[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: updated talk for PID
Hall D PID Mail List:
Hi Matt,
Thanks --
To be clear, I don't think your study was presented disingenuously you
have certainly fully qualified your assumptions. This has been a hot
topic of interest on peoples minds for almost a year now, and it is
great to see things rolling. My beef with what was written in the
document was that it presents very solid (extraordinary) quantitative
numbers for what your more ideal case. It then concludes with a
softer statement about backgrounds and the possibility of doing
strangeonium physics with GlueX.
It is hard to extrapolate those extraordinary to a realistic case
since we just have one data point. I think the "right answer" is
somewhere in between what you and Eugene have done. If, say the final
answer is a S/B of 3/1 (much better than Eugene, but more of the scale
we are talking) instead of 1000/1 as presented in your ideal case,
then what can one concretely conclude other than ideal is really
ideal? The doc as is now doesn't have Eugene's other end of the
spectrum. The trick is getting the right balance now during the final
moments of writing, especially when are are certain that we will not
have the right answer in time to put in the document -- those studies
to break away from the ideal case just won't be done in the next 24
hours.
(Fortunately there is a week to "fine-tune" this message in the talks.)
Let's give Curtis a chance to have his beer and nuts and see where we
stand.
-Matt
On Mar 19, 2008, at 6:34 PM, Matt wrote:
> Hi all,
>
> The results from the pi/K separation study I did were never
> meant to imply that we do not need additional PID. And every time I
> presented these results I have emphasized that there is other
> background that was not considered that could leak in. Also, that we
> are working from an idealized case where all tracks are
> reconstructed (4-C fit) and the errors are known perfectly.
>
> The goal was to show what we start with in an ideal world and go
> from there. That's where most of these types of studies start. I
> completely agree that in the real world we will never see quite that
> performance.
>
> However, I feel that Eugene's talk completely neglects a huge
> amount of effort that has gone into developing a legitimate
> kinematic fitter and other PID tools. If the goal of these talks is
> to present to the review committee where we stand, then we are
> selling ourselves short by not discussing this other work. The
> kinematic fitter that Eugene mentions is, by his own admission, not
> a fully functioning fitter....though there is one which we have
> checked into our repository and which has been used by both myself
> and the IU group (for 2-photon fits).
>
> I hope that no one feels that my study was presented
> disingenuously. I have always tried to stress the limitations of
> what I have done just as much, if not more so, than whatever results
> I have shown. I think the fact that we have started to incorporate
> timing information as part of our global event fitting is a *huge*
> step forward....even for perfect world cases and it shows where we
> plan to go for PID.
>
> I am just concerned that Eugene's talk misses out on at least
> mentioning where some of our efforts have been and paints a more
> pessimistic view of the situation than may actually be.
>
>
> Matt
>
>
>
> On Wed, Mar 19, 2008 at 5:54 PM, Matthew Shepherd <mashephe@indiana.edu
> > wrote:
>
> Hi Curtis,
>
> I strongly agree with backing off those numbers in the document.
>
> My concern is that you are presenting only the good part of the full
> story. We wouldn't want to give the impression that additional PID is
> unnecessary because we've already solved the problem with kinematic
> fitting. There is not enough time to fully understand the rest of the
> story regarding which backgrounds leak in. I think this means you
> should avoid such quantitative statements in the text.
>
> Perhaps there are less provocative aspects of Matt's work that could
> be presented to demonstrate it is something we are hard at work it?
> Showing nearly perfect signal purity with 50% efficiency in the "key
> channel for strangeonium hybrids" would be great ammo for shooting
> down any claim we need supplementary PID.
>
> -Matt
>
>
> On Mar 19, 2008, at 5:47 PM, Curtis A. Meyer wrote:
>
> > Hi Matt (Shepherd),
> >
> > yes, they do seem to imply that with kinematic fitting and
> > strangeness
> > conservation, we are done. I do not believe this. I suspect that
> if we
> > were to include channels that had hyperons (Lambda, Sigma), and a
> > single kaon, we would have more problems. I would guess that if we
> > had a missing neutron (or proton), the 1-C fits would not be nearly
> > so powerful.
> >
> > What I take from these is that kinematic fitting will help us a
> lot
> > (every experiment I have done has seen this), but even though I
> > put the numbers in the document, I think I want to back off on this
> > at the moment until we have a better understanding, or at least
> > recast Matt Bellis's work as a hypothical??
> >
> > Opionions anyone?
> >
> > thanks -- Curtis
> > On Wed March 19 2008, Matthew Shepherd wrote:
> >> Hall D PID Mail List:
> >>
> >>
> >> Hi Eugene and Matt,
> >>
> >> I was also struck by this passage in the document and the studies
> >> that
> >> Matt did. I don't doubt the power of kinematic fitting, but these
> >> results are quite surprising. In fact, taken at face value, they
> >> would say forget any other PID -- we don't need it. For some of
> the
> >> high multiplicity all neutral channels we were happy with S:B of
> >> about
> >> 5:1 and 10% efficiency. You have strange channels with S:B =
> >> infinity
> >> and ~50% efficiency -- that's quite good.
> >>
> >> I can't help but wonder what the background will be from broken
> >> events. For example, if you have an event where miss a pion you
> can
> >> make another one or two pions in the event a kaon to recover some
> of
> >> the lost four momentum.
> >>
> >> Matt
> >>
> >> On Mar 19, 2008, at 5:25 PM, Eugene Chudakov wrote:
> >>
> >>> Hall D PID Mail List:
> >>>
> >>> Matt,
> >>>
> >>> thanks for sending me your talk on the event fitting.
> >>> For my current studies I (in fact) did not use any fitting. I just
> >>> selected particle combinations, balancing the initial and final 3-
> >>> momentum and
> >>> the energy. Again, it was very simple: I selected the same
> absolute
> >>> cuts
> >>> in GeV for all events, without calculating the individual
> >>> covariances for each event.
> >>> Therefore, there must be a room for improvement.
> >>> I am sure you have already perfected this method.
> >>>
> >>> If I did any new step with this study, it is considering the
> minimum
> >>> bias
> >>> background from PYTHIA.
> >>> The pion suppression is only a factor of 0.4-0.6 for a kaon
> >>> candidate,
> >>> the overall suppression is about 0.2. The proton PID gives another
> >>> factor of 4.
> >>> Indeed, a factor of 3 comes from combinatorics, while some events
> >>> may have
> >>> no proton as well.
> >>>
> >>> Eugene
> >>>
> >>> ------------------------------------------------------
> >>> Eugene Chudakov
> >>> http://www.jlab.org/~gen
> >>> phone (757) 269 6959 fax (757) 269 5703
> >>> Thomas Jefferson National Accelerator Facility
> >>> 12000 Jefferson Ave, Newport News, VA 23606 USA
> >>>
> >>> On Wed, 19 Mar 2008, Matt wrote:
> >>>
> >>>> Hi Eugene,
> >>>>
> >>>> The kinematic fit mentioned in your talk, is that the fitter
> >>>> that I
> >>>> checked into the repository under $HALLD_HOME/src/libraries/PID?
> >>>> Or a
> >>>> different one? I see you're doing a different study than what I
> >>>> did, but I
> >>>> found that using time-of-flight info from the BCAL and forward
> TOF
> >>>> got you
> >>>> quite far in trying to do kaon physics, though I had a simpler
> >>>> physics
> >>>> background than the full pythia spectrum.
> >>>>
> >>>> Matt
> >>>>
> >>>>
> >>>> On Wed, Mar 19, 2008 at 3:54 PM, Eugene Chudakov gen <gen@jlab.org
> >
> >>>> wrote:
> >>>>
> >>>>> Hall D PID Mail List:
> >>>>>
> >>>>>
> >>>>> I updated the slides, including new pictures and adding more
> info.
> >>>>>
> >>>>> It can be found at:
> >>>>> http://www.jlab.org/~gen/gluex/talk_pid_rev.pdf<http://www.jlab.org/%7Egen/gluex/talk_pid_rev.pdf
> >>>>>>
> >>>>>
> >>>>> The talk source is located at
> >>>>> jlabl1:/home/gen/tex/GLUEX/talk_pid_march_2008.tex
> >>>>> The pictures are stored at
> >>>>> jlabl1:/home/gen/tex/GLUEX/pictures/plot_glx_*.pdf
> >>>>> For each picture there are also .epsi and .ps files.
> >>>>>
> >>>>> Eugene
> >>>>>
> >>>>> ----------------------------
> >>>>> Eugene Chudakov
> >>>>> JLab
> >>>>>
> >>>>>
> >>>>
> >>>>
> >>>> --
> >>>> --
> >>>> ----------------------------
> >>>> Matt Bellis
> >>>> Carnegie Mellon University
> >>>> (office) 412-268-6949
> >>>> (cell) 412-310-4586
> >>>> ----------------------------
> >>>>
> >>>
> >>
> >
> >
> >
> > --
> > Professor Curtis A. Meyer Department of Physics
> > Phone: (412) 268-2745 Carnegie Mellon University
> > Fax: (412) 681-0648 Pittsburgh PA 15213-3890
> > cmeyer@ernest.phys.cmu.edu http://www.curtismeyer.com/
>
>
>
>
> --
> --
> ----------------------------
> Matt Bellis
> Carnegie Mellon University
> (office) 412-268-6949
> (cell) 412-310-4586
> ----------------------------