed as random_seed %}

Audit {% if random_seed %}Selection {% endif %}Report for {{ contest }}

Proportion of the total number of votes for this contest that were cast within this county: {{ contest.proportion|floatformat:1 }}%

See additional statistical information below...

{% if random_seed %}

The audit units are reported for selection in priority order, based on the thresholds below, and the "Sum of Square Roots" pseudorandom number generator using the random seed value {{ random_seed }} combined with the Batch sequence numbers.

Select the top {{ stats.negexp_precincts|floatformat:0 }} audit units.

{% endif %} {% endwith %} {% for contestbatch in contest.contestbatch_set.all %} {% if forloop.first %} {% for votecount in contestbatch.votecount_set.all|dictsort:"choice.name"%} {% endfor %} {% for votecount in contestbatch.votecount_set.all|dictsort:"choice.name"%} {% endfor %} {% endif %} {% endfor %} {% for contestbatch in contest_batches %} {% with contestbatch.batch as batch %} {% for votecount in contestbatch.votecount_set.all|dictsort:"choice.name"%} {% endfor %} {% endwith %} {% endfor %}
Batch Seq Threshold Random Priority Type Batches Ballots Contest Ballots{{ votecount.choice }}
Totals {{ votecount.choice.votes }}
{{ batch.id|stringformat:"06d" }} {{ contestbatch.threshhold|stringformat:"05f" }} {{ contestbatch.ssr|stringformat:"05f" }} {{ contestbatch.priority|stringformat:"05f" }} {{ batch.type }} {{ batch.name }} {{ batch.ballots }} {{ contestbatch.contest_ballots }}{{ votecount.votes }}

Audit statistics

The "Number of audit units to audit" is based on the NEGEXP method, which is very efficient and requires selecting larger audit units with higher probability than smaller ones. The numbers given here are based on a confidence level of {{ contest.confidence }}%. I.e. they are designed so the audit will either 1) find a discrepancy and call for an escalation or full hand recount, or 2) reduce the risk of confirming an incorrect outcome to (100 - {{ contest.confidence }})%, even if the tally system has been manipulated. A maximum "within-precinct-miscount" of {{ stats.wpm|floatformat:0 }}% is assumed. See On Auditing Elections When Precincts Have Different Sizes, by Javed A. Aslam, Raluca A. Popa and Ronald L. Rivest

{{ stats.prose }}
{% endblock %}