Knox Study
Region: GB
Monday 01 September 2025 09:23:56 GMT
131188
4778
95
191
Music
Download
Comments
Djjdkxkdkd :
i was in that 0.1% of false positives once and it was traumatic 😍
2025-09-01 12:05:09
680
ThomasSmith8582959 :
this assumes that everyone indiscriminately receives a test, without any reason
2025-09-01 19:26:12
23
Jay :
This is assuming that 10k people take the test. Most people that take a test have a reason to believe they might have it, likely what you describe as high risk
2025-09-01 16:35:50
162
Brian :
This is not how Bayes’ theorem works on an individual result my guy
2025-09-02 21:26:47
66
J 🏴 :
if you never had a risky behaviour your chances are nil
2025-09-02 20:17:39
3
thelidds :
this is making my head spin 😂 Question though, are you assuming that the 0.1% is referring to a false positive? Is it possible the accuracy figure includes false negatives as well or even in its entirety? What would this do to the figures?
2025-09-01 10:35:59
39
David Khudaverdyan :
The accuracy of the test should be represented by more natural definition: Assuming you are truly positive, what are the odds, that the test shows positive. And respectively: if you are truly negative, what are the odds you trsted negative. These two independent numbers don’t bear surprises mentioned in the video.
2025-09-01 17:16:37
5
Fardarterfu :
What are the chances of getting two false positives in a row?
2025-10-07 19:22:50
0
Narcissus :
in the low risk cohort , if you showed symptoms and took the test would this not change the prior distribution ? I've always wondered this
2025-09-01 14:38:03
12
jamalabdisalam :
in summary test accuracy becomes a bit of a limiting factor when you're in a low risk population
2025-09-01 14:30:06
2
BeanBoi :
Oh my days. Best ever explanation of bayes thereom🔥🔥🔥🔥
2025-09-01 13:34:41
14
yamalpina :
oh wow. oh wow. this is bonkers.
2025-09-01 13:44:51
7
fattyaeroplanes :
This assumes that you are randomly tested.
2025-10-27 22:02:43
2
Slinks :
i remember doing this in A level math
2025-09-01 12:14:17
4
VZ :
This is a new type of math- denialism🥲
2025-10-28 01:00:40
0
46 75 63 6B 20 79 6F 75 :
Give them something useful like markov models and how that can help them build critical mass in their lives
2025-09-01 19:46:20
0
Oblation :
it's a brilliant insight, game changing.
2025-09-01 09:37:34
4
Nicholas (Sedgewick) Thompson :
I’m curious about correlation in sequential false positives how that can be studied… presumably a physical reason some people more likely to be false positive for a given test so for them for a given test whoa crazy situation🤯
2025-09-01 12:42:30
0
Be Real English :
fascinating 🙏🙏🙏 good to remember 👍
2025-09-01 10:45:25
5
Max :
This really is a beautiful constructed example. I hope I remember this when it comes up some time down the line (inevitable as my field is 90% Bayesian stats)
2025-09-01 16:00:59
0
Yannick Bégin :
Well, in médecine we don’t use simply the term accuracy. we use sensibility for the accuracy to find true positive and specificity for accuracy to find true négative. It’s Even moré complexe than what you said. Thx for the vulgarisation
2025-10-19 17:25:17
0
Zowee :
It makes much more sense once you display the probability as a graph with a horizontal scale covering risk factors (e.g. needle sharing, promiscuity, etc. ) You will see a curve with a very flat slate for the most part and a steep increase and peak at the right edge. So if you are anywhere within the 75% lower risk range, your probability is very low because it's more likely to be a false positive.
2025-09-01 15:04:49
0
koostkoost :
A LOT of confidently wrong comments in here
2025-09-02 20:14:22
3
Things Can Only Get Vetter :
This is an enormous over-simplification of sensitivity and specificity. If the test has a very high specificity, and is 99.9% accurate, you have HIV.
2025-09-01 11:43:20
7
To see more videos from user @knoxstudy, please go to the Tikwm
homepage.
