In 2020, as in 2016, the Trafalgar Group provided accurate presidential election polls for the swing states. However, despite remarkably consistent precision, Trafalgar’s win predictions for the swing states were, on average, incorrect. At the same time, the conventional pollsters, even though far less precise, were correct as to their predictions for the winner in the swing states. This seems on the surface to be rather anomalous: why would the more precise pollster get the wrong overall answer?
One answer to this question could be that fraud conducted during the election occurred on such a large scale that the resulting distortion of the vote caused otherwise very accurate polling to appear to be wrong, when it really was not. There is already plenty of evidence that substantial election fraud occurred, especially in the swing states. [1,2] In particular, the numerical analyses of voting patterns are quite disturbing. [3,4,5] The purpose of this article is to show that Trafalgar polling was better in the swing states Trump won, and significantly worse in swing states Trump lost. Similarly, the reverse was true for the conventional pollsters. Both of these circumstances support the hypothesis of election fraud, since this sort of polling dichotomy would not be expected for pollsters who routinely produce consistent results. Election fraud could explain this unanticipated difference, because it would cause the “actual” vote to depart significantly from accurate pre-election polls.
The metric employed in this analysis is the margin between the two candidates, calculated as (Trump percentage) – (Biden percentage). A positive number therefore indicates a Trump advantage, and a negative number indicates a Biden advantage. The data used were taken from polls reported during the last week before the election, reflecting the assumptions that polling data collected in the last week of the election would most accurately predict voter preferences, and that a week’s worth of polls would provide enough data to calculate reasonable averages. The conventional or main stream pollsters were represented by those with an A rating (A+, A, or A-) as determined by FiveThirtyEight. 
The Trafalgar swing state poll average shows that Trump was predicted to win the 10 swing states listed in the table by an average overall margin of about 1 ½ points. Note that Trafalgar actually predicted that Trump would lose 3 of those states: Minnesota, Nevada, and Wisconsin.
Table 1: Trafalgar Final Week Swing State Poll Results 
By contrast, the A-rated pollsters predicted that Trump would lose these 10 swing states by an average overall margin of about 4 points, and that Trump would lose all 10 states.
Table 2: A-rated Polls – Final Week Swing State Results 
|Mon-mouth||ABC/ Wash||NYT/ Siena||Emer- son||Survey USA|
The next table compares the Trafalgar (TRF) and A-rated pollsters’ final week swing state projections of the candidate margins to the actual margins, in order to find the average error for both sets of polls. The actual margins were calculated from the data at Decision Desk HQ, accessed on 25 November 2020.
Table 3: Final Week Poll Results Compared to “Actual” Vote Margins
Margins (Table 1)
Margins (Table 2)
This table shows that Trafalgar overstated Trump’s performance by an average of 1 ¾ points, while the A-rated pollsters overstated Biden’s performance by an average of just less than 4 points. For these swing states, Trafalgar polling was better than twice as precise as was that of the A-Rated pollsters, while failing to predict the overall winner.
With the relative accuracy of Trafalgar vs conventional polls in 2020 established, it is then important to confirm Trafalgar’s consistency between 2016 and 2020. The following table compares Trafalgar’s performance in 2016 and 2020 (for the swing states for which I could find 2016 data).
Table 4: Trafalgar Swing State Polling Performance: 2016 vs. 2020
This table shows that Trafalgar’s accuracy was admirably uniform between the 2016 and 2020 Presidential elections. In both elections, Trafalgar overstated the the Trump margin by about 1 ½ points overall for these swing states. And that begs the obvious question: why did Trafalgar, with almost the same level of accuracy as previously demonstrated in 2016, fail to reliably pick the winner in the swing states in 2020? The answer has to do with the “actual” candidate differences. The actual difference between the two candidates in these swing states was about 2 ½ points overall in 2016, a difference for which a precision level of 1 ½ points was more than sufficient to produce an overall prediction matching the actual election results. However, in 2020, the “actual” results were much closer, less than 1 point on average for this particular set of swing states (0.87). Therefore, in 2020, a precision level of 1 ½ points was not sufficient to support reliable win predictions.
But if in fact the vote was deliberately distorted, and therefore did not match the true voter distribution, shouldn’t Trafalgar’s accuracy have experienced a greater decrease than 0.15 points (1.36 – 1.51) in the 2020 swing states relative to 2016? To investigate this question, another look at the 2020 swing state results is in order. If the assumption is made that the swing states which Trump lost were the likely states in which election fraud occurred, then the expectation would be that Trafalgar’s accuracy would decline in the states Trump lost, because the vote was distorted, while improving in the states which Trump won, because the vote was not distorted in those states. At the same time, the conventional polls, overstated as usual in the Democrat’s favor, should show improved accuracy in the states which Biden won, because the distortion in the actual vote in favor of Biden decreased the difference between their poll averages and the recorded vote. The following table presents the polling error data from Table 3 but this time split between the states which Trump lost as opposed to those he won.
Table 5: 2020 Comparison of Pollster Errors: States Trump Won vs States Trump Lost
Margin (Table 3, Column 2)
Error/ Trump Loss
(Table 3, Column 3)
(Table 3, Column 3)
Error/ Trump Loss
(Table 3, Column 6)
(Table 3, Column 6)
This table shows exactly what would be expected if both Trafalgar and conventional polling expertise remained consistent between 2016 and 2020, but votes were manipulated in favor of Biden in the states which Trump lost. Trafalgar accuracy increased to just about a single point in the states which Trump won, while it doubled from 1.5 to 3 points in the states which Trump lost. The reverse pattern occured with the A-rated pollsters. In the states which Trump won, their accuracy became even worse than it was overall, increasing from 4 to almost 7 points. Yet in the states which Trump lost, those states in which election fraud probably occurred, the A-rated pollsters improved their accuracy to below 3 points, and in these states their accuracy was actually better than Trafalgar’s. These are striking results: the error averages should be roughly the same no matter which group of states is selected, and they are not.
But it all makes sense if large scale election fraud occurred in those key swing states where Trump lost. That would explain why Trafalgar’s average error doubled in those states, and the conventional poll errors suddenly and miraculously improved in those same states, but not in the states in which Trump won. It was not that Trafalgar’s precision was any worse, or that the conventional pollsters were any better: rather, the other side of the equation, the distorted vote, did not reflect voter preferences which were quite accurately measured by Trafalgar.
1. Sheffield, Carrie. “FEC Chairman – Trump campaign bringing legitimate accusations of election fraud to court.” Just The News. November 23, 2020. https://justthenews.com/politics-policy/elections/federal-election-commission-chairman-trump-campaign-bringing-legitimate
2. Trusdell, Brian and Mack, Eric. “Detroit Elections Worker Says She Was Instructed to Change Ballot Dates.” Newsmax. November 9, 2020. https://www.newsmax.com/us/detroit-affidavit-vote-fraud/2020/11/09/id/996212
3. Hounsell, Scott “Excuse Me While I Call BS – In Wisconsin.” RedState. November 9th, 2020. https://redstate.com/scotthounsell/2020/11/09/excuse-me-while-i-call-bs-in-wisconsin-n276884
4. Mangiaracina, Emily. “Analysis shows Biden got statistically impossible percentage of votes in Georgia.” Lifesite News. November 25, 2020. https://www.lifesitenews.com/news/analysis-shows-biden-got-statistically-impossible-percentage-of-votes-in-georgia
5. Hoft, Joe. “WE CAUGHT THEM Part – SAME IMPOSSIBLE BALLOT RATIO Found in Milwaukee.” The Gateway Pundit. November 15, 2020. https://www.thegatewaypundit.com/2020/11/caught-part-3-impossible-ballot-ratio-found-milwaukee-results-change-wisconsin-election-30000-votes-switched-president-trump-biden
6. Silver, Nate. “FiveThirtyEight’s Pollster Ratings.” FiveThirtyEight. Updated May 19, 2020. https://projects.fivethirtyeight.com/pollster-ratings/
7. The Trafalgar Group. https://thetrafalgargroup.org
8. “Latest 2020 General Election Polls.” RealClear Politics. Accessed November 15, 2020. https://www.realclearpolitics.com/epolls/latest_polls/general_election
9. “2020 General Elections.” Decision Desk HQ. Accessed at November 25, 2020. https://results.decisiondeskhq.com/
10. “Election Results 2016 By States and County.” Politico. November 8, 2016. https://www.politico.com/story/2016/10/election-results-2016-by-state-and-country-229735h
11. “Florida Presidential Election Survey.” The Trafalgar Group. November 6, 2016. https://www.realclearpolitics.com/docs/2016/Trafalgar_FL_November_2016.pdf
12. “Georgia Presidential Election Survey.” The Trafalgar Group. November 6, 2016. https://www.realclearpolitics.com/docs/2016/Trafalgar_Georgia_November.pdf
13. Munro, Neil. “Pollster Says Donald Trump Ahead in Pennsylvania, Michigan.” Breitbart. Nov 7, 2016. https://www.breitbart.com/politics/2016/11/07/poll-trump-michigan-pennsylvania
14. “2016 Presidential Election Polls, The Trafalgar Group, Nevada.” 270towin. Accessed 25 November 2020. https://www.270towin.com/2016-polls-clinton-trump
15. Cahaly, Robert. “New NC Poll Showing Trump Lead 4.97%.” The Trafalgar Group. November 2, 2016. https://www.fitsnews.com/wp-content/uploads/2016/11/NC-Gen-Prez-Poll-Release-and-Report.pdf
16. “Ohio Presidential Election Survey.” The Trafalgar Group. October 27, 2016. https://www.realclearpolitics.com/docs/2016/Trafalgar_Ohio_October_27th_2016.pdf