|
Honj Steak posted:I miss the Babylon central bank’s monetary policy Ea-nāṣir had some interesting "buy now, pay later deals" but then he just didn't deliver the goods
|
# ? Oct 4, 2023 19:13 |
|
|
# ? May 28, 2024 04:10 |
|
|
# ? Oct 8, 2023 02:07 |
|
Yes this is definitely awful
|
# ? Oct 8, 2023 02:12 |
|
I found the source. they've plotted the x-axis as a log plot for the data, but then plopped that equation on top of it as if it was linear and each bar was 1 minute. I.e. as if x=131,072 was x=19 and x=never was x=20. Or in other words, what they've plotted isn't y = 23.124 e^[-0.120 x], it's y = 23.124 * x^[-0.120/(loge2)] or something like that. In the text, they do recognize that it is a power-law distribution, so I have no idea why they decided to put that wrong equation on the plot. How they came to this fitting function is anyone's guess; it is descrived as an "estimated curve" rather than a fitting function. The x=0, y=25 bar should not appear on this graph (as well as the NEVER bar, but that one's more obvious). Each bar is twice the time of the previous, and half the time of the next. Except the first bar is infinity times less than the second bar. Lastly, I was confused about the sample size. The data clearly shows the survival rate at each time for an original sample size of N=25; i.e. of 25 vandalisms, 5 get corrected and 20 survive within the first minute, etc. So how on earth can that data be collected from "100 random articles"? Surely 25 vandalisms are on at most 25 different articles. It turns out that they looked at 100 random articles' edit history, found the most recent vandalism, and used that (which doesn't seem like a statistically sound method, because articles get vandalized at different rates). So of the 100 they looked at, only 25 had ever been vandalized. So that explains that. The raw data is in there if anyone wants to take a crack at plotting it and coming up with a better fitting function.
|
# ? Oct 8, 2023 03:36 |
|
DontMockMySmock posted:I found the source. i appreciate how you tried to explain it two ways and both were indecipherable
|
# ? Oct 8, 2023 13:48 |
|
I think it's funnier to imagine they selected 100 random articles, vandalized them, and then waited 91 days to see if they got fixed. Edit: (they repeated this 25 times to be scientific )
|
# ? Oct 8, 2023 15:49 |
|
DontMockMySmock posted:In the text, they do recognize that it is a power-law distribution, so I have no idea why they decided to put that wrong equation on the plot. How they came to this fitting function is anyone's guess; it is descrived as an "estimated curve" rather than a fitting function. Search your feelings, you know what the truth for both of these is. They put it into Excel
|
# ? Oct 8, 2023 16:04 |
|
Regarde Aduck posted:i appreciate how you tried to explain it two ways and both were indecipherable You're welcome
|
# ? Oct 8, 2023 20:59 |
|
|
# ? Oct 9, 2023 03:38 |
|
But why the U?
|
# ? Oct 9, 2023 12:26 |
|
Language acquires more u's with increasing proximity to France. Known fact.
|
# ? Oct 9, 2023 13:11 |
|
Alternatively, q doesn't exist in English without the u. Except for recent loanwords like Qatar 0
|
# ? Oct 9, 2023 16:30 |
|
then why don't we treat it as a letter like how the dutch have I, J, and IJ
|
# ? Oct 9, 2023 18:52 |
|
BonHair posted:Alternatively, q doesn't exist in English without the u. Except for recent loanwords like Qatar 0 And proper names hardly count
|
# ? Oct 9, 2023 19:00 |
|
maybealabia posted:And proper names hardly count I think there are a few others, like qi Mister Olympus posted:then why don't we treat it as a letter like how the dutch have I, J, and IJ Because a) English is a gently caress and b) unlike Dutch ij, English q is pretty uncommon, with (at a guess) 95% of uses being loans from romance languages (French and Latin).
|
# ? Oct 9, 2023 19:08 |
|
BonHair posted:I think there are a few others, like qi Qi is a good Scrabble word but I think most people use "chi" Still obviously a foreign loan word
|
# ? Oct 9, 2023 19:20 |
|
Mister Olympus posted:then why don't we treat it as a letter like how the dutch have I, J, and IJ Because no one wants to rewrite the alphabet song.
|
# ? Oct 9, 2023 20:15 |
|
Mister Olympus posted:then why don't we treat it as a letter like how the dutch have I, J, and IJ Why would we give any credence to the dutch?
|
# ? Oct 9, 2023 20:25 |
|
Even Afrikaners realised 'y' was always the way to go. But now the Dutch don't want to give any credence to South Africans.
|
# ? Oct 10, 2023 09:49 |
|
Paladinus posted:Even Afrikaners realised 'y' was always the way to go. But now the Dutch don't want to give any credence to South Africans. Nobody should give credence to South Africans.
|
# ? Oct 10, 2023 13:13 |
|
https://twitter.com/DKThomp/status/1713937759113498642 Soon, the rest of the world will join Danish mothers in spending infinite time with their children. Only in France will you be able to see a mother without her child
|
# ? Oct 17, 2023 04:30 |
|
I’m the 1965 parent somehow keeping a child alive spending only 25 minutes a day with them.
|
# ? Oct 17, 2023 06:48 |
|
Ah, those perfect days when you had to let the local drunk watch your kids in the street while you got some work done.
|
# ? Oct 17, 2023 06:50 |
|
Adhemar posted:I’m the 1965 parent somehow keeping a child alive spending only 25 minutes a day with them. Dr Spock's classic guide to modern scientific parenting, "The Common Sense Guide to Faun and Rabbit Kit Care".
|
# ? Oct 17, 2023 07:50 |
|
Adhemar posted:I’m the 1965 parent somehow keeping a child alive spending only 25 minutes a day with them. There must be a method error here. Hypothesis: Families were larger, and they've divided the "time with children" metric by the number of kids to come up with the plotted "time spent with child" metric? My mom's generation grew up with 4 siblings and a stay-at-home mom, there's no way they had less time with their mothers than my daughter does, but if we split my time by 5, this plot might look more reasonable.
|
# ? Oct 17, 2023 08:09 |
|
Hippie Hedgehog posted:There must be a method error here. The error come with the extrapolation from a handful of data points in some cases to two at the lowest. https://twitter.com/cheeze_squeeze/status/1713983577178947918?t=urGQQf6fmwU4ymfjOc5JRw&s=19
|
# ? Oct 17, 2023 12:11 |
|
Also there's a big difference between a 3 year old and a 10 year old, let alone teenagers. You can safely ignore the 10 year old if you need to and leave them with friends, but the 3 year old is gonna hurt themself or something before an hour has passed. If your low sample sizes also don't account for age, that's barely even worthwhile data.
|
# ? Oct 17, 2023 13:54 |
|
Love to look at charts meant to show a trend and see all the data points have overlapping confidence intervals before nodding at a job well done. Love to read in a meta analysis "research has confirmed our memories are not distorted." Definitely reviewing literature with an open mind.
|
# ? Oct 17, 2023 14:28 |
|
Thinking about stepping outside the office to yell at some researchers to come look at this poo poo.
|
# ? Oct 17, 2023 14:39 |
|
My parents chose to both "unschool" and practice "benevolent absence" when raising me and my two siblings. It's a miracle I know how to use a toilet.
|
# ? Oct 17, 2023 16:29 |
|
First of May posted:My parents chose to both "unschool" and practice "benevolent absence" when raising me and my two siblings. It's a miracle I know how to use a toilet. But do you know not to use an exponential regression for two data points?
|
# ? Oct 17, 2023 16:35 |
|
ikanreed posted:But do you know not to use an exponential regression for two data points? Oh yeah, guh-doi, obviously.
|
# ? Oct 17, 2023 16:39 |
|
First of May posted:It's a miracle I know how to use a toilet. Actually, we've been meaning to talk to you about that. Because what you're doing is, uh....it's weird.
|
# ? Oct 17, 2023 16:40 |
|
Captain Hygiene posted:Actually, we've been meaning to talk to you about that. Because what you're doing is, uh....it's weird. What are you, some kind of expert?
|
# ? Oct 17, 2023 17:36 |
|
Telegnostic posted:What are you, some kind of expert?
|
# ? Oct 17, 2023 17:44 |
|
Telegnostic posted:What are you, some kind of expert? Do you not see the user name? C'mon, son...
|
# ? Oct 17, 2023 17:52 |
|
|
# ? Oct 17, 2023 18:50 |
|
It's... some kind of chart, at least.
|
# ? Oct 18, 2023 20:13 |
|
I am a blind homo cockroach (Ask Me Anything)
|
# ? Oct 18, 2023 20:43 |
|
|
# ? May 28, 2024 04:10 |
|
Yeah, yeah, I'm gonna become a rat, whatever, but I've heard that there's apparently bad food in the sewers? Is it true?
|
# ? Oct 18, 2023 21:21 |