13:18:30	 From Anna S : I can be the host, yes maybe 4-5? We should maybe also tell them to talk with each other in those groups and discuss the answers?
13:18:41	 From Anna S : Sorry ;(
13:30:33	 From Kasper Hede Nielsen : Should expect uniform distribution of the numbers
13:30:48	 From Marta Mrozowska : I noticed when I was typing in the numbers, I typically typed 1-5 together and 6-0 together
13:31:08	 From Marta Mrozowska : because of how my keyboard is set up
13:31:43	 From Jaime Caballer Revenga : maybe operating with another true random number (irrational, e.g. pi, e), and see if the output shows a non-random pattern?
13:32:04	 From Liam : fourier transform
13:32:06	 From Laurits Høgel : You can check every 2 digit sequence
13:32:06	 From Kasper Hede Nielsen : For enough data you should expect runs longer than 1
13:32:22	 From Kasper Hede Nielsen : of each digit
13:32:46	 From Laurits Høgel : not second digit, but 2-digit-sequence
13:32:47	 From Emily Claire Winther Sørensen : check patterns to check if distribution of most probable next digit is also uniform
13:33:08	 From Kasper Hede Nielsen : 1/10 ?
13:33:48	 From Anna S : Isn't 3 in a row 1/1000?
13:34:25	 From Aske Luja Lehmann Rosted : feed all the different random digit sets to an ML algorithm that tries to predict the next number and see if it has any luck with any of them?
13:36:55	 From Troels Christian Petersen  To  Amalie Paulsen(privately) : Ikke i en gruppe?!?
13:37:21	 From Troels Christian Petersen : I’ll be back in a minute...
13:37:33	 From Amalie Paulsen  To  Troels Christian Petersen(privately) : Jeg kunne ikke lige logge på zoom, så jeg ved ikke lige hvad vi laver, men har vi regnetime nu? 
13:39:11	 From Troels Christian Petersen  To  Amalie Paulsen(privately) : Nej… stadig forelæsning… til 14::00-15
13:40:25	 From Efthymios Siamos : maybe compare the likelihood of the numbers ?
13:40:56	 From Hjalte : we talked about summing up numbers and seeing if they would be gaussian distributed
13:42:00	 From Katriona Mai Landau Gould : Sum all of the numbers in each column, if the numbers are randomly distributed the sums should give a uniform distribution, otherwise it may reveal certain orders people use e.g. switching between sides of the keyboard or moving around parts of the keyboard
13:42:06	 From lucaalessi : Shift the numbers in the sequence by 1 step and plot the shifted number sequence vs the original one
13:42:35	 From Marcus Nørgaard Weng : Very simple, but might catch something: Compare the averages of all the integers. For so many numbers, the averages should be very close to np.avg(range(10))
13:42:52	 From David : making a histogram of all tripplets (or larger groups) of numbers should look like the histogram of a uniformly distribution. If there are repeating sequences the histogram will be uneven, no?
13:43:26	 From Lasse Bonn : is there an approach where we consider the entropy/information of the data
13:44:20	 From Marta Mrozowska : Troels, Iminuit was updated 13 h ago and the new version makes your scripts not work; for erda users, install iminuit==1.5.4, then the scripts still work
13:46:50	 From Norman Pedersen : would it be unreasonable to filter out the drunk data ?
13:49:45	 From Marcus Nørgaard Weng : Do you think there is a correlation between "good"/precise measurements and low self-proclaimed uncertainties. Do people know when they made a precise measurement?
13:50:55	 From Marcus Nørgaard Weng : I was thinking about the situation where you've measured the table with the ruler and have to guesstimate your own uncertainty
13:53:51	 From Norman Pedersen : what if one has a data set of such size that manual inspection is not possible ?
13:54:01	 From jmlin : how to choose 3/4 sigma to reject data? is there any criteria or experience?
13:54:53	 From David John Willumsen : Does Chauvenet's criterion have to be applied to one point at a time, or can multiple points be compared to the same 1/2N?
13:55:06	 From Peter Andresen : This discarding of data assumes that it follows a normal distribution, but in reality that is not for sure right? So you are trimming the data to look like the distribution you expect?
13:59:33	 From David : about slide 15: could you repeat why there is no RMS value in this condition?
14:00:30	 From David : thanks
14:07:07	 From Peter Andresen : This discarding of data assumes that it follows a normal distribution, but in reality that is not for sure right? So you are trimming the data to look like the distribution you expect?
14:07:59	 From Peter Andresen : sounds good, thanks
14:08:15	 From Efthymios Siamos : so how can you be objective when peer reviewing others work ? 
14:08:47	 From Efthymios Siamos : I mean in the statistics as it seems each person does different statistic analysis