Replies:

1) Group name:

Supplementary information:

• Description of required data samples: kinematic features relevant for triggering, limitations on online cuts (e.g. need for side-band regions, cuts that would give unacceptable bias etc.) control samples required etc.
• Plots of kinematics of physics channel e.g. p_T or E_T distribution of final state particles/jets.
• What aspects of current physics analyses are limited by the trigger:
  • Is there particular physics which can not be addressed because triggers are too tight?
  • Are there some systematic uncertainties which could be improved if the statistics of control samples were increased?

2) List of current triggers:

Possible supplimentary information:

• Details of offline selection cuts, comparison of online and offline cut values

3) Quantify the impact of increasing the thresholds of the above triggers to the level needed to maintain current trigger rates at a luminosity of 2x10³⁴. Please see the accompanying example menu for indicative L1 threshold values at 2x10³⁴. Are there any cases where it be preferable to apply a pre-scale factor or run the trigger as an end-of-fill trigger (e.g. cases that are sytematics rather than statistically limited).

Possible supplementary information:

• Plots of efficiency for signal channel as a function of threshold

4) Tightening of selection cuts and addition of new cuts: Based on a comparison of online and offline selection cuts, could you suggest cuts that could be tightened online? Are there cuts currently only applied offline, that could be moved online? (e.g. Isolation, Mass cuts, Decay length cuts, angular cuts, Primary vertex requirements, Secondary vertex requirements...)?

Supplementary information:

• Please supply a table or plot of rejection as a function of cut value.

5) Apart from raising thresholds, what additional/alternative strategies are there for controlling the rates of the triggers listed in (2) at a luminosity of 2x10³⁴? Would it be possible to replace single object primary triggers by multi-object triggers, or OR combinations of single and multi-object triggers? For example, could the lowest EM threshold be removed (or raised) with the addition of e+jet and e+met triggers.

Supplementary information:

• How would the threshold for the multi-object trigger be optimized? (e.g. relative priority to lepton threshold vs jet threshold for a l+jet(s) trigger, complementarity to single object trigger thresholds).
• What are the implications in terms of control samples? Do new triggers need to be implemented for those?

6) Topological triggers: Are there features of the topology of multi-object triggers that could be used in the trigger - in the HLT? Already at L1? i.e. triggers based on two or more trigger objects with e.g. mass cuts, angular requirements, scaler or vector ET or pT sums, etc.

Supplementary information :

• Please give tables or plots of rejection as a function of cut value.

7) Please compare the advantages and disadvantages of a strategy which prioritizes single-object triggers with raised thresholds(3) versus a menu which priorities multi-object triggers(5). Would you prefer one or other approach? To what extent are these approached complementary i.e. would you prefer a raised-threshold single-object trigger used in conjunction with a multi-object trigger (given the limit on total rates, thresholds would need to be higher than for using (3) or (5) alone). Are there any cases where it be preferable to apply a prescale factor or run the trigger as an end-off fill trigger?

-- EmanuelStrauss - 27-Sep-2012