Choose timezone
Your profile timezone:
45FS Recommendation Meeting
→
Europe/Zurich
EVO
EVO
, , ,
Minutes of the 4-5-flavor cross section discussion meeting, April 7th 2011
==========================================================================
Present (among others, not all of them the whole time): Monica Vazquez, Chiara Mariotti, Reisaburo Tanaka, Giampiero Passarino, Michael Spira, Georg Weiglein, Sven Heinemeyer, Markus Schumacher, Robert Harlander, Sasha Nikitenko, Markus Klute, Sandra Horvat, Michael Krämer, Robert Thorne, Stefan Dittmaier, Jürgen Kroseberg, Koji Nakamura, Sally Dawson - my apologies to everybody I might have forgotten, I didn't look at EVO all time.
A special discussion session on a possible recommendation on inclusive cross sections for b-associated neutral Higgs boson production (important for current MSSM h/H/A->tautau analyses) was held.
A short introduction by Markus Warsinsky was given, where he reminded everybody that in the CERN Yellow Report the cross sections were given both for the 4-flavor scheme (with scale unc., but w/o PDF unc. - these are expected to be small compared to scale unc.) and for the 5-flavor scheme were given. For the Moriond analyses (which were utilizing no cuts on jets or b-tagging), the CMS results wre interpreted in the 5FS, while ATLAS used the 4FS to show exclusion contours in the (mA,tan(beta)) plane. The purpose of this meeting was set to examine wether one can come to a clear recommendation on which calculation or procedure to recommend from the LHC Higgs cross section group for the future.
Robert Harlander briefly discussed a document which he had prepared together with Michael Krämer and Markus Schumacher. In it he discussed the advantages and disadvantages of both calculation approaches. To summarize quickly: Large logarithms that appear in the 4FS calculation are resummed automatically in the 5FS calculation. It can thus be expected that the 4FS calculation underestimates the cross section for large Higgs boson masses. It should be noted that the error bands of both calculations overlap, especially at low masses. He then introduced a heuristic interpolation scheme between the two calculations, based on an expression that is logarithmic in M_H. This so-called "Santander matching" would give about equal weight to both calculations at MH=100 GeV and about 1/3 4FS and 2/3 5FS at MH=300 GeV.
This heuristic proposal was discussed. Questions came up on what would be the alternatives and in how far this proposal would be better than these. The alternatives would be to always show the limits with both schemes for calculation, or to use the complete envelope of the two calculations.
The first alternative (showing both curves) was discussed. Some criticism was made that this would result in overcrowded plots and that if one would make two plots one would still have to quote one as the central result.
Taking the envelope was in wide agreement among theorists deemed to be overly conservative, while the "Santander-matched" values are expected to be closer to the "true" result and the errors, though smaller than the envelope, would be close to the 4FS numbers. Michael Spira commented about this about that from the behaviour of the transverse mass distribution of the b quarks gives indications supporting the numerical results that will be obtained from the Santander matching. Some more details of this argument will be linked from the agenda page.
It was asked what the effect of 4FS vs. 5FS in the limit calculation is and it was mentioned that this will currently make a difference of about 10% in tan(beta) in the limit.
It was emphasised by Sasha Nikitenko that such a matchings scheme should not be enforced by the experimentalists, but rather based on the opinion of the invloved theorists that this procedure is scientifically sound and expected to give a better estimate. The involved theorists agreed widely to the second part and all confirmed that in their mind this is the best thing one can do at the moment.
Then Robert Thorne discussed a similar document in which he discussed another interpolation scheme. He also emphasised that similar interpolations are also made on a regular basis in the extraction of PDFs. The alternative scheme is based on estimating the resummed terms in the 5FS calculation and making a weight based on these. Robert emphasised that the numerical difference to the Santander-matching will be small and that he would also agree with this alternative. It was discussed that some of the coefficients used in the weighting also seem to be arbitrary and heuristic.
In the discussion it became clear that a clear preference for a simple, more heuristic matching scheme (i.e. the Santander-matching) was present among the invloved theorists.
Reisaburo Tanaka showed in one slide a possibility for a numerical definition of an envelope between the two calculations. It was discussed an again the feeling was that this is overly conservative.
The final agreement among all people connected was to recommend the Santander matching as a prescription for an inclusive cross section for b-associated Higgs boson production. The theory uncertainties should be scaled in the same way, but as they are asymmetric for the 5FS calculation, this should be done independently in both directions. In the weighting formula (given in the available document), the bottom pole mass should be used, and to be consistent with the used PDF (MSTW2008) one should use 4.75 GeV.
It was pointed out by Robert Thorne that in the future studies should be made on the uncertainty of the cross section in dependence of the bottom quark mass should be made. PDF sets with varying quark masses only became available after work for the CERN yellow report had begun, so this was not done for the YR. This is some area of future work, and as mentioned by Sven Heinemeyer one will have to consider also the branching fractions, as in the final sigma x BR, especially for the tautau final state, most of these uncertainties will drop out again.
In addition it was decided to recommend an upper value of tan(beta) to be shown in the future based on the region where the bottom Yukawa coupling becomes non-perturbative. It was decided to recommend showing only up to tan(beta) of 60.
Markus Warsinsky, April 13th 2011
==========================================================================
Present (among others, not all of them the whole time): Monica Vazquez, Chiara Mariotti, Reisaburo Tanaka, Giampiero Passarino, Michael Spira, Georg Weiglein, Sven Heinemeyer, Markus Schumacher, Robert Harlander, Sasha Nikitenko, Markus Klute, Sandra Horvat, Michael Krämer, Robert Thorne, Stefan Dittmaier, Jürgen Kroseberg, Koji Nakamura, Sally Dawson - my apologies to everybody I might have forgotten, I didn't look at EVO all time.
A special discussion session on a possible recommendation on inclusive cross sections for b-associated neutral Higgs boson production (important for current MSSM h/H/A->tautau analyses) was held.
A short introduction by Markus Warsinsky was given, where he reminded everybody that in the CERN Yellow Report the cross sections were given both for the 4-flavor scheme (with scale unc., but w/o PDF unc. - these are expected to be small compared to scale unc.) and for the 5-flavor scheme were given. For the Moriond analyses (which were utilizing no cuts on jets or b-tagging), the CMS results wre interpreted in the 5FS, while ATLAS used the 4FS to show exclusion contours in the (mA,tan(beta)) plane. The purpose of this meeting was set to examine wether one can come to a clear recommendation on which calculation or procedure to recommend from the LHC Higgs cross section group for the future.
Robert Harlander briefly discussed a document which he had prepared together with Michael Krämer and Markus Schumacher. In it he discussed the advantages and disadvantages of both calculation approaches. To summarize quickly: Large logarithms that appear in the 4FS calculation are resummed automatically in the 5FS calculation. It can thus be expected that the 4FS calculation underestimates the cross section for large Higgs boson masses. It should be noted that the error bands of both calculations overlap, especially at low masses. He then introduced a heuristic interpolation scheme between the two calculations, based on an expression that is logarithmic in M_H. This so-called "Santander matching" would give about equal weight to both calculations at MH=100 GeV and about 1/3 4FS and 2/3 5FS at MH=300 GeV.
This heuristic proposal was discussed. Questions came up on what would be the alternatives and in how far this proposal would be better than these. The alternatives would be to always show the limits with both schemes for calculation, or to use the complete envelope of the two calculations.
The first alternative (showing both curves) was discussed. Some criticism was made that this would result in overcrowded plots and that if one would make two plots one would still have to quote one as the central result.
Taking the envelope was in wide agreement among theorists deemed to be overly conservative, while the "Santander-matched" values are expected to be closer to the "true" result and the errors, though smaller than the envelope, would be close to the 4FS numbers. Michael Spira commented about this about that from the behaviour of the transverse mass distribution of the b quarks gives indications supporting the numerical results that will be obtained from the Santander matching. Some more details of this argument will be linked from the agenda page.
It was asked what the effect of 4FS vs. 5FS in the limit calculation is and it was mentioned that this will currently make a difference of about 10% in tan(beta) in the limit.
It was emphasised by Sasha Nikitenko that such a matchings scheme should not be enforced by the experimentalists, but rather based on the opinion of the invloved theorists that this procedure is scientifically sound and expected to give a better estimate. The involved theorists agreed widely to the second part and all confirmed that in their mind this is the best thing one can do at the moment.
Then Robert Thorne discussed a similar document in which he discussed another interpolation scheme. He also emphasised that similar interpolations are also made on a regular basis in the extraction of PDFs. The alternative scheme is based on estimating the resummed terms in the 5FS calculation and making a weight based on these. Robert emphasised that the numerical difference to the Santander-matching will be small and that he would also agree with this alternative. It was discussed that some of the coefficients used in the weighting also seem to be arbitrary and heuristic.
In the discussion it became clear that a clear preference for a simple, more heuristic matching scheme (i.e. the Santander-matching) was present among the invloved theorists.
Reisaburo Tanaka showed in one slide a possibility for a numerical definition of an envelope between the two calculations. It was discussed an again the feeling was that this is overly conservative.
The final agreement among all people connected was to recommend the Santander matching as a prescription for an inclusive cross section for b-associated Higgs boson production. The theory uncertainties should be scaled in the same way, but as they are asymmetric for the 5FS calculation, this should be done independently in both directions. In the weighting formula (given in the available document), the bottom pole mass should be used, and to be consistent with the used PDF (MSTW2008) one should use 4.75 GeV.
It was pointed out by Robert Thorne that in the future studies should be made on the uncertainty of the cross section in dependence of the bottom quark mass should be made. PDF sets with varying quark masses only became available after work for the CERN yellow report had begun, so this was not done for the YR. This is some area of future work, and as mentioned by Sven Heinemeyer one will have to consider also the branching fractions, as in the final sigma x BR, especially for the tautau final state, most of these uncertainties will drop out again.
In addition it was decided to recommend an upper value of tan(beta) to be shown in the future based on the region where the bottom Yukawa coupling becomes non-perturbative. It was decided to recommend showing only up to tan(beta) of 60.
Markus Warsinsky, April 13th 2011
There are minutes attached to this event.
Show them.
