run12 pp200 UE in unfolding
Updated on Thu, 2018-04-19 12:01. Originally created by veprbl on 2018-03-22 01:41.
Errata
- Last slide of iu_local_meeting_2018_03_22_v2.pdf showed all towers even those with 0 energy after the hadronic subtraction, this was fixed in iu_local_meeting_2018_03_22_v2_fixed.pdf
Hi all,
In the meeting I realized that people forgot what the VPDMB cut was for. Let me recap the logics of what is being done. We are looking to unfold VPDMB jet spectrum because that sample provides unbiased spectrum of the jets at low p_T. Usually, when we select events to construct the detector response matrix, we ask for triggerXXX->shouldFire(), but that is not implemented for VPDMB. We also know that hardware triggering condition is implemented as a cut on vpdtof value such that it should effectively require |vpdvz| < 30 cm. Again, the simulation of vpdtof/vpdvz is *not* implemented in our embedding (my best understanding is that we only simulate VPD hits). This is why we can't implement "shouldFire" condition for VPDMB directly. To overcome this we can impose what I call a "VPDMB cut".
This cut just needs to satisfy two main requirements:
1) VPDMB cut when applied to data and embedding has to pass events with the same jet properties (for example it should not skew R_T or UE p_T too much). This is a prerequisite for correct detector effect unfolding.
2) VPDMB cut must impose a stronger requirement than the VPDMB trigger condition. Similarly to requirement 1 this is also to prevent any biases that VPDMB trigger can introduce.
The claim is that the VPDMB cut, as presented on slide 2, is a good candidate for the role. In today's presentation and in my previous study [1] I tried to evaluate that claim. Requirement 1) from the above implies that sample of jet events satisfying "VPDMB NOBSMD did fire && JP0 should fire && VPDMB cut" should be comparable to a sample "JP0 did fire && VPDMB cut". That is what you can check by comparing leftmost and rightmost plots in slides [1] and this is why I compare blue and dashed red curves in today's slides. I hope this clears things up a bit.
Best,
Dmitry
[1] https://drupal.star.bnl.gov/STAR/system/files/vpdmb_integration.pdf
In the meeting I realized that people forgot what the VPDMB cut was for. Let me recap the logics of what is being done. We are looking to unfold VPDMB jet spectrum because that sample provides unbiased spectrum of the jets at low p_T. Usually, when we select events to construct the detector response matrix, we ask for triggerXXX->shouldFire(), but that is not implemented for VPDMB. We also know that hardware triggering condition is implemented as a cut on vpdtof value such that it should effectively require |vpdvz| < 30 cm. Again, the simulation of vpdtof/vpdvz is *not* implemented in our embedding (my best understanding is that we only simulate VPD hits). This is why we can't implement "shouldFire" condition for VPDMB directly. To overcome this we can impose what I call a "VPDMB cut".
This cut just needs to satisfy two main requirements:
1) VPDMB cut when applied to data and embedding has to pass events with the same jet properties (for example it should not skew R_T or UE p_T too much). This is a prerequisite for correct detector effect unfolding.
2) VPDMB cut must impose a stronger requirement than the VPDMB trigger condition. Similarly to requirement 1 this is also to prevent any biases that VPDMB trigger can introduce.
The claim is that the VPDMB cut, as presented on slide 2, is a good candidate for the role. In today's presentation and in my previous study [1] I tried to evaluate that claim. Requirement 1) from the above implies that sample of jet events satisfying "VPDMB NOBSMD did fire && JP0 should fire && VPDMB cut" should be comparable to a sample "JP0 did fire && VPDMB cut". That is what you can check by comparing leftmost and rightmost plots in slides [1] and this is why I compare blue and dashed red curves in today's slides. I hope this clears things up a bit.
Best,
Dmitry
[1] https://drupal.star.bnl.gov/STAR/system/files/vpdmb_integration.pdf
»
- veprbl's blog
- Login or register to post comments