5 ms·

I'm writing this in response to the originator of this thread @beedeebeedee[1]. I'm posting it as a top-level comment because I believe it's applicable to the o

I'm writing this in response to the originator of this thread @beedeebeedee[1]. I'm posting it as a top-level comment because I believe it's applicable to the overall conversation. > According to the theory, in fair elections, turnout indicators typically follow a regular graphical representation that resembles a bell shape. If anomalies appear in the data -- for example, forms different from the bell shape or the bell "grows a tail" -- this indicates unfair elections. There is *no* respectable statistician who would *ever* draw a conclusion from data. You could say 'may indicate evidence of unfair elections,' but it is impossible to deem whether something occurred or not with 100% certainty. Now to address their "indicators." > While in cities the data representation mostly follows a bell shape, in the regions it is anomalously deviated All statistics is just a representation of data, it tells us nothing of logic. Logic is also based in axioms and it's important to identify what those are. Here the author is saying the data is "anomalously deviated" with the axiomatic assumption that all voters must follow the same distribution, but why should this assumption be held true? It ignores other logical factors that may sway these distributions. If I'm a candidate for a political position that would give a rural precinct free money, wouldn't that heavily affect that population's voting outcomes? Surely what's said in a political campaign can sway these things, yet there's nothing in the article that mentions that. > According to the theory, if a number of stations emerge where one candidate -- the "beneficiary" -- is unusually high, this might indicate falsification. There's a high risk of manipulation where increased overall voter turnout is reflected only in support of the "beneficiary." So the belief is that because there are 2 distributions for the number of precincts with a certain percentage of "yes" votes, that automatically "indicates falsification"? There needs to be more information than this to make a conclusion. What one holds as axioms are a matter of opinion. The only way to get something useful out of a counterargument to this is for a proponent to exhaustively list all the votes they think are legitimate, and for the opponent to show the same patterns can occur there as well. Similarly for a proponent to convince detractors, they'd have to show that other elections they see as legitimate *never* indicate those patterns. Otherwise this is just an opinion piece on data. [1] https://news.ycombinator.com/item?id=42996898 https://news.ycombinator.com/item?id=42996898

beedeebeedee 2y ago
Hi derangedHorse, thanks for your insight. The anomaly is that there appears to be a heavy shift in votes only in (1) early voting, and not in (2) mail-in ballots or (3) Election Day voting. Additionally, the anomaly only appears for all tabulators after 300 or more votes are cast. Here's the image that represents the anomaly in early voting: https://img1.wsimg.com/isteam/ip/9087f51c-d3bd-4002-9943-79706c6e82a3/somethingiswrong2024-04%20(3)-16ff714.png/:/cr=t:0%25,l:0%25,w:100%25,h:100%25/rs=w:1209,cg:true https://img1.wsimg.com/isteam/ip/9087f51c-d3bd-4002-9943-797... > There is no respectable statistician who would ever draw a conclusion from data. You could say 'may indicate evidence of unfair elections,' but it is impossible to deem whether something occurred or not with 100% certainty. The originators of the analysis are claiming that this indicates manipulation. Perhaps they should have given more qualifiers in that image of their analysis, but the data is alarming and it is an up-hill battle even to get their analysis noticed (other than the press release to WCIA, only Newsweek has mentioned it). We can say 'maybe' but given the current political and media environment right now, it is likely that that qualification would have doomed their analysis from getting any coverage at all. Here is a video from the group that discusses these issues and is worth watching because it speaks directly to your concerns and others: https://www.youtube.com/watch?v=WOQ-GxJyJN4 https://www.youtube.com/watch?v=WOQ-GxJyJN4 Edit: I haven't fully absorbed their analysis, so I've gone back and they do make those qualifications to the data ("Data Suggestive of Vote Manipulation", etc) within the PDF: https://img1.wsimg.com/blobby/go/9087f51c-d3bd-4002-9943-79706c6e82a3/Election%20Truth%20Alliance_Clark%20County%20N-a246330.pdf https://img1.wsimg.com/blobby/go/9087f51c-d3bd-4002-9943-797... "The pattern above shows an inexplicable spike in vote distribution that is statistically unlikely based on typical human voting behavior." And while discussing alternate causes ("Deep Red" Areas) "While such a deviation in the data could potentially still emerge, this does not explain why the pattern is limited to Early Voting, as in that scenario it would be reasonable to expect the same deviation to appear in Election Day results. Instead, Election Day results indicate a normal, expected variation."
- derangedHorse 2y ago
  6:11 I don't find this particularly strange. I think they'd make a much stronger argument if they showed the drop-off vote for 2020. I found it strange that they only picked 2012 and 2016 given my theory that drop-off votes are probably more prominent in controversial elections. Especially given the fact Kamala entered the race somewhat last minute. I also find it odd that there are no links to explore the previous years that were shown in the video from the website linked in the video's description [1]. 10:10 Another claim of irregularity that doesn't quite support their conclusion. It seems they're indicating that both plots seem irregular, which I agree, but there thesis of fraud doesn't completely address 2020's case. They could say republican cheated in 2020 as well, but then they'd have to explain why this election went any different. 11:36 Here they're trying to support the irregularity of the data with anecdote. "Does this match from your opinion with the actual demographic there?" is not a methodical approach to figuring out what's wrong. Surely not every republican would be okay with fraudulent votes regardless of who won. A contrary voice in these video would lend some credence to their claims. 13:28 Voting against one's registered party also doesn't seem crazy to me as I've had many friends who have this past election. My own anecdote doesn't affect the data and may not sway arguments, nor does my theory of controversial elections breeding "irregular" results, but hopefully it can create an equally valid opinion than the one presented in the video. 15:05 If someone was rigging the election in what way? Any security model being claimed as breached needs to define what constitutes a break and how the attack could create that break in security. 16:30 I don't understand why a '2014 Wisonsin Gov. Race' (the title of the graph) is applicable here. Maybe she just used it as a lead-in for the next part? 18:37 Why is she bringing up the random distribution of inadvertent errors? After spending about an hour watching the video and observing the results, I'm even more skeptical of their claims. [1] https://smartelections.us/dropoff https://smartelections.us/dropoff

beedeebeedee 2y ago

Hi derangedHorse, thanks for your insight. The anomaly is that there appears to be a heavy shift in votes only in (1) early voting, and not in (2) mail-in ballots or (3) Election Day voting. Additionally, the anomaly only appears for all tabulators after 300 or more votes are cast. Here's the image that represents the anomaly in early voting: https://img1.wsimg.com/isteam/ip/9087f51c-d3bd-4002-9943-79706c6e82a3/somethingiswrong2024-04%20(3)-16ff714.png/:/cr=t:0%25,l:0%25,w:100%25,h:100%25/rs=w:1209,cg:true https://img1.wsimg.com/isteam/ip/9087f51c-d3bd-4002-9943-797... > There is no respectable statistician who would ever draw a conclusion from data. You could say 'may indicate evidence of unfair elections,' but it is impossible to deem whether something occurred or not with 100% certainty. The originators of the analysis are claiming that this indicates manipulation. Perhaps they should have given more qualifiers in that image of their analysis, but the data is alarming and it is an up-hill battle even to get their analysis noticed (other than the press release to WCIA, only Newsweek has mentioned it). We can say 'maybe' but given the current political and media environment right now, it is likely that that qualification would have doomed their analysis from getting any coverage at all. Here is a video from the group that discusses these issues and is worth watching because it speaks directly to your concerns and others: https://www.youtube.com/watch?v=WOQ-GxJyJN4 https://www.youtube.com/watch?v=WOQ-GxJyJN4 Edit: I haven't fully absorbed their analysis, so I've gone back and they do make those qualifications to the data ("Data Suggestive of Vote Manipulation", etc) within the PDF: https://img1.wsimg.com/blobby/go/9087f51c-d3bd-4002-9943-79706c6e82a3/Election%20Truth%20Alliance_Clark%20County%20N-a246330.pdf https://img1.wsimg.com/blobby/go/9087f51c-d3bd-4002-9943-797... "The pattern above shows an inexplicable spike in vote distribution that is statistically unlikely based on typical human voting behavior." And while discussing alternate causes ("Deep Red" Areas) "While such a deviation in the data could potentially still emerge, this does not explain why the pattern is limited to Early Voting, as in that scenario it would be reasonable to expect the same deviation to appear in Election Day results. Instead, Election Day results indicate a normal, expected variation."

derangedHorse 2y ago
6:11 I don't find this particularly strange. I think they'd make a much stronger argument if they showed the drop-off vote for 2020. I found it strange that they only picked 2012 and 2016 given my theory that drop-off votes are probably more prominent in controversial elections. Especially given the fact Kamala entered the race somewhat last minute. I also find it odd that there are no links to explore the previous years that were shown in the video from the website linked in the video's description [1]. 10:10 Another claim of irregularity that doesn't quite support their conclusion. It seems they're indicating that both plots seem irregular, which I agree, but there thesis of fraud doesn't completely address 2020's case. They could say republican cheated in 2020 as well, but then they'd have to explain why this election went any different. 11:36 Here they're trying to support the irregularity of the data with anecdote. "Does this match from your opinion with the actual demographic there?" is not a methodical approach to figuring out what's wrong. Surely not every republican would be okay with fraudulent votes regardless of who won. A contrary voice in these video would lend some credence to their claims. 13:28 Voting against one's registered party also doesn't seem crazy to me as I've had many friends who have this past election. My own anecdote doesn't affect the data and may not sway arguments, nor does my theory of controversial elections breeding "irregular" results, but hopefully it can create an equally valid opinion than the one presented in the video. 15:05 If someone was rigging the election in what way? Any security model being claimed as breached needs to define what constitutes a break and how the attack could create that break in security. 16:30 I don't understand why a '2014 Wisonsin Gov. Race' (the title of the graph) is applicable here. Maybe she just used it as a lead-in for the next part? 18:37 Why is she bringing up the random distribution of inadvertent errors? After spending about an hour watching the video and observing the results, I'm even more skeptical of their claims. [1] https://smartelections.us/dropoff https://smartelections.us/dropoff