| ... | @@ -22,7 +22,7 @@ Main results review page: https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3 |
... | @@ -22,7 +22,7 @@ Main results review page: https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3 |
|
|
|No.| NAME | Comments on config ini file|Comments on results | Reviewees comments |Final status|
|
|
|No.| NAME | Comments on config ini file|Comments on results | Reviewees comments |Final status|
|
|
|
|---| ---- | ---- | ---- | --- |--- |
|
|
|---| ---- | ---- | ---- | --- |--- |
|
|
|
|1. |[GW150914](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW150914) <br> person responsible: Krishnendu| |<br>Anuradha: No comparison-with-prod-run-plots, no ks plots.<br> ks_pos/ks_sym: Why is log L different? why a1 and a2 are oppositely unconstrained? Why is optimal (or matched-filter) SNR so different than BBH case? <br> Anuradha: <br> BBH: Why are ra, dec, DL, MF-SNR, OP-SNR are different? Is comparison with L1 samples? <br> ks_pos/ks_sym: Why are ra, dec, DL differnt? ks plots missing.|<br>Krishnendu: Comparison pages added now <br> As Anuradha pointed out the coherence test was not done on this event and there are still some chains found to be not-converged. <br> rerun would be necessary and started these runs| Runs are over. Results ready for review|
|
|
|1. |[GW150914](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW150914) <br> person responsible: Krishnendu| |<br>Anuradha: No comparison-with-prod-run-plots, no ks plots.<br> ks_pos/ks_sym: Why is log L different? why a1 and a2 are oppositely unconstrained? Why is optimal (or matched-filter) SNR so different than BBH case? <br> Anuradha: <br> BBH: Why are ra, dec, DL, MF-SNR, OP-SNR are different? Is comparison with L1 samples? <br> ks_pos/ks_sym: Why are ra, dec, DL differnt? ks plots missing.|<br>Krishnendu: Comparison pages added now <br> As Anuradha pointed out the coherence test was not done on this event and there are still some chains found to be not-converged. <br> rerun would be necessary and started these runs| Runs are over. Results ready for review|
|
|
|
|2. |[GW151226](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW151226) <br> person responsible: Krishnendu| |<br>Anuradha: No comparison-with-prod-run-plots, no ks plots.<br> ks_pos/ks_sym: Why is log L very very different? why a1 and a2 are totally unconstrained? Why is optimal (or matched-filter) SNR so different than BBH case? Mc, Mtot and m1 are railing at the boundary. <br> <br> Archisman: spin angles are different (generic feature); mtot railing. <br><br> Anuradha: Log is different for BBH and prod runs. Please confirm the Bayes factors are similar.|<br>Krishnendu: Comparison pages added now <br><br> Krishnendu (on the other comments): <br> [compare mtotal_source and mtotal_detector](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW151226_G211117/pv2_c01_prod/comp_PE_Prod0_cbcBayesCompPos/bh_bh/) <br> [config](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW151226_G211117/config/GW151226_C02_IMRPv2_bh.ini) <br> mtotal should not just bounded bcz of the mass prior ```distance-max=1500 chirpmass-min=9.5 chirpmass-max=10.5 q-min=0.0555555555555556 q-max=1.0 comp-max=54.398 comp-min=3.022 ```` <br> Any suggestion on the rerun settings?|Rerun? Need to decide|
|
|
|2. |[GW151226](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW151226) <br> person responsible: Krishnendu| |<br>Anuradha: No comparison-with-prod-run-plots, no ks plots.<br> ks_pos/ks_sym: Why is log L very very different? why a1 and a2 are totally unconstrained? Why is optimal (or matched-filter) SNR so different than BBH case? Mc, Mtot and m1 are railing at the boundary. <br> <br> Archisman: spin angles are different (generic feature); mtot railing. <br><br> Anuradha: Log is different for BBH and prod runs. Please confirm the Bayes factors are similar.|<br>Krishnendu: Comparison pages added now <br><br> Krishnendu (on the other comments): <br> [compare mtotal_source and mtotal_detector](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW151226_G211117/pv2_c01_prod/comp_PE_Prod0_cbcBayesCompPos/bh_bh/) <br> [config](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW151226_G211117/config/GW151226_C02_IMRPv2_bh.ini) <br> mtotal should not just bounded bcz of the mass prior ```distance-max=1500 chirpmass-min=9.5 chirpmass-max=10.5 q-min=0.0555555555555556 q-max=1.0 comp-max=54.398 comp-min=3.022 ```` <br> Any suggestion on the rerun settings? <br><br> BF from our BH run 49.71 and from the [PEPROD run 51.072](https://ldas-jobs.ligo.caltech.edu/~christopher.berry/O1/PE/G211117/C02/IMRP/lalinferencenest/IMRPhenomPv2pseudoFourPN/1135136350.65-211117/H1L1/posplots.html) |Rerun? Need to decide|
|
|
|
|3. |[GW170104](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW170104) <br> person responsible: Krishnendu| |<br>Anuradha: No comparison-with-prod-run-plots, no ks plots. <br> ks_pos: Why is log L different? Why a1 and a2 are totally unconstrained? <br> ks_sym: DL and q go beyond prior limits. Why Mc is so different? <br> <br> Archisman: does distance hit the upper bound or is that just the shape of the posterior? |<br>Krishnendu: Comparison pages added now <br><br> No distance prior issue on the [BH run](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW170104_G268556/pv2_c01_prod/comp_PE_Prod0_cbcBayesCompPos/bh_ks_pos/) <br> When we compare the BH and ks_sym case [here](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW170104_G268556/pv2_c01_prod/comp_PE_Prod0_cbcBayesCompPos/bh_ks_sym/), when we include ks distance, mchirp (hence mtotal, m1, m2), chi_eff are over-estimated and both optimal and matched filter SNR are peaked to a lower value <br> [base config](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW170104_G268556/config/GW170104_C02_IMRPv2.ini) uses no distance prior hence the sharp boundary at the right end of the distance posterior should also coming from the introduction of this extra parameter <br><br> these effects are not present with ks_sym <br><br> rerun only needed for [ks_sym case](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW170104_G268556/pv2_c01_prod/post-plots/)? |Results ready for review|
|
|
|3. |[GW170104](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW170104) <br> person responsible: Krishnendu| |<br>Anuradha: No comparison-with-prod-run-plots, no ks plots. <br> ks_pos: Why is log L different? Why a1 and a2 are totally unconstrained? <br> ks_sym: DL and q go beyond prior limits. Why Mc is so different? <br> <br> Archisman: does distance hit the upper bound or is that just the shape of the posterior? |<br>Krishnendu: Comparison pages added now <br><br> No distance prior issue on the [BH run](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW170104_G268556/pv2_c01_prod/comp_PE_Prod0_cbcBayesCompPos/bh_ks_pos/) <br> When we compare the BH and ks_sym case [here](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW170104_G268556/pv2_c01_prod/comp_PE_Prod0_cbcBayesCompPos/bh_ks_sym/), when we include ks distance, mchirp (hence mtotal, m1, m2), chi_eff are over-estimated and both optimal and matched filter SNR are peaked to a lower value <br> [base config](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW170104_G268556/config/GW170104_C02_IMRPv2.ini) uses no distance prior hence the sharp boundary at the right end of the distance posterior should also coming from the introduction of this extra parameter <br><br> these effects are not present with ks_sym <br><br> rerun only needed for [ks_sym case](https://ldas-jobs.ligo.caltech.edu/~nv.krishnendu/LSC/lalinference/O3/o3_a_tgr/O1_O2/GW170104_G268556/pv2_c01_prod/post-plots/)? |Results ready for review|
|
|
|
|4. |[GW170608](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW170608)<br> person responsible: Krishnendu | |<br> Anuradha: <br> BBH: Why log L is different? <br>ks_pos: Why log L is different? a2 is unconstrained. m1, m2, and mtotal are railing at the boundary. 1D ks plot is missing on the posplots.html page. <br>ks_sym: Why log L is different? a2 is unconstrained. m1, m2 and mtotal are railing at the boundary. 1D ks plot is missing on the posplots.html page. <br> <br> Archisman: spin angles are different (generic); mtot severely railing. | Saleem: There is no Mtot prior imposed in the config and there is no railing too (it appears to have so due to the small number of bins). The same applies for m1 and m2 railing as well. The priors on m1 and m2 are [1, 40] and the so-called railing is way too inside this limits. So, a rerun may not be required. <br> <br> Compared to the PE ini, the mc prior is wider for siqm: [8,9] vs [5,20], with the latter being the same as that used for TIGER. For the given mass ratio prior, this allows wider range of mtot to be sampled. <br> <br> The 1D pdf is available in the pesummary page which is linked from the event page.|Results ready for review|
|
|
|4. |[GW170608](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW170608)<br> person responsible: Krishnendu | |<br> Anuradha: <br> BBH: Why log L is different? <br>ks_pos: Why log L is different? a2 is unconstrained. m1, m2, and mtotal are railing at the boundary. 1D ks plot is missing on the posplots.html page. <br>ks_sym: Why log L is different? a2 is unconstrained. m1, m2 and mtotal are railing at the boundary. 1D ks plot is missing on the posplots.html page. <br> <br> Archisman: spin angles are different (generic); mtot severely railing. | Saleem: There is no Mtot prior imposed in the config and there is no railing too (it appears to have so due to the small number of bins). The same applies for m1 and m2 railing as well. The priors on m1 and m2 are [1, 40] and the so-called railing is way too inside this limits. So, a rerun may not be required. <br> <br> Compared to the PE ini, the mc prior is wider for siqm: [8,9] vs [5,20], with the latter being the same as that used for TIGER. For the given mass ratio prior, this allows wider range of mtot to be sampled. <br> <br> The 1D pdf is available in the pesummary page which is linked from the event page.|Results ready for review|
|
|
|
|5. |[GW170814](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW170814)<br> person responsible: Krishnendu | |<br>Anuradha: No comparison-with-prod-run-plots, no ks plots.<br> ks_pos/ks_sym: Why a1 and a2 are totally unconstrained? <br> <br> Archisman: spin angles are different (generic); mtot is fine. Why is dL so different? Why are sky angles not estimated in the SIQM run? Was a 2 detector run launched by mistake? <br> Anuradha: same question about DL and sky angles. Ks plots missing. |<br>Krishnendu: Comparison pages added now|Results ready for review|
|
|
|5. |[GW170814](https://git.ligo.org/nv.krishnendu/lalsuite/-/wikis/O3a-Spin-Quadrupole-BH-mimicker-test-results/GW170814)<br> person responsible: Krishnendu | |<br>Anuradha: No comparison-with-prod-run-plots, no ks plots.<br> ks_pos/ks_sym: Why a1 and a2 are totally unconstrained? <br> <br> Archisman: spin angles are different (generic); mtot is fine. Why is dL so different? Why are sky angles not estimated in the SIQM run? Was a 2 detector run launched by mistake? <br> Anuradha: same question about DL and sky angles. Ks plots missing. |<br>Krishnendu: Comparison pages added now|Results ready for review|
|
| ... | | ... | |
