This post is tearing me in half, forcing me to decide if my dislike of ICE is greater than my dislike of the baserate fallacy.
EXCLUSIVE | The number of migrants who claim ICE is detaining them unlawfully has increased nearly 10,000 percent in the first year of Trump’s second term.
Staggering New Figures Blow Up ICE Barbie’s Biggest Claim
The Trump administration says it’s detaining and deporting the “worst of the worst.”
trib.al
February 11, 2026 at 7:35 PM
This post is tearing me in half, forcing me to decide if my dislike of ICE is greater than my dislike of the baserate fallacy.
As someone quick to point out baserate fallacies, looks like the oil price jump is not one.
My usual rebuttal to “biggest percent change!” is “yeah, but what abt the absolute change?”
Which here? Also the biggest of all time.
My usual rebuttal to “biggest percent change!” is “yeah, but what abt the absolute change?”
Which here? Also the biggest of all time.
We're in uncharted territory- if WTI crude holds these gains tomorrow- the current rise of 29.9% or $27.16/bbl would be its largest single day price jump- in both percent and amount- ever. Ever. EVER.
March 9, 2026 at 2:42 AM
As someone quick to point out baserate fallacies, looks like the oil price jump is not one.
My usual rebuttal to “biggest percent change!” is “yeah, but what abt the absolute change?”
Which here? Also the biggest of all time.
My usual rebuttal to “biggest percent change!” is “yeah, but what abt the absolute change?”
Which here? Also the biggest of all time.
Ok, that’s it. I don’t think I’ll ever find a better example of the baserate fallacy than this.
Video ad in Houston airport for Texit coin, the “future of money in Texas.”
September 9, 2025 at 5:27 PM
Ok, that’s it. I don’t think I’ll ever find a better example of the baserate fallacy than this.
Also, if they were in fact up 5000%, that would mean there were ~0 to start with.
I swear to God, percentages and the baserate problem are one of the single greatest threats to sound policy, and I’m not really being hyperbolic.
I swear to God, percentages and the baserate problem are one of the single greatest threats to sound policy, and I’m not really being hyperbolic.
SENATOR @MarshaBlackburn: “We need the National guard in MEMPHIS because boxcar/train burglaries are up 5000%.”
MEMPHIS POLICE: “Those crimes are actually down.”
🤷🏼♀️ 🚂 (h/t @NoApologies901)
MEMPHIS POLICE: “Those crimes are actually down.”
🤷🏼♀️ 🚂 (h/t @NoApologies901)
September 23, 2025 at 2:19 PM
Also, if they were in fact up 5000%, that would mean there were ~0 to start with.
I swear to God, percentages and the baserate problem are one of the single greatest threats to sound policy, and I’m not really being hyperbolic.
I swear to God, percentages and the baserate problem are one of the single greatest threats to sound policy, and I’m not really being hyperbolic.
And yeah, these are absolutes, so long-term numbers say less.
But the absolute decline *just since 2022* has been the biggest in modern history—even compared to the 2000s, which has a comparable baserate.
And crickets on that too.
So it’s not a rates-vs-absolutes issue—it’s a deeper problem.
But the absolute decline *just since 2022* has been the biggest in modern history—even compared to the 2000s, which has a comparable baserate.
And crickets on that too.
So it’s not a rates-vs-absolutes issue—it’s a deeper problem.
December 23, 2025 at 10:40 PM
And yeah, these are absolutes, so long-term numbers say less.
But the absolute decline *just since 2022* has been the biggest in modern history—even compared to the 2000s, which has a comparable baserate.
And crickets on that too.
So it’s not a rates-vs-absolutes issue—it’s a deeper problem.
But the absolute decline *just since 2022* has been the biggest in modern history—even compared to the 2000s, which has a comparable baserate.
And crickets on that too.
So it’s not a rates-vs-absolutes issue—it’s a deeper problem.
Few words make me fear “this is going to be misleading” more than “the fasting growing ….”
The baserate fallacy is real and omnipresent.
(The post I’m subtweeting isn’t worth flagging, but any time I see that line I feel the need to raise the baserate fallacy alarm.)
The baserate fallacy is real and omnipresent.
(The post I’m subtweeting isn’t worth flagging, but any time I see that line I feel the need to raise the baserate fallacy alarm.)
November 23, 2025 at 5:18 PM
Few words make me fear “this is going to be misleading” more than “the fasting growing ….”
The baserate fallacy is real and omnipresent.
(The post I’m subtweeting isn’t worth flagging, but any time I see that line I feel the need to raise the baserate fallacy alarm.)
The baserate fallacy is real and omnipresent.
(The post I’m subtweeting isn’t worth flagging, but any time I see that line I feel the need to raise the baserate fallacy alarm.)
Baserate fallacy? ✅
Jonathan Haidt support? ✅
Article later admits most kids are actually taking types of non-home classes, just in a different form, bc admit officers don’t trust parent grades? ✅
Clear selection bias of wealthy parents w big tutoring and college counseling budgets? ✅
Jonathan Haidt support? ✅
Article later admits most kids are actually taking types of non-home classes, just in a different form, bc admit officers don’t trust parent grades? ✅
Clear selection bias of wealthy parents w big tutoring and college counseling budgets? ✅
Looking for an edge in college admissions? Just pull your kid from school. Homeschooling, long associated with hippies and religious conservatives in the U.S., is in the middle of a rebrand and a boom.
The Homeschooling Hack
Looking for an edge in college admissions? Just pull your kid from school. Homeschooling, long associated with hippies and religious conservatives in the U.S., is in the middle of a rebrand and a boom.
nymag.com
February 12, 2026 at 1:15 PM
Baserate fallacy? ✅
Jonathan Haidt support? ✅
Article later admits most kids are actually taking types of non-home classes, just in a different form, bc admit officers don’t trust parent grades? ✅
Clear selection bias of wealthy parents w big tutoring and college counseling budgets? ✅
Jonathan Haidt support? ✅
Article later admits most kids are actually taking types of non-home classes, just in a different form, bc admit officers don’t trust parent grades? ✅
Clear selection bias of wealthy parents w big tutoring and college counseling budgets? ✅
I also miss the account that would quote-tweet science headlines that just included the relative effect (“5000% increase in the risk of dying!”) with the absolute effect as well (“risk of death rises by 0.0002%”).
Pure uncut baserate fallacy correction.
Pure uncut baserate fallacy correction.
Aw, man, this is the rare post that makes me miss Twitter, where there was an account that would regularly quote tweet scientific studies that failed to include “in mice” in the headline and just add “… in mice.”
It was (no snark) a HUGE public service.
It was (no snark) a HUGE public service.
"In mice"
April 17, 2024 at 2:23 AM
I also miss the account that would quote-tweet science headlines that just included the relative effect (“5000% increase in the risk of dying!”) with the absolute effect as well (“risk of death rises by 0.0002%”).
Pure uncut baserate fallacy correction.
Pure uncut baserate fallacy correction.
Perhaps in the policing context it's a slightly better predictor of violence, simply because there are fewer cops (but then the baserate is higher than the general pop's, which undermines its predictive power).
But it's still likely a pretty inefficient one (in part bc of that baserate issue).
But it's still likely a pretty inefficient one (in part bc of that baserate issue).
July 17, 2026 at 4:38 AM
Perhaps in the policing context it's a slightly better predictor of violence, simply because there are fewer cops (but then the baserate is higher than the general pop's, which undermines its predictive power).
But it's still likely a pretty inefficient one (in part bc of that baserate issue).
But it's still likely a pretty inefficient one (in part bc of that baserate issue).
In standard scientific reports in psych 96% of first mentioned main hypotheses are supported. In Registered Reports, this is 46%. The 96% is clearly biased (given true H1 rate and power). Lack of transparency means we do not know the true baserate.
journals.sagepub.com/doi/10.1177/...
journals.sagepub.com/doi/10.1177/...
February 14, 2026 at 6:08 PM
In standard scientific reports in psych 96% of first mentioned main hypotheses are supported. In Registered Reports, this is 46%. The 96% is clearly biased (given true H1 rate and power). Lack of transparency means we do not know the true baserate.
journals.sagepub.com/doi/10.1177/...
journals.sagepub.com/doi/10.1177/...
There are two statistical hills I am always ready to die on:
1. Stop making baserate errors (big percent change likely means NOT that important).
2. Stop selecting on the dependent variable.
(My 1Ls this afternoon are about to get an unexpected stats class.)
1. Stop making baserate errors (big percent change likely means NOT that important).
2. Stop selecting on the dependent variable.
(My 1Ls this afternoon are about to get an unexpected stats class.)
September 5, 2024 at 11:37 AM
There are two statistical hills I am always ready to die on:
1. Stop making baserate errors (big percent change likely means NOT that important).
2. Stop selecting on the dependent variable.
(My 1Ls this afternoon are about to get an unexpected stats class.)
1. Stop making baserate errors (big percent change likely means NOT that important).
2. Stop selecting on the dependent variable.
(My 1Ls this afternoon are about to get an unexpected stats class.)
Any large percent change is almost always from negligible to still-negligible.
When this baserate issue doesn't hold--when a huge percent change = huge absolute change--it's something huge and catastrophic we don't need data to know (see: Covid).
Also, "assault" can be defined REALLY broadly.
When this baserate issue doesn't hold--when a huge percent change = huge absolute change--it's something huge and catastrophic we don't need data to know (see: Covid).
Also, "assault" can be defined REALLY broadly.
July 1, 2025 at 4:04 PM
Any large percent change is almost always from negligible to still-negligible.
When this baserate issue doesn't hold--when a huge percent change = huge absolute change--it's something huge and catastrophic we don't need data to know (see: Covid).
Also, "assault" can be defined REALLY broadly.
When this baserate issue doesn't hold--when a huge percent change = huge absolute change--it's something huge and catastrophic we don't need data to know (see: Covid).
Also, "assault" can be defined REALLY broadly.
Huh. Trump is a uniter, not a divider.
This is one of the only polls in recent memory that I can think of where the results aren’t just “% of respondents who are Dem [or GOP] +/- a small error.”
For once, “twice as many GOPers!” is a baserate issue, not a “The City and The City” issue.
This is one of the only polls in recent memory that I can think of where the results aren’t just “% of respondents who are Dem [or GOP] +/- a small error.”
For once, “twice as many GOPers!” is a baserate issue, not a “The City and The City” issue.
The poll can be found here: today.yougov.com/topics/inter...
And the partisan splits (or relative lack there of) manage to be even more impressive
And the partisan splits (or relative lack there of) manage to be even more impressive
June 23, 2025 at 3:33 AM
Huh. Trump is a uniter, not a divider.
This is one of the only polls in recent memory that I can think of where the results aren’t just “% of respondents who are Dem [or GOP] +/- a small error.”
For once, “twice as many GOPers!” is a baserate issue, not a “The City and The City” issue.
This is one of the only polls in recent memory that I can think of where the results aren’t just “% of respondents who are Dem [or GOP] +/- a small error.”
For once, “twice as many GOPers!” is a baserate issue, not a “The City and The City” issue.
Updatedletter, I know go into soem emperical results as well: zenodo.org/records/1800...
The Apperent Negative Associations Between Early and Adult Performance in elite performers Arise from Colider Selection Bias and Baserate Neglect
Güllich et al. argue that among elite performers there is a negative association between early and adult performance, a pattern they link to distinct developmental causal mechanisms for early, and adu...
zenodo.org
December 21, 2025 at 12:01 PM
Updatedletter, I know go into soem emperical results as well: zenodo.org/records/1800...
There are many various policy hills I am willing to die on in the public, all of which I think have merit, but some more than others.
Fighting the baserate fallacy ("big percent change means it matters the most!") is probably one of the easiest to justify.
Fighting the baserate fallacy ("big percent change means it matters the most!") is probably one of the easiest to justify.
June 18, 2024 at 8:02 PM
There are many various policy hills I am willing to die on in the public, all of which I think have merit, but some more than others.
Fighting the baserate fallacy ("big percent change means it matters the most!") is probably one of the easiest to justify.
Fighting the baserate fallacy ("big percent change means it matters the most!") is probably one of the easiest to justify.
Hm. AI speeds up history by a bunch, humanity would've died out in history far away, which is brought closer by said speedup.
& species often get outcompeted by being reconnected with near-speciated rivals, I tried to get a baserate for this (endogenous vs. exogenous) but failed
& species often get outcompeted by being reconnected with near-speciated rivals, I tried to get a baserate for this (endogenous vs. exogenous) but failed
September 16, 2026 at 6:56 PM
Hm. AI speeds up history by a bunch, humanity would've died out in history far away, which is brought closer by said speedup.
& species often get outcompeted by being reconnected with near-speciated rivals, I tried to get a baserate for this (endogenous vs. exogenous) but failed
& species often get outcompeted by being reconnected with near-speciated rivals, I tried to get a baserate for this (endogenous vs. exogenous) but failed
I believe this makes it now three esoteric, somewhat quixotic hills on which I am willing to die:
1. Baserate fallacies.
2. Selecting on the dependent variable fallacies.
3. Use of the word "pre-print."
Guess I need to find 4 more to satisfy the "all men think abt the Roman Empire daily" rule.
1. Baserate fallacies.
2. Selecting on the dependent variable fallacies.
3. Use of the word "pre-print."
Guess I need to find 4 more to satisfy the "all men think abt the Roman Empire daily" rule.
I've seen "pre-print" twice in the past two days, so feel like I need to repost this.
"Pre-print" is a terribly misleading term which elevates working papers to something more, and then makes all empirical work look less reliable when those results shift.
Working paper.
Call it a working paper.
"Pre-print" is a terribly misleading term which elevates working papers to something more, and then makes all empirical work look less reliable when those results shift.
Working paper.
Call it a working paper.
Just saw "preprint" in the wild today, and so just re-upping my regular plea to please, please PLEASE stop using this word.
It's "working paper." The word we NEED to use is "working paper."
"Preprint" sounds like it's been accepted but not yet printed. It ... has not. And it misleads the public.
It's "working paper." The word we NEED to use is "working paper."
"Preprint" sounds like it's been accepted but not yet printed. It ... has not. And it misleads the public.
October 16, 2024 at 6:31 PM
I believe this makes it now three esoteric, somewhat quixotic hills on which I am willing to die:
1. Baserate fallacies.
2. Selecting on the dependent variable fallacies.
3. Use of the word "pre-print."
Guess I need to find 4 more to satisfy the "all men think abt the Roman Empire daily" rule.
1. Baserate fallacies.
2. Selecting on the dependent variable fallacies.
3. Use of the word "pre-print."
Guess I need to find 4 more to satisfy the "all men think abt the Roman Empire daily" rule.
#quotes #books #booksky #excerpts #citations #delusionsofcrowds #bernstein #authority #expert #predict #prediction #politics #economics #forecast #baserate #delusions #crowds #psychology
October 27, 2025 at 10:30 PM
Anyway, tl;dr: from a baserate of ~4,000 monthly impressions (which, even on collapsing Xitter, is a drop in the ocean), it jumped to 375K. Still … not much, even if not good.
The article didn’t give the 4,000. Just the 375,000 in the body, and just the 9500% in the title.
Not-great-Bob-dot-gif.
The article didn’t give the 4,000. Just the 375,000 in the body, and just the 9500% in the title.
Not-great-Bob-dot-gif.
December 8, 2023 at 3:42 AM
Anyway, tl;dr: from a baserate of ~4,000 monthly impressions (which, even on collapsing Xitter, is a drop in the ocean), it jumped to 375K. Still … not much, even if not good.
The article didn’t give the 4,000. Just the 375,000 in the body, and just the 9500% in the title.
Not-great-Bob-dot-gif.
The article didn’t give the 4,000. Just the 375,000 in the body, and just the 9500% in the title.
Not-great-Bob-dot-gif.
Key result # 1: From natural user behavior as input, both passive and active engagement algorithms amplified IME posts disproportionately higher than their baserate in the network➡️ overrepresentation of IME content in feeds
February 14, 2025 at 3:27 PM
Key result # 1: From natural user behavior as input, both passive and active engagement algorithms amplified IME posts disproportionately higher than their baserate in the network➡️ overrepresentation of IME content in feeds
If my genuine joy at this seems confusing, (1) welcome, you clearly are someone who didn't follow me over at the Other Place, glad to see you here.
(2) Almost always, "HUGE PERCENT CHANGE" means ... DOESN'T REALLY MATTER.
Because the baserate needs to be SMALL to get a BIG percent change.
(2) Almost always, "HUGE PERCENT CHANGE" means ... DOESN'T REALLY MATTER.
Because the baserate needs to be SMALL to get a BIG percent change.
June 18, 2024 at 7:46 PM
If my genuine joy at this seems confusing, (1) welcome, you clearly are someone who didn't follow me over at the Other Place, glad to see you here.
(2) Almost always, "HUGE PERCENT CHANGE" means ... DOESN'T REALLY MATTER.
Because the baserate needs to be SMALL to get a BIG percent change.
(2) Almost always, "HUGE PERCENT CHANGE" means ... DOESN'T REALLY MATTER.
Because the baserate needs to be SMALL to get a BIG percent change.
I imagine its just kinda a baserate thing of like most americans are either not leftists or, if they are a leftist, don't have the means to firebomb a walmart
April 27, 2026 at 8:19 PM
I imagine its just kinda a baserate thing of like most americans are either not leftists or, if they are a leftist, don't have the means to firebomb a walmart
Not looking to stan for sports betting--I'm sure it's going to be a big net social harm, like pretty much all gambling is.
But I can't not point out baserate issues whenever I see them in the wild.
Baserates and survivorship-biased claims. Almost reflexive responses.
But I can't not point out baserate issues whenever I see them in the wild.
Baserates and survivorship-biased claims. Almost reflexive responses.
a cat is sitting in front of a door with a sign that says zebrani
Alt: cat swatting at a ball.
media.tenor.com
September 23, 2024 at 3:30 PM
Not looking to stan for sports betting--I'm sure it's going to be a big net social harm, like pretty much all gambling is.
But I can't not point out baserate issues whenever I see them in the wild.
Baserates and survivorship-biased claims. Almost reflexive responses.
But I can't not point out baserate issues whenever I see them in the wild.
Baserates and survivorship-biased claims. Almost reflexive responses.
Luck/randomness and a low baserate offense type.
November 12, 2024 at 8:00 PM
Luck/randomness and a low baserate offense type.
People's intuitions about #news shifted toward the baserate of true #headlines in a #socialMedia feed among 2.8k (2.5k excluded).
A mostly true feed biased intuitions towards 'true' (and a mostly false feed biased intuitions towards 'false') ratings.
doi.org/10.31234/osf...
A mostly true feed biased intuitions towards 'true' (and a mostly false feed biased intuitions towards 'false') ratings.
doi.org/10.31234/osf...
April 4, 2025 at 10:28 AM
People's intuitions about #news shifted toward the baserate of true #headlines in a #socialMedia feed among 2.8k (2.5k excluded).
A mostly true feed biased intuitions towards 'true' (and a mostly false feed biased intuitions towards 'false') ratings.
doi.org/10.31234/osf...
A mostly true feed biased intuitions towards 'true' (and a mostly false feed biased intuitions towards 'false') ratings.
doi.org/10.31234/osf...