Talk:1132: Frequentists vs. Bayesians - Revision history

146.70.174.163 at 14:44, 25 January 2026

2026-01-25T14:44:41Z

172.71.154.215 at 21:33, 24 March 2023

2023-03-24T21:33:51Z

BradleyRoss: Names such as Bayesian can be associated with multiple concepts. Turing used as example.

2023-02-23T01:32:33Z

Names such as Bayesian can be associated with multiple concepts. Turing used as example.

BradleyRoss: Names such as Bayesian can be associated with multiple concepts. Turing used as example.

2023-02-23T01:31:26Z

Names such as Bayesian can be associated with multiple concepts. Turing used as example.

BradleyRoss: Added comment

2023-02-23T01:26:09Z

Added comment

BradleyRoss: Adding some comments

2023-02-23T01:24:13Z

Adding some comments

108.162.229.54: Extraordinary claims require extraordinary evidence

2020-10-12T10:42:08Z

Extraordinary claims require extraordinary evidence

108.162.237.16 at 07:05, 21 August 2020

2020-08-21T07:05:58Z

172.69.69.244: Seeing if I missed anything about the humor in the final part of the punchline.

2019-12-03T14:46:39Z

Seeing if I missed anything about the humor in the final part of the punchline.

Kazvorpal: OMG WTF No Sig!!!

2019-11-11T21:57:11Z

OMG WTF No Sig!!!

← Older revision		Revision as of 14:44, 25 January 2026
Line 95:		Line 95:
	3: If the 95% confidence interval includes the value of 1/36 then it supports the null hypothesis that there is no correlation is suggested between the sun and the positive detector results.		3: If the 95% confidence interval includes the value of 1/36 then it supports the null hypothesis that there is no correlation is suggested between the sun and the positive detector results.
	Bayesian and Frequentist Statistics should yield the same result if handled correctly, because they are basically algebraic rearrangements of each other. --[[Special:Contributions/172.71.154.215\|172.71.154.215]] 21:33, 24 March 2023 (UTC)		Bayesian and Frequentist Statistics should yield the same result if handled correctly, because they are basically algebraic rearrangements of each other. --[[Special:Contributions/172.71.154.215\|172.71.154.215]] 21:33, 24 March 2023 (UTC)
		+	: So... explode the sun several dozen times? :P [[Special:Contributions/146.70.174.163\|146.70.174.163]] 14:44, 25 January 2026 (UTC)

← Older revision		Revision as of 21:33, 24 March 2023
Line 89:		Line 89:

	With regard to Bayesian having multiple meanings, this is probably similar to there being a Turing test, a Turing machine, and Turing Complete. [[User:BradleyRoss\|BradleyRoss]] ([[User talk:BradleyRoss\|talk]]) 01:32, 23 February 2023 (UTC)		With regard to Bayesian having multiple meanings, this is probably similar to there being a Turing test, a Turing machine, and Turing Complete. [[User:BradleyRoss\|BradleyRoss]] ([[User talk:BradleyRoss\|talk]]) 01:32, 23 February 2023 (UTC)
		+
		+	No-one has discussed how to properly apply frequentists statistics to the problem.
		+	1: Rerun the test several dozen times.
		+	2: Find the 95% confidence interval of the generated data (A Poisson distribution is most appropriate for the modeling of event frequency).
		+	3: If the 95% confidence interval includes the value of 1/36 then it supports the null hypothesis that there is no correlation is suggested between the sun and the positive detector results.
		+	Bayesian and Frequentist Statistics should yield the same result if handled correctly, because they are basically algebraic rearrangements of each other. --[[Special:Contributions/172.71.154.215\|172.71.154.215]] 21:33, 24 March 2023 (UTC)

← Older revision		Revision as of 01:32, 23 February 2023
Line 88:		Line 88:
	If the eventist says "Detector! What would the Bayesian statistician say if I asked him whether I would say the sun had exploded", the Bayesian doesn't know what the detector would say. (I am changing the wording slightly, but it doesn't make sense to me as stated.) Therefore, the Bayesian can't give an answer. The Bayesian's answer would therefore be "I am a neutrino detector (answers are sometimes true and sometimes false), not a labyrinth guard (answers are always true or always false)". He then predicts that the Bayesian would say "Seriously, did your brain fall out?" After somebody hits the button, the detector answers truthfully (the likeliest option), and gives his opinion "YES". [[User:BradleyRoss\|BradleyRoss]] ([[User talk:BradleyRoss\|talk]]) 01:26, 23 February 2023 (UTC)		If the eventist says "Detector! What would the Bayesian statistician say if I asked him whether I would say the sun had exploded", the Bayesian doesn't know what the detector would say. (I am changing the wording slightly, but it doesn't make sense to me as stated.) Therefore, the Bayesian can't give an answer. The Bayesian's answer would therefore be "I am a neutrino detector (answers are sometimes true and sometimes false), not a labyrinth guard (answers are always true or always false)". He then predicts that the Bayesian would say "Seriously, did your brain fall out?" After somebody hits the button, the detector answers truthfully (the likeliest option), and gives his opinion "YES". [[User:BradleyRoss\|BradleyRoss]] ([[User talk:BradleyRoss\|talk]]) 01:26, 23 February 2023 (UTC)

−	With regard to Bayesian having multiple meanings, this is probably similar to there being a Turing test, a Turing machine, and Turing Complete.	+	With regard to Bayesian having multiple meanings, this is probably similar to there being a Turing test, a Turing machine, and Turing Complete. [[User:BradleyRoss\|BradleyRoss]] ([[User talk:BradleyRoss\|talk]]) 01:32, 23 February 2023 (UTC)

← Older revision		Revision as of 01:31, 23 February 2023
Line 87:		Line 87:

	If the eventist says "Detector! What would the Bayesian statistician say if I asked him whether I would say the sun had exploded", the Bayesian doesn't know what the detector would say. (I am changing the wording slightly, but it doesn't make sense to me as stated.) Therefore, the Bayesian can't give an answer. The Bayesian's answer would therefore be "I am a neutrino detector (answers are sometimes true and sometimes false), not a labyrinth guard (answers are always true or always false)". He then predicts that the Bayesian would say "Seriously, did your brain fall out?" After somebody hits the button, the detector answers truthfully (the likeliest option), and gives his opinion "YES". [[User:BradleyRoss\|BradleyRoss]] ([[User talk:BradleyRoss\|talk]]) 01:26, 23 February 2023 (UTC)		If the eventist says "Detector! What would the Bayesian statistician say if I asked him whether I would say the sun had exploded", the Bayesian doesn't know what the detector would say. (I am changing the wording slightly, but it doesn't make sense to me as stated.) Therefore, the Bayesian can't give an answer. The Bayesian's answer would therefore be "I am a neutrino detector (answers are sometimes true and sometimes false), not a labyrinth guard (answers are always true or always false)". He then predicts that the Bayesian would say "Seriously, did your brain fall out?" After somebody hits the button, the detector answers truthfully (the likeliest option), and gives his opinion "YES". [[User:BradleyRoss\|BradleyRoss]] ([[User talk:BradleyRoss\|talk]]) 01:26, 23 February 2023 (UTC)
		+
		+	With regard to Bayesian having multiple meanings, this is probably similar to there being a Turing test, a Turing machine, and Turing Complete.

← Older revision		Revision as of 01:26, 23 February 2023
Line 86:		Line 86:
	When he Bayesian says "Bet you $50 it hasn't.", he is saying that he will probably win the bet. However, he isn't saying he knows whether the sun has exploded or whether the detector is lying. What he is saying is roughly equivalent to "If we are playing Texas Holdem and I have a royal flush while you have nothing showing, I am probably going to win and might as well bet what I can.		When he Bayesian says "Bet you $50 it hasn't.", he is saying that he will probably win the bet. However, he isn't saying he knows whether the sun has exploded or whether the detector is lying. What he is saying is roughly equivalent to "If we are playing Texas Holdem and I have a royal flush while you have nothing showing, I am probably going to win and might as well bet what I can.

−	If the eventist says "Detector! What would the Bayesian statistician say if I asked him whether I would say the sun had exploded", the Bayesian doesn't know what the detector would say. (I am changing the wording slightly, but it doesn't make sense to me as stated.) Therefore, the Bayesian can't give an answer. The Bayesian's answer would therefore be "I am a neutrino detector (answers are sometimes true and sometimes false), not a labyrinth guard (answers are always true or always false)". He then predicts that the Bayesian would say "Seriously, did your brain fall out?" After somebody hits the button, the detector answers truthfully (the likeliest option), and gives his opinion "YES". 01:24, 23 February 2023 (UTC)	+	If the eventist says "Detector! What would the Bayesian statistician say if I asked him whether I would say the sun had exploded", the Bayesian doesn't know what the detector would say. (I am changing the wording slightly, but it doesn't make sense to me as stated.) Therefore, the Bayesian can't give an answer. The Bayesian's answer would therefore be "I am a neutrino detector (answers are sometimes true and sometimes false), not a labyrinth guard (answers are always true or always false)". He then predicts that the Bayesian would say "Seriously, did your brain fall out?" After somebody hits the button, the detector answers truthfully (the likeliest option), and gives his opinion "YES". [[User:BradleyRoss\|BradleyRoss]] ([[User talk:BradleyRoss\|talk]]) 01:26, 23 February 2023 (UTC)

← Older revision		Revision as of 01:24, 23 February 2023
Line 83:		Line 83:

	As per Sagan, "Extraordinary claims require extraordinary evidence". [[Special:Contributions/108.162.229.54\|108.162.229.54]] 10:42, 12 October 2020 (UTC)		As per Sagan, "Extraordinary claims require extraordinary evidence". [[Special:Contributions/108.162.229.54\|108.162.229.54]] 10:42, 12 October 2020 (UTC)
		+
		+	When he Bayesian says "Bet you $50 it hasn't.", he is saying that he will probably win the bet. However, he isn't saying he knows whether the sun has exploded or whether the detector is lying. What he is saying is roughly equivalent to "If we are playing Texas Holdem and I have a royal flush while you have nothing showing, I am probably going to win and might as well bet what I can.
		+
		+	If the eventist says "Detector! What would the Bayesian statistician say if I asked him whether I would say the sun had exploded", the Bayesian doesn't know what the detector would say. (I am changing the wording slightly, but it doesn't make sense to me as stated.) Therefore, the Bayesian can't give an answer. The Bayesian's answer would therefore be "I am a neutrino detector (answers are sometimes true and sometimes false), not a labyrinth guard (answers are always true or always false)". He then predicts that the Bayesian would say "Seriously, did your brain fall out?" After somebody hits the button, the detector answers truthfully (the likeliest option), and gives his opinion "YES". 01:24, 23 February 2023 (UTC)

← Older revision		Revision as of 10:42, 12 October 2020
Line 81:		Line 81:

	Is there a statistical angle I'm missing to the final part of the mouseover text 'did your brain fall out? [roll] yes...' Or is is purely linguistic between literal and figurative i.e. if his brain has fallen out as in he has made a careless error, then that's true. If it's literally did his brain fall out, is the 'yes' the 97% chance that it's talking about his mistake, or the ~3% chance that it's lying about the literal truth? [[Special:Contributions/172.69.69.244\|172.69.69.244]] 14:46, 3 December 2019 (UTC)		Is there a statistical angle I'm missing to the final part of the mouseover text 'did your brain fall out? [roll] yes...' Or is is purely linguistic between literal and figurative i.e. if his brain has fallen out as in he has made a careless error, then that's true. If it's literally did his brain fall out, is the 'yes' the 97% chance that it's talking about his mistake, or the ~3% chance that it's lying about the literal truth? [[Special:Contributions/172.69.69.244\|172.69.69.244]] 14:46, 3 December 2019 (UTC)
		+
		+	As per Sagan, "Extraordinary claims require extraordinary evidence". [[Special:Contributions/108.162.229.54\|108.162.229.54]] 10:42, 12 October 2020 (UTC)

← Older revision		Revision as of 07:05, 21 August 2020
Line 1:		Line 1:
		+	I just sort of assumed he bet 50 dollars because if the sun had exploded, they'd be dead and therefore wouldn't need the machine. [[Special:Contributions/108.162.237.16\|108.162.237.16]] 07:05, 21 August 2020 (UTC)
		+
	Something should be added about the prior probability of the sun going nova, as that is the primary substantive point. "The neutrino detector is evidence that the Sun has exploded. It's showing an observation which is 35 times more likely to appear if the Sun has exploded than if it hasn't (likelihood ratio of 35:1). The Bayesian just doesn't think that's strong enough evidence to overcome the prior odds, i.e., after multiplying the prior odds by 35 they still aren't very high." - http://lesswrong.com/r/discussion/lw/fe5/xkcd_frequentist_vs_bayesians/ [[Special:Contributions/209.65.52.92\|209.65.52.92]] 23:51, 9 November 2012 (UTC)		Something should be added about the prior probability of the sun going nova, as that is the primary substantive point. "The neutrino detector is evidence that the Sun has exploded. It's showing an observation which is 35 times more likely to appear if the Sun has exploded than if it hasn't (likelihood ratio of 35:1). The Bayesian just doesn't think that's strong enough evidence to overcome the prior odds, i.e., after multiplying the prior odds by 35 they still aren't very high." - http://lesswrong.com/r/discussion/lw/fe5/xkcd_frequentist_vs_bayesians/ [[Special:Contributions/209.65.52.92\|209.65.52.92]] 23:51, 9 November 2012 (UTC)

← Older revision		Revision as of 14:46, 3 December 2019
Line 77:		Line 77:

	It is my feeling that sloppy or machiavellian academics have come to use the term "Bayesian" to mean something more like "we adjusted it to what we felt was most reasonable", which introduces so much bias that it actually leaves one unable to determine the scientific validity of the results. I was reading [https://academic.oup.com/beheco/article/23/3/619/224216 a publication], today, that made me think of that and look up this comic. —[[User:Kazvorpal\|Kazvorpal]] ([[User talk:Kazvorpal\|talk]]) 21:57, 11 November 2019 (UTC)		It is my feeling that sloppy or machiavellian academics have come to use the term "Bayesian" to mean something more like "we adjusted it to what we felt was most reasonable", which introduces so much bias that it actually leaves one unable to determine the scientific validity of the results. I was reading [https://academic.oup.com/beheco/article/23/3/619/224216 a publication], today, that made me think of that and look up this comic. —[[User:Kazvorpal\|Kazvorpal]] ([[User talk:Kazvorpal\|talk]]) 21:57, 11 November 2019 (UTC)
		+
		+	Is there a statistical angle I'm missing to the final part of the mouseover text 'did your brain fall out? [roll] yes...' Or is is purely linguistic between literal and figurative i.e. if his brain has fallen out as in he has made a careless error, then that's true. If it's literally did his brain fall out, is the 'yes' the 97% chance that it's talking about his mistake, or the ~3% chance that it's lying about the literal truth? [[Special:Contributions/172.69.69.244\|172.69.69.244]] 14:46, 3 December 2019 (UTC)

← Older revision		Revision as of 21:57, 11 November 2019
Line 76:		Line 76:
	This comic hurts my head. [[Special:Contributions/173.245.54.7\|173.245.54.7]] 21:44, 12 August 2019 (UTC)		This comic hurts my head. [[Special:Contributions/173.245.54.7\|173.245.54.7]] 21:44, 12 August 2019 (UTC)

−	It is my feeling that sloppy or machiavellian academics have come to use the term "Bayesian" to mean something more like "we adjusted it to what we felt was most reasonable", which introduces so much bias that it actually leaves one unable to determine the scientific validity of the results. I was reading [https://academic.oup.com/beheco/article/23/3/619/224216 a publication], today, that made me think of that and look up this comic.	+	It is my feeling that sloppy or machiavellian academics have come to use the term "Bayesian" to mean something more like "we adjusted it to what we felt was most reasonable", which introduces so much bias that it actually leaves one unable to determine the scientific validity of the results. I was reading [https://academic.oup.com/beheco/article/23/3/619/224216 a publication], today, that made me think of that and look up this comic. —[[User:Kazvorpal\|Kazvorpal]] ([[User talk:Kazvorpal\|talk]]) 21:57, 11 November 2019 (UTC)