Gerrymandering Big Data: Online Rigged Election
After just a month and a half since South Korea’s 21st general elections, waves of South Korean citizens, led by Min Kyung-wook, continue together protesting the results: They shout, “The election was rigged”, and “Follow the Party”. What is going on here? And what does “Follow the Party” mean?
I am Roy Kim. I found the phrase “Follow the Party” hidden in the election data released by the South Korean National Election Commission (NEC). In my analysis of the N value I noticed that it was a distributed value and not just a random one. If it is natural data, both positive and negative numbers exist. However, what I discovered was unnatural and abnormal.
Based on the Democratic Party’s 50% of votes earned form election day voting, if the difference between the election day voting and pre-voting was less than 50%, the results were “all positive” and if it was more than 50%, the results were “all negative”. And if we combine the two sets of positive and negative numbers, we get 0. It is unusual to find a rule divided around a certain datum point. Moreover, only negative numbers exist in all 89 districts with more than 50% of the votes earned on election day. In the other 164 districts, only positive numbers exist. Again, it was not just one district but all 253 districts that had either “all positive” or “all negative” numbers with no mixture of both. This is solid evidence of manipulation.
It is similar to the concept of “Gerrymandering”: the legal process of dividing up districts. The goal is to turn favorable districts into unfavorable districts, and unfavorable districts into favorable ones. The goal is to win an election where you might normally lose one. But now we live in the era of Big Data and we use an election system in which the central server of the NEC collects and processes the data for every district. With all of this Big Data in one central location, I believe it is possible to use excess ballots from favorable districts to win in unfavorable districts.
< Based on 50% of the same-day vote earned rate, blue is the same day’s proportion, red is the graph after applying the shift value, and red is the same as the pre-result proportion graph.>
If you look at the elections statistically, the unfavorable districts with early voting were advantageously changed so the ruling party could win. The ruling party won over 50% in only 89 out of 253 districts on election day voting. In early voting however, the ruling party won over 50% of the vote in 102 of 191 districts. When you combine early voting with election day voting, the results were actually 164 total seats.
We have found other evidence that proves the 2020 South Korean general election was manipulated. In one district, more than 18,000 citizens supposedly voted, with only one available ballot box. This figure is only possible if one person votes every 4.6 seconds, for two days straight. In some places, more voters voted than the entire voting population. How can this happen?
I discovered” Follow the Party” by finding that the regularity of the values shifted to prove gerrymandering. While looking for the regularity of the complicated order of the values, I created characters by converting them to 16 bytes. This was done by obtaining a value close to the printable number 100 with 1 byte. This is just one of the common hacking contest problem-solving skills.
I tried to make words and sentences from the letter table that I generated. It was within this table that I found the following terms and phrases: “Follow the Party” (the Chinese Communist Party slogan), “Ghost” (also known as spectre to communists), “Hippo”, “Harpy”, and “Riots”. All these terms have a strong relationship to communist ideology: Augustine of Hippo’s historical view greatly influenced Karl Marx, Mao Zedong was also called “Hippo”, “Harpy” is a Greek mythical bird and a monster that has a human figure that symbolizes Marxism, and “Riots” is reminiscent of communist riots. One might say that it is possible for these words to appear randomly. But having four words that symbolize communism all together is not very random at all. Finally, to find “Follow the Party” embedded in the data, it is hard not to become suspicious.
IBM semiconductor designer Benjamin Wilkerson also released a picture of a computer-like CPU built into the counting system. “This is a supercomputer specification that’s higher than a typical laptop specification,” Benjamin Walker explained. Due to the nature of a simple counting system, such a high-performance system is not required. Contrary to the explanation made by the South Korean National Election Commission, high-tech communication devices were also found in the laptops and other devices used at the polling places. China’s new Huawei communication equipment was also used.
Video footage of abstentions without invalid or stamped votes entering the first ruling party counter (ballot box) was also reported at various polling stations. Even stiff (brand new) ballots were found as if they had just come from the print shop. Normally, voters are told to fold their ballots. But those ballots were unfolded like brand new bills, hot off the presses. The South Korean National Election Commission unconvincingly announced that it was just a simple error.
When then-National Assemblyman Min Kyung-wook announced the discovery of “Follow the Party” in the data, all hell broke loose. Even though I published my work under the pseudonym Roy Kim, I was quickly and severely doxxed online: netizens dug up my real name, as well as personal messages from over a decade ago. National Assemblyman Ha Tae-kyung, responded with a public announcement dismissing “Follow the Party” as a fake “Ghost” story. It should be noted that South Korean National Assembly Ha Tae-kyung graduated from a school in Jilin, China, with funding from the Chinese Communist Party.
We also discovered that the data used differed from the NEC data, and nobody pointed this out for nearly two months. That difference in the data means the original dataset includes the other election day votes. It is impossible to determine if the manipulated votes or blank votes were from early voting or from election day voting, even though the NEC distinguished and marked the other votes and the votes of the election day.
With the election day voting data, I have analyzed all the statistical data released in South Korea over the past month and a half. Even Professor Mebane of Michigan has analyzed and spoken about the results. I used the very same data. But no one attacked or questioned the data in Professor Mebane’s analysis for being different from the data of the NEC. Also, since the number of those other votes is not that great, it does not affect the order of generating “Follow the Party”.
This is the reality of the situation. I want to ask people: Did I hit a nerve, or do you just want to dismiss this as a ghost story? Is it also a ghost story that pre-voting in the 21st South Korean election was 10% higher than election day voting? It’s not just a coincidence, so stop attacking me with “Follow the Ghost” and recount them. There will be a lot of ghost votes in the ruling party’s pre-voting box. I promise you.