Percentages in summary .csv file always add up to less than 100%

[chughes297]

See attached. I assume this is because we truncate all percentage values at the 4th decimal place? I found this stack overflow post that appears to have a few different possible solutions to this.

FWIW I'm glad we're winding up with less than 100% and not more than 100%.

2024-06-12_14-23-47_summary.csv

[artoonie]

Yes, and this is pretty common -- I'm not sure we want any of the "fixes," since they all "fix" the issue by creating incorrect percentages. We can either have the numbers add up to 100 exactly, or we can have the percentages be exactly correct, but not both.

The phrase "Percentages may not add up to 100 because of rounding." is common: https://www.google.com/search?hl=en&q=%22Percentages%20may%20not%20add%20up%20to%20100%20because%20of%20rounding.%22

e.g. this article on the NYT

[chughes297]

Makes sense. I lean towards just telling people that percentages may not add up to 100 because of rounding. Would be curious to get input from Rosemary/George/Tiffany/other admins on that.

[yezr]

added needs-rcvrc-clarification tag. Will review at next bi-weekly review (aka the RCTab Download™) on July 9 and update here after

[yezr]

After discussion we're going to leave this as is. No amount of showing more decimal places will get us to exactly 100%, and any method to match 100% could show a candidate with more or less support than they actually have. For the extremely determined observer who wants to know exactly what percentage a candidate has they have the vote tallies to do that in the next column over.