For each non-GR injection, perform a GR run to compare posteriors in the two cases.
Two subpopulations:
1. q~1, spins~0, face on, M~70 (reproduce known LIGO observed population);
Simulation parameters:
MISSING
2. q>1, spins>0, not face on, M \in [150, 250] (reproduce possible IMBH population);
Simulation parameters:
MISSING
10 injections for each subpopulation, each of the injections to be runned with SNR=15,30,50.
Different values of parameters for NR and GR-template injections, to explore a larger parameter space.
Prioritise given CPU hours constraints.
- GR-template injections: we expect to recover 0-centered posteriors on deviation parameters and all other parameters consistent with GR. Also useful to study correlations with the other parameters.
How many? Which population?
- NR injections: we expect to recover 0-centered posteriors on deviation parameters and all other parameters consistent with GR. Also useful to study correlations with the other parameters.
How many? Which population?
- Non-GR template injections: inject either pEOB+domega/=0 and/or a physically motivated model (e.g. a model with modified flux as in the IMR consistency test review).
How many? Which population?
Same parameters as the GR cases with non-zero (domega220, tau220) for the lower-mass population, non-zero (domega220, tau220, domega330, tau330) for the higher mass population.
**Waveform infrastructure call**
...
...
@@ -93,15 +105,21 @@ How many? Which population?
*Attendees*
Abhirup, Roland, Richard, Max, Gregorio
*Minutes*
*Action items*
- Abhirup to fill in info on how to run the test;
- Abhirup to fill in info on real events results;
- Reviewers to have a careful read of the [paper](https://journals.aps.org/prd/abstract/10.1103/PhysRevD.98.084038)
- ...
- Reviewers to fill in info on how to run the test;
- Reviewers to fill in info on real events results;
- Reviewers to have a careful read of the [paper](https://journals.aps.org/prd/abstract/10.1103/PhysRevD.98.084038) ;
- Write down a proposal on the injection set values along the lines discussed in the telecon;
- Reviewees to ask the TGR chairs how many CPU hours we can use;
- Abhirup to create:
1. Corner plots with (Mf,af,domega)
2. A plot showing (Mf,af) from a GR run vs the ones recovered when domega was varying.
(These plots are to check that a deviation in Mf does not compensate a deviation in domega in such a way that omega(Mf,af)*(1+domega) still matches the GR prediction)
