[Halld_src] [EXTERNAL] ALP Analysis Note: Comments requested by next Monday

[Dipangkar Dutta]

Hi Jackson,
The analysis note is very detailed and clear. I also think it is
essentially ready for unblinding.
I just have a couple of questions which you can quickly answer during our
meeting today.

1. Could you briefly describe how the detection efficiency epsilon_a  (Fig
31 right panel) and epsilon_eta were determined? I guess the ratio of these
efficiencies is the primary factor determining the shape of Fig 32.

2. Please check if the narrow sharp feature at the very right edge of the
dark bands in Fig 36 and 37 are real.

I also second Axel's concern about downstream etas, and the suggested quick
test.

Cheers
Dipangkar




On Sun, Dec 11, 2022 at 10:20 PM Axel Schmidt via halld_src <
halld_src at jlab.org> wrote:

> Dear all,
> Thank you Jackson, for putting this incredibly detailed and well-written
> note together. I've finished going through it. Basically I have two
> concerns before we unblind. I hope you can address them easily. Let's
> discuss in tomorrow's meeting.
>
>
> 1) etas produced downstream
>
> My biggest fear is that we'll "discover" an ALP that is actually the peak
> from etas produced on some downstream material. Can you convince me that
> this cannot happen? I don't know the best way to study it but some ideas
> are:
> - Figure out the total thickness of any foils or windows after the target
> ladder and where they occur.  (I don't care about air, since it will not
> make a bump)
> - Simulate a fake signal with an eta mass, but a displaced vertex.
>
> Eventually this effect must limit our sensitivity to ALPs. I would be very
> satisfied if you could show where this limit ought to be, especially if
> it's 4 orders of magnitude from our expected sensitivity.
>
>
>
> 2) Effective # of searches
>
> My second concern is about your data-driven method of using up-crossings
> to estimate the effective number of independent searches. If I understand
> things correctly, the effective number of search regions should really only
> be a function of your resolution. Maybe the shape of the smooth background
> comes in as a second order effect, but really it's an effect of the width
> of any bumps you might find.
>
> However, for any given random data set, the precise number of up crossings
> might fluctuate due to statistics. With 10% of the data, we see 3
> up-crossings. When we unblind we expect to still see 3, plus or minus one.
> My question to you is: what should we do if we see many many more?
>
> In some sense, we could study this by looking at only 1% of the data. Or
> in simulation. But I don't want to waste time on this. I just would like to
> hear your thoughts. Is there anyway to do a sanity check on the full data
> set before we look for bumps? Would our seeing 27 up crossings invalidate a
> discovery?
>
>
>
>
>
> If you can give answers to these concerns, I would feel comfortable
> unblinding.
> Looking forward to discussing tomorrow,
> Axel
>
>
> ________________
>
> Typos to fix whenever you are bored in the next few months:
>
> Line 89
> ! GeV
>
> Line 106
> Three criteria vs four criteria
>
> Line 192
> a -> an
>
> ________________
>
> On Dec 5, 2022, at 15:00 , Jackson Reeves Pybus via halld_src <
> halld_src at jlab.org> wrote:
>
> Hello all,
>
> As discussed in today’s SRC-CT analysis meeting, I am now distributing our
> internal analysis note for the ongoing Axion-Like Particle search using our
> Carbon data. As we hope to unblind the data from our current 10% soon to
> examine our full statistics, we would ask that anyone with interest in the
> analysis read over the note and provide comments in the following week,
> before next SRC-CT meeting. Comments relevant to the decision to unblind
> are the priority at this time. Next week, an overall description of the
> analysis will be presented to the group.
>
> Any comments should be added to the following Google Doc, which will be
> addressed Monday:
>
> https://urldefense.proofpoint.com/v2/url?u=https-3A__docs.google.com_document_d_1LxsqEbn4JKX9AqlFElJ59TWd-2DN09to4w6Xj833Tv3cA_edit-3Fusp-3Dsharing&d=DwIFaQ&c=CJqEzB1piLOyyvZjb8YUQw&r=66Sb4pp0d4diwFuNQ_jTiKyZDsB_FG2X3Hlmk9_wuSY&m=vgHXnKtbu9dzjstZaSnaMJ-foYO3EnzZDK3SzGgencR38xSbz0-4X9o5f2o7XtEO&s=2v--PpuimTTRqKXK4b0uPHJx3UvOTGmAtxu0vEDZnjQ&e= 
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__docs.google.com_document_d_1LxsqEbn4JKX9AqlFElJ59TWd-2DN09to4w6Xj833Tv3cA_edit-3Fusp-3Dsharing&d=DwMFaQ&c=CJqEzB1piLOyyvZjb8YUQw&r=66Sb4pp0d4diwFuNQ_jTiKyZDsB_FG2X3Hlmk9_wuSY&m=N1TqBAlSCOi_HY3Ublw9Syo_-QbsiFIAlZoRAatnsA4yBIYMpkVO0TZaxmOD3aSv&s=C7WSuOX1eWH7BEYtQQMxCLtqL0Se2-6SaxEXuL1_b-0&e=>
>
> Thank you,
> Jackson
>
> <SRC_CT_Analysis_Note_for_ALP_Search.pdf>
> _______________________________________________
> halld_src mailing list
> halld_src at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld_src
>
>
> _______________________________________________
> halld_src mailing list
> halld_src at jlab.org
>
> https://urldefense.proofpoint.com/v2/url?u=https-3A__secure-2Dweb.cisco.com_12i0HzdbGeYBpx8E50ty9hk3Tbt1HekTPTHIG9yOnYkxe0qB9EwBvLSviRUWxgiHpl9g3c8VEwwm2fTQu-5FRhGUsv5Tqat2KmSad7GRa0YfFSFH32Bv0ZzPvlamLkKQZ7GF4Aqeas8ByXUJp-2DJea5Ed8rzJH59iLluyE-5FIdNA0fZeO3weN7B-2DlLIcp9jKUBFSVgUdr-5FUhmyWGrdLV6JOienBH016hWk6RAazkZvDYzg7gtNxJfM6Yos8GN-5FsIlGIRCY8mjDmVWsBxk2-5Fgo6eHxvspw-5FI1XvmutaeE2q06oWZVud-5FMNiM3g7hQJxQpZnIez_https-253A-252F-252Fmailman.jlab.org-252Fmailman-252Flistinfo-252Fhalld-5Fsrc&d=DwIFaQ&c=CJqEzB1piLOyyvZjb8YUQw&r=66Sb4pp0d4diwFuNQ_jTiKyZDsB_FG2X3Hlmk9_wuSY&m=vgHXnKtbu9dzjstZaSnaMJ-foYO3EnzZDK3SzGgencR38xSbz0-4X9o5f2o7XtEO&s=htvhIjj7C9tgu-z2iQVDd1KAgH-vfn2ZddOc-_ys-5I&e= 
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld_src/attachments/20221212/b95b8e6e/attachment.html>