09:19:49	 From Mirjam Partovi Dilami : When do we get feedback ?
09:24:00	 From Jakob Riber Rasmussen : The what about error propagation formulas, integrals and so on? Should we state those?
09:24:42	 From Jakob Riber Rasmussen : Ah yeah, thanks
09:30:10	 From Markus : When you say a value for the separation do you mean a two sample z test?
09:30:55	 From Markus : Thank you
09:31:52	 From Peter Andresen : In the problem set, wrote in Latex, which was fine but time consuming. Do you have a suggestion on how to present the exam? :) 
09:32:57	 From Peter Andresen : Cool, thanks!
09:33:55	 From Sebastian Øllegaard Utecht : You can use the markdown cells in jupyter and hide the code cells. Makes it a bit more coherent to write. It was faster for me that way
09:33:55	 From Amalie Paulsen : that would be very nice ;) 
09:36:37	 From Kiril Klein : Is it sufficient to give the answers in a table, like the one you just showed, without writing additional explanations?
09:37:41	 From Kiril Klein : Ok, thanks
09:37:53	 From Peter Andresen : Could you perhaps explain why a p-value of 0.999 for instance is a big problem?'
09:38:37	 From Qinyi Chen : Too good to be true
09:39:20	 From Peter Andresen : Makes sense, but the conclusion would still be that the model fits the data very well, given the uncertainties?
09:43:23	 From Peter Andresen : could you also say that the probability of not hitting should be less than 10%?
09:49:58	 From Peter Andresen : Do you need uncertainty on which d gives the lowest uncertainty?
09:51:44	 From Sofie Castro Holbæk : Just to make sure. If you have a mean, a standard error on the mean and a std, to calculate how many sigmas something is away you use the std?
09:56:30	 From Steen Bender : Is there away to see that you should use double gaussian? or is that just trial and error? 
09:57:42	 From Sebastian Øllegaard Utecht : What about when you propagate error and one of the variables is achieved through meaning some data. Should you use the error on the mean for that propagation or the std?
09:58:40	 From Peter Andresen : Is there an easy way to get an error on a calculated std?
10:00:00	 From Peter Andresen : Is an error on the std not required in for instance the Monte Carlo problem? :) 
10:00:10	 From Camilla : The limits there are different than the ones we were given right?
10:00:27	 From Nicolai Ree : The function here is also wrong
10:00:43	 From Camilla : Cool, thanks :-)
10:00:53	 From Camilla : Thought it looked different :-)
10:01:00	 From Steen Bender : There must be something i have missed.. When did RMS go from Root Mean Squared to become Std? 
10:02:06	 From Jakob Riber Rasmussen : And I guess the "resolution" also describes this right?
10:03:10	 From Peter Andresen : How do you get an error on this?
10:05:03	 From Jakob Riber Rasmussen : Should we in general put error bars on all histograms?
10:08:12	 From Sarah Jane Stapleton : In qu. 5, what would you say about the chi square when it is too large?
10:09:14	 From Sarah Jane Stapleton : Yep, thanks
10:14:16	 From Bia Fonseca : are we getting our problem sets back with comments?
10:17:05	 From Mikkel Liisberg : Will it be okay to submit multiple code-files, e.g., for each problem-section?
10:17:42	 From Mikkel Liisberg : Cool, thanks
10:18:02	 From Sebastian Øllegaard Utecht : Are we allowed to use stuff from the code library that you have given us? I.e. the external functions?
10:18:51	 From Nikolay Georgiev : where’s the x axis label? :D
10:19:02	 From Mads Engstrøm : What is the 3 yellow problem?
10:21:17	 From Zuzana Moravcova : 3rd yellow is 3.1.5
10:21:32	 From Mads Engstrøm : thank you :)
10:21:52	 From Mathias Labonte : What does an 80 correspond to on the 12 pt scale?
10:23:13	 From Mikkel Liisberg : Can you ask about the grading of a specific problem?
10:24:53	 From Zuzana Moravcova : Sorry, you’re not getting back your problem sets with comments. But if you believe you should get more points for a specific problem (based on the table and the problem set you uploaded), feel free to get back to us