The election was not counted, it was calculated, and we were on cruise control.”

— Draza Smith summarizes her findings, in interview with David Clements

Draza Smith is a computer engineer with two master’s degrees, in electrical engineering and in computer engineering. She completed her doctoral work in computer engineering, and awaits defense of her PhD dissertation. She has worked for Sandia National Labs focusing on Cyber and Grid Security.

She explored the 2020 election count data from The New York Times (supplied mostly by Edison Research, with additions by The Associated Press) which shows the proportion of ballots voting for each candidate as the results came in, batch by batch. Her investigations indicate that results in many states may potentially be being manipulated or controlled via a computer algorithm.

She believes the algorithm has been programmed to achieve a specific percentage of votes in each location, adjusting the votes up or down during the count in order to reach that target. She likens this algorithm to that of a “PID Controller”, or a car’s cruise control, where, after a specific target speed has been set, the car’s engine adjusts up or down to reach that target. She identified several cases where the target was adjusted mid-way through the ballot counting process (potentially based on the results in other states, such as after Florida was announced as a win for Trump) and it appears that all batches following this adjustment show a shift in the results as the algorithm now aims for the new target value.

Since most election systems in the US are “closed source” — meaning that their source code cannot be easily audited — it’s not clear exactly how this manipulation is occurring, but it deserves deeper investigation, and is yet another example of why closed-source election software run by private companies should not be used in elections.

Draza posts her analyses, charts and graphs on her Telegram Channel.

Articles & Reports

Video Presentations

Raw Data

You can view the ballot counting time-series graphs as well as access the raw time-series data used for many of Draza’s statistical analyses in our Data Explorer.

Visitor Comments

Show Comments