Shawn Briquelet
@shawn-briquelet.bsky.social
Political Junkie. No time for other stuff till we get rid of the Nazi's.
Update:
Jeffrey Epstein emailed Steven Bannon the day before Trump nominated Bill Barr to be AG with a subject line:
Do you know bill barr. CIA.
Jeffrey Epstein emailed Steven Bannon the day before Trump nominated Bill Barr to be AG with a subject line:
Do you know bill barr. CIA.
November 25, 2025 at 11:23 PM
Update:
Jeffrey Epstein emailed Steven Bannon the day before Trump nominated Bill Barr to be AG with a subject line:
Do you know bill barr. CIA.
Jeffrey Epstein emailed Steven Bannon the day before Trump nominated Bill Barr to be AG with a subject line:
Do you know bill barr. CIA.
Charlie McCarthy and Howdy Doody had a love child, and his name is Scott Bessent.
November 23, 2025 at 3:28 PM
Charlie McCarthy and Howdy Doody had a love child, and his name is Scott Bessent.
Do we know as fact that the death threats are coming from Americans? Not just on Marge. Could they be coming from the Russian government pretending they are Americans.
I don't think we should assume that these are Americans doing this.
I don't think we should assume that these are Americans doing this.
November 20, 2025 at 12:54 AM
Do we know as fact that the death threats are coming from Americans? Not just on Marge. Could they be coming from the Russian government pretending they are Americans.
I don't think we should assume that these are Americans doing this.
I don't think we should assume that these are Americans doing this.
Here's what happens when someone's cognitive dissonance collapses. They lash out.
"Please do a statistical analysis of the data in columns b and c. Each row is a data point. Tell me about a correlation between these two data sets."
That's just math, bro.
"Please do a statistical analysis of the data in columns b and c. Each row is a data point. Tell me about a correlation between these two data sets."
That's just math, bro.
November 19, 2025 at 10:42 PM
Here's what happens when someone's cognitive dissonance collapses. They lash out.
"Please do a statistical analysis of the data in columns b and c. Each row is a data point. Tell me about a correlation between these two data sets."
That's just math, bro.
"Please do a statistical analysis of the data in columns b and c. Each row is a data point. Tell me about a correlation between these two data sets."
That's just math, bro.
The next group 57% Trump with 45% Turnout
58-24
45-02
58-01
45-12
66-37
58-27
66-41
58-03
58-20
58-35
55% Trump, 48% Turnout
58-38
58-25
63-06
66-32
26-16
66-15
58-08
58-02
58-05
58-19
54% Trump, 51% Turnout
26-22
45-06
58-33
66-05
63-11
26-10
66-12
66-08
45-07
63-12
58-24
45-02
58-01
45-12
66-37
58-27
66-41
58-03
58-20
58-35
55% Trump, 48% Turnout
58-38
58-25
63-06
66-32
26-16
66-15
58-08
58-02
58-05
58-19
54% Trump, 51% Turnout
26-22
45-06
58-33
66-05
63-11
26-10
66-12
66-08
45-07
63-12
November 19, 2025 at 10:28 PM
The next group 57% Trump with 45% Turnout
58-24
45-02
58-01
45-12
66-37
58-27
66-41
58-03
58-20
58-35
55% Trump, 48% Turnout
58-38
58-25
63-06
66-32
26-16
66-15
58-08
58-02
58-05
58-19
54% Trump, 51% Turnout
26-22
45-06
58-33
66-05
63-11
26-10
66-12
66-08
45-07
63-12
58-24
45-02
58-01
45-12
66-37
58-27
66-41
58-03
58-20
58-35
55% Trump, 48% Turnout
58-38
58-25
63-06
66-32
26-16
66-15
58-08
58-02
58-05
58-19
54% Trump, 51% Turnout
26-22
45-06
58-33
66-05
63-11
26-10
66-12
66-08
45-07
63-12
Look at these precincts:
They went from Trump 37%, and 35% of the registered voters turned out.
45-18
45-09
63-07
45-10
45-14
45-19
45-17
45-13
58-44
58-18
These went 42% with 50% Turnout.
45-08
45-11
58-40
45-21
58-15
58-36
58-31
63-22
45-16
63-25
They went from Trump 37%, and 35% of the registered voters turned out.
45-18
45-09
63-07
45-10
45-14
45-19
45-17
45-13
58-44
58-18
These went 42% with 50% Turnout.
45-08
45-11
58-40
45-21
58-15
58-36
58-31
63-22
45-16
63-25
November 19, 2025 at 10:24 PM
Look at these precincts:
They went from Trump 37%, and 35% of the registered voters turned out.
45-18
45-09
63-07
45-10
45-14
45-19
45-17
45-13
58-44
58-18
These went 42% with 50% Turnout.
45-08
45-11
58-40
45-21
58-15
58-36
58-31
63-22
45-16
63-25
They went from Trump 37%, and 35% of the registered voters turned out.
45-18
45-09
63-07
45-10
45-14
45-19
45-17
45-13
58-44
58-18
These went 42% with 50% Turnout.
45-08
45-11
58-40
45-21
58-15
58-36
58-31
63-22
45-16
63-25
No dude, this analysis only covers the Wards Trump won. As shown in red.
November 19, 2025 at 10:13 PM
No dude, this analysis only covers the Wards Trump won. As shown in red.
ChatGPT says:
"Your observed correlation is far outside the realm of chance.
It lies: More than 10 standard deviations away from the null mean. Far beyond the maximum correlation seen in 20,000 random trials. Literally off the chart compared to random permutations"
"Your observed correlation is far outside the realm of chance.
It lies: More than 10 standard deviations away from the null mean. Far beyond the maximum correlation seen in 20,000 random trials. Literally off the chart compared to random permutations"
November 19, 2025 at 10:11 PM
ChatGPT says:
"Your observed correlation is far outside the realm of chance.
It lies: More than 10 standard deviations away from the null mean. Far beyond the maximum correlation seen in 20,000 random trials. Literally off the chart compared to random permutations"
"Your observed correlation is far outside the realm of chance.
It lies: More than 10 standard deviations away from the null mean. Far beyond the maximum correlation seen in 20,000 random trials. Literally off the chart compared to random permutations"
Not identical? All of them are in Philadelphia County. They voted for Trump as follows:
November 19, 2025 at 10:04 PM
Not identical? All of them are in Philadelphia County. They voted for Trump as follows:
We ran 20,000 random permutations of column C relative to column B.
This creates a null distribution where B and C have no relationship.
Then we compared your actual correlation to this null world.
This creates a null distribution where B and C have no relationship.
Then we compared your actual correlation to this null world.
November 19, 2025 at 9:58 PM
We ran 20,000 random permutations of column C relative to column B.
This creates a null distribution where B and C have no relationship.
Then we compared your actual correlation to this null world.
This creates a null distribution where B and C have no relationship.
Then we compared your actual correlation to this null world.
Again, we are only looking at the precincts that Trump won. The demographics are identical.
In those 163 precincts, why would the relationship between turnout and Trump's vote share behave like a function?
To be naturally occurring. The voters would have had to coordinate it in advance.
In those 163 precincts, why would the relationship between turnout and Trump's vote share behave like a function?
To be naturally occurring. The voters would have had to coordinate it in advance.
November 19, 2025 at 9:57 PM
Again, we are only looking at the precincts that Trump won. The demographics are identical.
In those 163 precincts, why would the relationship between turnout and Trump's vote share behave like a function?
To be naturally occurring. The voters would have had to coordinate it in advance.
In those 163 precincts, why would the relationship between turnout and Trump's vote share behave like a function?
To be naturally occurring. The voters would have had to coordinate it in advance.
If you want best fit from these options, use the cubic model:
f(b)≈18.775b3−31.144b2+17.617b−2.796
You can plug any b in [min(b), max(b)] into either of those to get an estimated c.
f(b)≈18.775b3−31.144b2+17.617b−2.796
You can plug any b in [min(b), max(b)] into either of those to get an estimated c.
November 19, 2025 at 9:32 PM
If you want best fit from these options, use the cubic model:
f(b)≈18.775b3−31.144b2+17.617b−2.796
You can plug any b in [min(b), max(b)] into either of those to get an estimated c.
f(b)≈18.775b3−31.144b2+17.617b−2.796
You can plug any b in [min(b), max(b)] into either of those to get an estimated c.
⭐ It is much less likely that this structure arises from uncontrolled real-world or behavioral processes.
⭐ The smoothness, predictability, and binning behavior are all signatures of a generated or model-driven relationship.
⭐ The smoothness, predictability, and binning behavior are all signatures of a generated or model-driven relationship.
November 19, 2025 at 9:29 PM
⭐ It is much less likely that this structure arises from uncontrolled real-world or behavioral processes.
⭐ The smoothness, predictability, and binning behavior are all signatures of a generated or model-driven relationship.
⭐ The smoothness, predictability, and binning behavior are all signatures of a generated or model-driven relationship.
It means:
✔ There is an underlying function
✔ But the noise around the function grows as B grows
This again fits a modeled or formula-based relationship.
✔ There is an underlying function
✔ But the noise around the function grows as B grows
This again fits a modeled or formula-based relationship.
November 19, 2025 at 9:28 PM
It means:
✔ There is an underlying function
✔ But the noise around the function grows as B grows
This again fits a modeled or formula-based relationship.
✔ There is an underlying function
✔ But the noise around the function grows as B grows
This again fits a modeled or formula-based relationship.
This is exactly what you would get if:
C is a smooth mathematical function of B (slightly curved, monotonic) plus random scatter.
C is a smooth mathematical function of B (slightly curved, monotonic) plus random scatter.
November 19, 2025 at 9:27 PM
This is exactly what you would get if:
C is a smooth mathematical function of B (slightly curved, monotonic) plus random scatter.
C is a smooth mathematical function of B (slightly curved, monotonic) plus random scatter.
The binned correlations are especially telling:
When you group the data into bins, the noise cancels out—and the underlying curve becomes almost perfectly monotonic and smooth
This is not typical for behavioral or social data
It is typical for data generated from a mathematical function plus noise
When you group the data into bins, the noise cancels out—and the underlying curve becomes almost perfectly monotonic and smooth
This is not typical for behavioral or social data
It is typical for data generated from a mathematical function plus noise
November 19, 2025 at 9:24 PM
The binned correlations are especially telling:
When you group the data into bins, the noise cancels out—and the underlying curve becomes almost perfectly monotonic and smooth
This is not typical for behavioral or social data
It is typical for data generated from a mathematical function plus noise
When you group the data into bins, the noise cancels out—and the underlying curve becomes almost perfectly monotonic and smooth
This is not typical for behavioral or social data
It is typical for data generated from a mathematical function plus noise
Summary of Findings
We tested whether the relationship between columns B and C is:
Linear
Quadratic (curved)
Cubic (more flexible curvature)
Based on model fits and R² values, the relationship is not perfectly linear, but also not strongly nonlinear — it shows only a mild upward curvature.
We tested whether the relationship between columns B and C is:
Linear
Quadratic (curved)
Cubic (more flexible curvature)
Based on model fits and R² values, the relationship is not perfectly linear, but also not strongly nonlinear — it shows only a mild upward curvature.
November 19, 2025 at 9:21 PM
Summary of Findings
We tested whether the relationship between columns B and C is:
Linear
Quadratic (curved)
Cubic (more flexible curvature)
Based on model fits and R² values, the relationship is not perfectly linear, but also not strongly nonlinear — it shows only a mild upward curvature.
We tested whether the relationship between columns B and C is:
Linear
Quadratic (curved)
Cubic (more flexible curvature)
Based on model fits and R² values, the relationship is not perfectly linear, but also not strongly nonlinear — it shows only a mild upward curvature.
Even at the coarsest level (5 bins), where noise in binning is highest, the correlation remains highly unusual (p ≈ 0.012).
The relationship between columns B and C is strong, smooth, and very unlikely to be due to random variation, whether you look at raw points or average them into groups.
The relationship between columns B and C is strong, smooth, and very unlikely to be due to random variation, whether you look at raw points or average them into groups.
November 19, 2025 at 9:20 PM
Even at the coarsest level (5 bins), where noise in binning is highest, the correlation remains highly unusual (p ≈ 0.012).
The relationship between columns B and C is strong, smooth, and very unlikely to be due to random variation, whether you look at raw points or average them into groups.
The relationship between columns B and C is strong, smooth, and very unlikely to be due to random variation, whether you look at raw points or average them into groups.
10-bin and 20-bin panels
Null distributions look roughly bell-shaped, centered near 0.
Your observed r ≈ 0.94 and 0.91 sit far beyond anything the null generates.
Null distributions look roughly bell-shaped, centered near 0.
Your observed r ≈ 0.94 and 0.91 sit far beyond anything the null generates.
November 19, 2025 at 9:19 PM
10-bin and 20-bin panels
Null distributions look roughly bell-shaped, centered near 0.
Your observed r ≈ 0.94 and 0.91 sit far beyond anything the null generates.
Null distributions look roughly bell-shaped, centered near 0.
Your observed r ≈ 0.94 and 0.91 sit far beyond anything the null generates.
10-bin correlation:
p_bin10 ≈ 0.0000
20-bin correlation:
p_bin20 ≈ 0.0000
p_bin10 ≈ 0.0000
20-bin correlation:
p_bin20 ≈ 0.0000
November 19, 2025 at 9:17 PM
10-bin correlation:
p_bin10 ≈ 0.0000
20-bin correlation:
p_bin20 ≈ 0.0000
p_bin10 ≈ 0.0000
20-bin correlation:
p_bin20 ≈ 0.0000
To test this, I:
Kept column B fixed.
Randomly shuffled C (breaking any structure) 10,000 times.
For each shuffle, I recomputed:
Raw correlation
5-bin correlation
10-bin correlation
20-bin correlation
This gives four null distributions—what each statistic looks like if B and C are unrelated.
Kept column B fixed.
Randomly shuffled C (breaking any structure) 10,000 times.
For each shuffle, I recomputed:
Raw correlation
5-bin correlation
10-bin correlation
20-bin correlation
This gives four null distributions—what each statistic looks like if B and C are unrelated.
November 19, 2025 at 9:17 PM
To test this, I:
Kept column B fixed.
Randomly shuffled C (breaking any structure) 10,000 times.
For each shuffle, I recomputed:
Raw correlation
5-bin correlation
10-bin correlation
20-bin correlation
This gives four null distributions—what each statistic looks like if B and C are unrelated.
Kept column B fixed.
Randomly shuffled C (breaking any structure) 10,000 times.
For each shuffle, I recomputed:
Raw correlation
5-bin correlation
10-bin correlation
20-bin correlation
This gives four null distributions—what each statistic looks like if B and C are unrelated.
So: as we aggregate data into bins, the relationship between B and C becomes even tighter (almost perfectly increasing).
November 19, 2025 at 9:15 PM
So: as we aggregate data into bins, the relationship between B and C becomes even tighter (almost perfectly increasing).
Here’s the multi-level Monte Carlo result for your data in columns B and C.
November 19, 2025 at 9:14 PM
Here’s the multi-level Monte Carlo result for your data in columns B and C.
✔ Confirms previous correlation results
This analysis supports the high Pearson and Spearman correlations you already observed.
This analysis supports the high Pearson and Spearman correlations you already observed.
November 19, 2025 at 9:14 PM
✔ Confirms previous correlation results
This analysis supports the high Pearson and Spearman correlations you already observed.
This analysis supports the high Pearson and Spearman correlations you already observed.
Bin Counts
These show how many points fall in each bin:
Observations:
The densest bins are in the middle (consistent with a unimodal distribution for B).
The highest bin contains only 1 point, so that final point should be interpreted cautiously (but it still fits the trend).
These show how many points fall in each bin:
Observations:
The densest bins are in the middle (consistent with a unimodal distribution for B).
The highest bin contains only 1 point, so that final point should be interpreted cautiously (but it still fits the trend).
November 19, 2025 at 9:13 PM
Bin Counts
These show how many points fall in each bin:
Observations:
The densest bins are in the middle (consistent with a unimodal distribution for B).
The highest bin contains only 1 point, so that final point should be interpreted cautiously (but it still fits the trend).
These show how many points fall in each bin:
Observations:
The densest bins are in the middle (consistent with a unimodal distribution for B).
The highest bin contains only 1 point, so that final point should be interpreted cautiously (but it still fits the trend).