Domenico Tiziani / ctools

Commit 1cfa265bc7a007f4b0ed228bf7acffaaf7672e8e

Authored by Jürgen Knödlseder
1 parent 350a888c

Add the shell width bias as a known issue.

Inline Side-by-side
Showing 1 changed file with 19 additions and 7 deletions
doc/source/issues.rst
  View file @1cfa265
... ... @@ -5,22 +5,23 @@ Known issues
5 5  
6 6 Below you will find a list of known issues.
7 7  
8   -- :ref:`Broken power law has unreliable errors <issues_bplaw>`
9   -- :ref:`Errors become unreliable when fitting the pivot energy <issues_pivot>`
10   -- :ref:`Binned analysis is biased when using coarse binning <issues_binned>`
  8 +- :ref:`Broken power law has unreliable errors <issue_bplaw>`
  9 +- :ref:`Errors become unreliable when fitting the pivot energy <issue_pivot>`
  10 +- :ref:`Binned analysis is biased when using coarse binning <issue_binned>`
  11 +- :ref:`Shell model is biased when width is comparable to angular resolution <issue_shell>`
11 12  
12 13  
13   -.. _issues_bplaw:
  14 +.. _issue_bplaw:
14 15  
15 16 .. topic:: Broken power law has unreliable errors
16 17  
17 18 The broken power law spectral model has unreliable errors, specifically
18 19 for the prefactor and the break value. Errors are in general too large,
19   - and this is related to the fact that the law's gradient is discontinous
  20 + and this is related to the fact that the law's gradient is discontinuous
20 21 in energy. There is not very much we can do about it, it's inherent in
21 22 the law.
22 23  
23   -.. _issues_pivot:
  24 +.. _issue_pivot:
24 25  
25 26 .. topic:: Errors become unreliable when fitting the pivot energy
26 27  
... ... @@ -34,7 +35,7 @@ Below you will find a list of known issues.
34 35 parameters of the model to assure non-degeneracy of the free
35 36 parameters.
36 37  
37   -.. _issues_binned:
  38 +.. _issue_binned:
38 39  
39 40 .. topic:: Binned analysis is biased when using coarse binning
40 41  
... ... @@ -45,3 +46,14 @@ Below you will find a list of known issues.
45 46 per pixel for the spatial binning and 10 bins per decade for the
46 47 spectral binning. If the binning is too coarse, the spectral parameters
47 48 that are fitted will be biased.
  49 +
  50 +.. _issue_shell:
  51 +
  52 +.. topic:: Shell model is biased when width is comparable to angular resolution
  53 +
  54 + When the width of the shell model becomes comparable to or smaller
  55 + than the angular resolution, the shell width tends to be overestimated
  56 + while the shell radius tends to be underestimated.
  57 + The fitted shell width and radius should thus not be overinterpreted
  58 + when the width is close to the angular resolution of CTA.
  59 +
... ...