Towards Jetography. Gavin Salam. LPTHE, CNRS and UPMC (Univ. Paris 6)

Transcription

1 Towards Jetography Gavin Salam LPTHE, CNRS and UPMC (Univ. Paris 6) Based on work with Jon Butterworth, Matteo Cacciari, Mrinal Dasgupta, Adam Davison, Lorenzo Magnea, Juan Rojo, Mathieu Rubin & Gregory Soyez C. N. Yang Institute for Theoretical Physics Stony Brook 9 February 2010

2 Towards Jetography, G. Salam (p. 2) Introduction Startup (again) for LHC

3 Towards Jetography, G. Salam (p. 2) Introduction Startup (again) for LHC October 2009

4 Towards Jetography, G. Salam (p. 2) Introduction Startup (again) for LHC 7 November: first beam (picture: CMS) October 2009

5 Towards Jetography, G. Salam (p. 2) Introduction Startup (again) for LHC December: 10 6 collisions at October s = 900 GeV2009

6 Towards Jetography, G. Salam (p. 2) Introduction Startup (again) for LHC December: collisions at s = October 2360 GeV2009

7 Towards Jetography, G. Salam (p. 3) Introduction Parton fragmentation quark Gluon emission: de dθ α s E θ 1 At low scales: α s 1

8 Towards Jetography, G. Salam (p. 3) Introduction Parton fragmentation quark θ gluon Gluon emission: de dθ α s E θ 1 At low scales: α s 1

9 Towards Jetography, G. Salam (p. 3) Introduction Parton fragmentation quark Gluon emission: de dθ α s E θ 1 At low scales: α s 1

10 Towards Jetography, G. Salam (p. 3) Introduction Parton fragmentation quark non perturbative hadronisation Gluon emission: de dθ α s E θ 1 At low scales: α s 1

11 Towards Jetography, G. Salam (p. 3) Introduction Parton fragmentation quark non perturbative hadronisation π + K L π 0 K + π Gluon emission: de dθ α s E θ 1 At low scales: α s 1

12 Towards Jetography, G. Salam (p. 3) Introduction Parton fragmentation quark non perturbative hadronisation π + K L π 0 K + π Gluon emission: de dθ α s E θ 1 At low scales: α s 1 This is a jet

13 Towards Jetography, G. Salam (p. 4) Introduction Seeing v. defining jets Jets are what we see. Clearly(?) 2 jets here How many jets do you see? Do you really want to ask yourself this question for 10 9 events?

14 Towards Jetography, G. Salam (p. 4) Introduction Seeing v. defining jets q q Jets are what we see. Clearly(?) 2 jets here How many jets do you see? Do you really want to ask yourself this question for 10 9 events?

15 Towards Jetography, G. Salam (p. 4) Introduction Seeing v. defining jets Jets are what we see. Clearly(?) 2 jets here How many jets do you see? Do you really want to ask yourself this question for 10 9 events?

16 Towards Jetography, G. Salam (p. 4) Introduction Seeing v. defining jets Jets are what we see. Clearly(?) 2 jets here How many jets do you see? Do you really want to ask yourself this question for 10 9 events?

17 Towards Jetography, G. Salam (p. 4) Introduction Seeing v. defining jets Jets are what we see. Clearly(?) 2 jets here How many jets do you see? Do you really want to ask yourself this question for 10 9 events?

18 Towards Jetography, G. Salam (p. 4) Introduction Seeing v. defining jets Jets are what we see. Clearly(?) 2 jets here How many jets do you see? Do you really want to ask yourself this question for 10 9 events?

19

20 Towards Jetography, G. Salam (p. 6) Introduction Jets as projections p π π φ K LO partons NLO partons parton shower hadron level Jet Def n Jet Def n Jet Def n Jet Def n jet 1 jet 2 jet 1 jet 2 jet 1 jet 2 jet 1 jet 2 Projection to jets should be resilient to QCD effects

21 Towards Jetography, G. Salam (p. 7) Introduction QCD jets flowchart Jet (definitions) provide central link between expt., theory and theory And jets are an input to almost all analyses

22 Towards Jetography, G. Salam (p. 7) Introduction QCD jets flowchart Jet (definitions) provide central link between expt., theory and theory And jets are an input to almost all analyses

23 Towards Jetography, G. Salam (p. 8) Introduction Talk Structure The different kinds of jet algorithm The historical problems with them ( Snowmass criteria ) and some of the solutions Speed, infrared safety Understanding the physics of jet algorithms the momentum of a jet v. the momentum of a parton Doing better physics with jets Dijet mass reconstruction Low-mass Higgs-boson search

24 Towards Jetography, G. Salam (p. 9) Two broad classes What jet algorithms are out there? 2 broad classes: 1. sequential recombination bottom up, e.g. k t, preferred by many theorists 2. cone type top down, preferred by many experimenters

25 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(kti 2,k2 tj ) R2 ij /R2 and d ib = ki 2 Recombine i,j (if ib: i jet) Repeat Bottom-up jets: Sequential recombination (attempt to invert QCD NB: hadron branching) collider variables Rij 2 = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle

26 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(k 2 ti,k2 tj ) R2 ij /R2 and d ib = k 2 i Recombine i,j (if ib: i jet) Repeat NB: hadron collider variables R 2 ij = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle

27 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(k 2 ti,k2 tj ) R2 ij /R2 and d ib = k 2 i Recombine i,j (if ib: i jet) Repeat NB: hadron collider variables R 2 ij = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle

28 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(k 2 ti,k2 tj ) R2 ij /R2 and d ib = k 2 i Recombine i,j (if ib: i jet) Repeat NB: hadron collider variables R 2 ij = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle

29 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(k 2 ti,k2 tj ) R2 ij /R2 and d ib = k 2 i Recombine i,j (if ib: i jet) Repeat NB: hadron collider variables R 2 ij = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle

30 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(k 2 ti,k2 tj ) R2 ij /R2 and d ib = k 2 i Recombine i,j (if ib: i jet) Repeat NB: hadron collider variables R 2 ij = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle

31 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(k 2 ti,k2 tj ) R2 ij /R2 and d ib = k 2 i Recombine i,j (if ib: i jet) Repeat NB: hadron collider variables R 2 ij = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle

32 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(k 2 ti,k2 tj ) R2 ij /R2 and d ib = k 2 i Recombine i,j (if ib: i jet) Repeat R ij > R NB: hadron collider variables R 2 ij = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle

33 Towards Jetography, G. Salam (p. 10) Two broad classes Sequential recombination algorithms k t algorithm Catani, Dokshizter, Olsson, Seymour, Turnock, Webber Ellis, Soper 93 Find smallest of all d ij = min(k 2 ti,k2 tj ) R2 ij /R2 and d ib = k 2 i Recombine i,j (if ib: i jet) Repeat R ij > R NB: hadron collider variables R 2 ij = (φ i φ j ) 2 + (y i y j ) 2 rapidity y i = 1 2 ln E i+p zi E i p zi R ij is boost invariant angle R sets minimal interjet angle NB: d ij distance QCD branching probability α s dk 2 tj dr2 ij d ij

34 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure Top-down jets: cone algorithms (energy flow conserved by QCD)

35 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

36 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

37 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

38 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

39 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

40 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

41 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

42 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

43 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

44 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure

45 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure What seeds do you use? All particles above some threshold Done originally [JetClu, Atlas] Additionally from midpoints between stable cones Midpoint cone [Tevatron Run II]

46 Towards Jetography, G. Salam (p. 11) Two broad classes Cones with Split Merge (SM) Tevatron & ATLAS cone algs have two main steps: Find some/all stable cones cone pointing in same direction as the momentum of its contents Found by iterating from some initial seed directions Resolve cases of overlapping stable cones By running a split merge procedure What seeds do you use? All particles above some threshold Done originally [JetClu, Atlas] Additionally from midpoints between stable cones Midpoint cone [Tevatron Run II]

47 Towards Jetography, G. Salam (p. 12) Snowmass Readying jet technology for the LHC era [a.k.a. satisfying Snowmass]

48 Towards Jetography, G. Salam (p. 13) Snowmass Snowmass accords Snowmass Accord (1990):

49 Towards Jetography, G. Salam (p. 13) Snowmass Snowmass accords Snowmass Accord (1990): Property 1 speed. (+other aspects) LHC events may have up to N = 4000 particles (at high-lumi) Sequential recombination algs. (k t ) slow, N 3 60s for N = 4000, not practical for O ( 10 9) events Can be reduced to N ln N (60s 20ms) Cacciari & GPS 05 + CGAL

50 Towards Jetography, G. Salam (p. 13) Snowmass Snowmass accords Snowmass Accord (1990): Property 4 Infrared and Collinear (IRC) Safety. It helps ensure: Soft (low-energy) emissions & collinear splittings don t change jets Each order of perturbation theory is smaller than previous (at high p t ) Wasn t satisfied by the cone algorithms

51 Towards Jetography, G. Salam (p. 14) Snowmass Cone IR issues JetClu (& Atlas Cone) in NLO jet jet W Snowmass issue #4 Cone algorithms and IR safety α 2 s α EW α 3 s α EW α 3 s α EW 1-jet + 2-jet O (1) 0 With these (& most) cone algorithms, perturbative infinities fail to cancel at some order IR unsafety

52 Towards Jetography, G. Salam (p. 14) Snowmass Cone IR issues JetClu (& Atlas Cone) in NLO jet jet W α 2 s α EW α 3 s α EW α 3 s α EW 1-jet + 2-jet O (1) 0 With these (& most) cone algorithms, perturbative infinities fail to cancel at some order IR unsafety

53 Towards Jetography, G. Salam (p. 14) Snowmass Cone IR issues JetClu (& Atlas Cone) in NLO jet jet jet jet soft divergence W W α 2 s α EW α 3 s α EW α 3 s α EW 1-jet + 2-jet O (1) 0 With these (& most) cone algorithms, perturbative infinities fail to cancel at some order IR unsafety

54 Towards Jetography, G. Salam (p. 14) Snowmass Cone IR issues JetClu (& Atlas Cone) in NLO jet jet jet jet jet soft divergence W W W α 2 s α EW α 3 s α EW α 3 s α EW 1-jet + 2-jet O (1) 0 With these (& most) cone algorithms, perturbative infinities fail to cancel at some order IR unsafety

55 Towards Jetography, G. Salam (p. 14) Snowmass Cone IR issues JetClu (& Atlas Cone) in NLO jet jet jet jet jet soft divergence W W W α 2 s α EW α 3 s α EW α 3 s α EW 1-jet + 2-jet O (1) 0 With these (& most) cone algorithms, perturbative infinities fail to cancel at some order IR unsafety

56 Towards Jetography, G. Salam (p. 15) Snowmass Cone IR issues IRC safety & real-life Real life does not have infinities, but pert. infinity leaves a real-life trace α 2 s + α3 s + α4 s α2 s + α3 s + α4 s lnp t/λ α 2 s + α3 s + α3 s }{{} BOTH WASTED Among consequences of IR unsafety: Last meaningful order JetClu, ATLAS MidPoint CMS it. cone Known at cone [IC-SM] [IC mp-sm] [IC-PR] Inclusive jets LO NLO NLO NLO ( NNLO) W/Z + 1 jet LO NLO NLO NLO 3 jets none LO LO NLO [nlojet++] W/Z + 2 jets none LO LO NLO [MCFM] m jet in 2j + X none none none LO NB: 50,000,000$/ /CHF/e investment in NLO Multi-jet contexts much more sensitive: ubiquitous at LHC And LHC will rely on QCD for background double-checks extraction of cross sections, extraction of parameters

57 Towards Jetography, G. Salam (p. 15) Snowmass Cone IR issues IRC safety & real-life Real life does not have infinities, but pert. infinity leaves a real-life trace α 2 s + α3 s + α4 s α2 s + α3 s + α4 s lnp t/λ α 2 s + α3 s + α3 s }{{} BOTH WASTED Among consequences of IR unsafety: Last meaningful order JetClu, ATLAS MidPoint CMS it. cone Known at cone [IC-SM] [IC mp-sm] [IC-PR] Inclusive jets LO NLO NLO NLO ( NNLO) W/Z + 1 jet LO NLO NLO NLO 3 jets none LO LO NLO [nlojet++] W/Z + 2 jets none LO LO NLO [MCFM] m jet in 2j + X none none none LO NB: 50,000,000$/ /CHF/e investment in NLO Multi-jet contexts much more sensitive: ubiquitous at LHC And LHC will rely on QCD for background double-checks extraction of cross sections, extraction of parameters

58 Towards Jetography, G. Salam (p. 15) Snowmass Cone IR issues IRC safety & real-life Real life does not have infinities, but pert. infinity leaves a real-life trace α 2 s + α3 s + α4 s α2 s + α3 s + α4 s lnp t/λ α 2 s + α3 s + α3 s }{{} BOTH WASTED Among consequences of IR unsafety: Last meaningful order JetClu, ATLAS MidPoint CMS it. cone Known at cone [IC-SM] [IC mp-sm] [IC-PR] Inclusive jets LO NLO NLO NLO ( NNLO) W/Z + 1 jet LO NLO NLO NLO 3 jets none LO LO NLO [nlojet++] W/Z + 2 jets none LO LO NLO [MCFM] m jet in 2j + X none none none LO NB: 50,000,000$/ /CHF/e investment in NLO Multi-jet contexts much more sensitive: ubiquitous at LHC And LHC will rely on QCD for background double-checks extraction of cross sections, extraction of parameters

59 Towards Jetography, G. Salam (p. 16) Snowmass Cone IR issues Two directions Fix stable-cone finding SISCone How do we solve cone IR safety problems? GPS & Soyez 07 Same family as Tev. Run II alg Invent "cone-like" alg. anti-kt Cacciari, GPS & Soyez 08

60 Towards Jetography, G. Salam (p. 17) Snowmass Cone IR issues Essential characteristic of cones? Cone (ICPR)

61 Towards Jetography, G. Salam (p. 17) Snowmass Cone IR issues Cone (ICPR) Essential characteristic of cones? (Some) cone algorithms give circular jets in y φ plane Much appreciated by experiments e.g. for acceptance corrections

62 Towards Jetography, G. Salam (p. 17) Snowmass Cone IR issues Cone (ICPR) Essential characteristic of cones? (Some) cone algorithms give circular jets in y φ plane Much appreciated by experiments e.g. for acceptance corrections k t alg.

63 Towards Jetography, G. Salam (p. 17) Snowmass Cone IR issues Cone (ICPR) Essential characteristic of cones? (Some) cone algorithms give circular jets in y φ plane Much appreciated by experiments e.g. for acceptance corrections k t alg. k t jets are irregular Because soft junk clusters together first: d ij = min(k 2 ti,k2 tj ) R2 ij Regularly held against k t

64 Towards Jetography, G. Salam (p. 17) Snowmass Cone IR issues Cone (ICPR) Essential characteristic of cones? (Some) cone algorithms give circular jets in y φ plane Much appreciated by experiments e.g. for acceptance corrections k t jets are irregular cone-based way of getting k t alg. Because soft junk clusters together first: d ij = min(k 2 ti,k2 tj ) R2 ij Regularly held against k t Is there some other, non circular jets?

65 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

66 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

67 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

68 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

69 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

70 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

71 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

72 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

73 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

74 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

75 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

76 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08

77 Towards Jetography, G. Salam (p. 18) Snowmass Cone IR issues Adapting seq. rec. to give circular jets Soft stuff clusters with nearest neighbour k t : d ij = min(k 2 ti, k2 tj ) R2 ij anti-k t: d ij = R 2 ij max(k 2 ti, k2 tj ) Hard stuff clusters with nearest neighbour Privilege collinear divergence over soft divergence Cacciari, GPS & Soyez 08 anti-k t gives cone-like jets without using stable cones

78 Towards Jetography, G. Salam (p. 19) Snowmass A collection of algs A full set of IRC-safe jet algorithms Generalise inclusive-type sequential recombination with d ij = min(k 2p ti,k 2p tj ) R 2 ij/r 2 d ib = k 2p ti Alg. name Comment time p = 1 k t Hierarchical in rel. k t CDOSTW 91-93; ES 93 N ln N exp. p = 0 Cambridge/Aachen Hierarchical in angle Dok, Leder, Moretti, Webber 97 Scan multiple R at once N ln N Wengler, Wobisch 98 QCD angular ordering p = 1 anti-k t Cacciari, GPS, Soyez 08 Hierarchy meaningless, jets reverse-k t Delsart like CMS cone (IC-PR) N 3/2 SC-SM SISCone Replaces JetClu, ATLAS GPS Soyez 07 + Tevatron run II 00 MidPoint (xc-sm) cones N 2 ln N exp. All these algorithms [& much more] coded in (efficient) C++ at (Cacciari, GPS & Soyez 05-09)

79 Towards Jetography, G. Salam (p. 20) Snowmass A collection of algs ATLAS: first dijet event, with anti-k t

80 Towards Jetography, G. Salam (p. 21) Snowmass A collection of algs CMS: first dijet event, with anti-k t

81 Towards Jetography, G. Salam (p. 22) Beyond Snowmass Snowmass is solved But it was a problem from the 1990s What are the problems we should be trying to solve for LHC?

82 Towards Jetography, G. Salam (p. 23) Beyond Snowmass Which jet definition(s) for LHC? Choice of algorithm (k t, SISCone,... ) Choice of parameters (R,... ) Can we address this question scientifically? Jetography

83 Towards Jetography, G. Salam (p. 23) Beyond Snowmass Which jet definition(s) for LHC? Choice of algorithm (k t, SISCone,... ) Choice of parameters (R,... ) Can we address this question scientifically? Jetography

84 Towards Jetography, G. Salam (p. 24) Physics of jets Jet def n differences Jet definitions }{{} alg + R differ mainly in: 1. How close two particles must be to end up in same jet [discussed in the 90s, e.g. Ellis & Soper] 2. How much perturbative radiation is lost from a jet [indirectly discussed in the 90s (analytic NLO for inclusive jets)] 3. How much non-perturbative contamination (hadronisation, UE, pileup) a jet receives [partially discussed in 90s Korchemsky & Sterman 95, Seymour 97]

85 Towards Jetography, G. Salam (p. 24) Physics of jets Jet def n differences Jet definitions }{{} alg + R differ mainly in: 1. How close two particles must be to end up in same jet [discussed in the 90s, e.g. Ellis & Soper] 2. How much perturbative radiation is lost from a jet [indirectly discussed in the 90s (analytic NLO for inclusive jets)] 3. How much non-perturbative contamination (hadronisation, UE, pileup) a jet receives [partially discussed in 90s Korchemsky & Sterman 95, Seymour 97]

86 Towards Jetography, G. Salam (p. 24) Physics of jets Jet def n differences Jet definitions }{{} alg + R differ mainly in: 1. How close two particles must be to end up in same jet [discussed in the 90s, e.g. Ellis & Soper] 2. How much perturbative radiation is lost from a jet [indirectly discussed in the 90s (analytic NLO for inclusive jets)] 3. How much non-perturbative contamination (hadronisation, UE, pileup) a jet receives [partially discussed in 90s Korchemsky & Sterman 95, Seymour 97]

87 Towards Jetography, G. Salam (p. 25) Physics of jets Perturbative p t Jet p t v. parton p t : perturbatively? The question s dangerous: a parton is an ambiguous concept Three limits can help you: Threshold limit e.g. de Florian & Vogelsang 07 Parton from color-neutral object decay (Z ) Small-R (radius) limit for jet One simple result p t,jet p t,parton p t = α { s π lnr 1.01CF quarks 0.94C A n f gluons + O (α s ) only O (α s ) depends on algorithm & process cf. Dasgupta, Magnea & GPS 07

88 Towards Jetography, G. Salam (p. 26) Physics of jets Non-perturbative p t Jet p t v. parton p t : hadronisation? Hadronisation: the parton-shower hadrons transition Method: infrared finite α s à la Dokshitzer & Webber 95 prediction based on e + e event shape data could have been deduced from old work Korchemsky & Sterman 95 Seymour 97 Main result 0.4 GeV p t,jet p t,parton shower R { CF quarks gluons C A cf. Dasgupta, Magnea & GPS 07 coefficient holds for anti-k t ; see Dasgupta & Delenda 09 for k t alg.

89 Towards Jetography, G. Salam (p. 27) Physics of jets Non-perturbative p t Underlying Event (UE) Naive prediction (UE colour dipole between pp): { p t 0.4 GeV R2 2 CF q q dipole gluon dipole C A DWT Pythia tune or ATLAS Jimmy tune tell you: p t GeV R2 2 This big coefficient motivates special effort to understand interplay between jet algorithm and UE: jet areas How does coefficient depend on algorithm? How does it depend on jet p t? How does it fluctuate? cf. Cacciari, GPS & Soyez 08

90 Towards Jetography, G. Salam (p. 27) Physics of jets Non-perturbative p t Underlying Event (UE) Naive prediction (UE colour dipole between pp): { p t 0.4 GeV R2 2 CF q q dipole gluon dipole C A DWT Pythia tune or ATLAS Jimmy tune tell you: p t GeV R2 2 This big coefficient motivates special effort to understand interplay between jet algorithm and UE: jet areas How does coefficient depend on algorithm? How does it depend on jet p t? How does it fluctuate? cf. Cacciari, GPS & Soyez 08

91 Towards Jetography, G. Salam (p. 28) Physics of jets Jet-properties summary Jet algorithm properties: summary k t Cam/Aachen anti-k t SISCone reach R R R (1 + p t2 p t2 )R p t,pt αsc i π lnr lnr lnr ln1.35r p t,hadr 0.4 GeVC i R 0.7? 1? area = πr ± ± πr 2 C i πb 0 ln αs(q 0) α s(rp t) 0.52 ± ± ± 0.07 In words: k t : area fluctuates a lot, depends on p t (bad for UE) Cam/Aachen: area fluctuates somewhat, depends less on p t anti-k t : area is constant (circular jets) SISCone: reaches far for hard radiation (good for resolution, bad for multijets), area is smaller (good for UE)

92 Towards Jetography, G. Salam (p. 29) Physics with jets Can we benefit from this understanding in our use of jets?

93 Towards Jetography, G. Salam (p. 30) Physics with jets Dijet resonances Jet momentum significantly affected by R So what R should we choose? Examine this in context of reconstruction of dijet resonance

94 Towards Jetography, G. Salam (p. 31) Physics with jets Dijet resonances PT radiation: q : p t α sc F π p t lnr What R is best for an isolated jet? E.g. to reconstruct m X (p tq + p t q ) q Hadronisation: q : p t C F R 0.4 GeV p q X q p Underlying event: q,g : p t R GeV 2 q Minimise fluctuations in p t Use crude approximation: p 2 t p t 2 in small-r limit (?!) cf. Dasgupta, Magnea & GPS 07

95 Towards Jetography, G. Salam (p. 31) Physics with jets Dijet resonances PT radiation: q : p t α sc F π p t lnr Hadronisation: q : p t C F R Underlying event: 0.4 GeV q,g : p t R GeV 2 Minimise fluctuations in p t Use crude approximation: p 2 t p t 2 What R is best for an isolated jet? δp t 2 pert + δp t 2 h + δp t 2 UE [GeV2 ] GeV quark jet LHC quark jets p t = 50 GeV δp t 2 h δp t 2 UE δp t 2 pert R in small-r limit (?!) cf. Dasgupta, Magnea & GPS 07

96 Towards Jetography, G. Salam (p. 31) Physics with jets Dijet resonances PT radiation: q : p t α sc F π p t lnr Hadronisation: q : p t C F R Underlying event: 0.4 GeV q,g : p t R GeV 2 Minimise fluctuations in p t Use crude approximation: p 2 t p t 2 What R is best for an isolated jet? δp t 2 pert + δp t 2 h + δp t 2 UE [GeV2 ] LHC quark jets p t = 1 TeV 1 TeV quark jet δp t 2 pert δp t 2 UE R in small-r limit (?!) cf. Dasgupta, Magnea & GPS 07

97 Towards Jetography, G. Salam (p. 31) Physics with jets Dijet resonances PT radiation: What R is best for an isolated jet? q : p t α sc F π p t lnr 40 At low p t, small R limits relative impact of UE Hadronisation: q : p t C 30 At high Fp t, perturbative effects dominate over 0.4 GeV R non-perturbative R best 1. Underlying event: q,g : p t R GeV 2 Minimise fluctuations in p t Use crude approximation: p 2 t p t 2 δp t 2 pert + δp t 2 h + δp t 2 UE [GeV2 ] LHC quark jets p t = 1 TeV 1 TeV quark jet δp t 2 pert δp t 2 UE R in small-r limit (?!) cf. Dasgupta, Magnea & GPS 07

98 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 0.3 qq, M = 100 GeV SISCone, R=0.3, f=0.75 Q w f=0.24 = 24.0 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets q q X q p dijet mass [GeV] q

99 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 0.3 qq, M = 100 GeV SISCone, R=0.3, f=0.75 Q w f=0.24 = 24.0 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

100 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 0.4 qq, M = 100 GeV SISCone, R=0.4, f=0.75 Q w f=0.24 = 22.5 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

101 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 0.5 qq, M = 100 GeV SISCone, R=0.5, f=0.75 Q w f=0.24 = 22.6 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

102 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 0.6 qq, M = 100 GeV SISCone, R=0.6, f=0.75 Q w f=0.24 = 23.8 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

103 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 0.7 qq, M = 100 GeV SISCone, R=0.7, f=0.75 Q w f=0.24 = 25.1 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

104 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 0.8 qq, M = 100 GeV SISCone, R=0.8, f=0.75 Q w f=0.24 = 26.8 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

105 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 0.9 qq, M = 100 GeV SISCone, R=0.9, f=0.75 Q w f=0.24 = 28.8 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

106 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 1.0 qq, M = 100 GeV SISCone, R=1.0, f=0.75 Q w f=0.24 = 31.9 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

107 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 1.1 qq, M = 100 GeV SISCone, R=1.1, f=0.75 Q w f=0.24 = 34.7 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

108 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 1.2 qq, M = 100 GeV SISCone, R=1.2, f=0.75 Q w f=0.24 = 37.9 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

109 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 1.3 qq, M = 100 GeV SISCone, R=1.3, f=0.75 Q w f=0.24 = 42.3 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: p Resonance X dijets jet q q X q p dijet mass [GeV] jet q

110 Towards Jetography, G. Salam (p. 32) Physics with jets Dijet resonances 1/N dn/dbin / R = 1.3 qq, M = 100 GeV SISCone, R=1.3, f=0.75 Q w f=0.24 = 42.3 GeV Dijet mass: scan over R [Pythia 6.4] arxiv: ρ L from Q w f= qq, M = 100 GeV SISCone, f=0.75 arxiv: dijet mass [GeV] R After scanning, summarise quality v. R. Minimum BEST picture not so different from crude analytical estimate

111 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 100 GeV qq, M = 100 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

112 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 150 GeV qq, M = 150 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

113 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 200 GeV qq, M = 200 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

114 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 300 GeV qq, M = 300 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

115 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 500 GeV qq, M = 500 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

116 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 700 GeV qq, M = 700 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

117 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 1000 GeV qq, M = 1000 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

118 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 2000 GeV qq, M = 2000 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

119 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 4000 GeV qq, M = 4000 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

120 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 4000 GeV qq, M = 4000 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

121 Towards Jetography, G. Salam (p. 33) Physics with jets Dijet resonances Scan through q q mass values ρ L from Q w f= m qq = 4000 GeV qq, M = 4000 GeV SISCone, f=0.75 arxiv: Best R is at minimum of curve Best R depends strongly on mass of system Increases with mass, just like crude analytical prediction NB: current analytics too crude BUT: so far, LHC s plans involve running with fixed smallish R values e.g. CMS arxiv: R NB: 100,000 plots for various jet algorithms, narrow qq and gg resonances from Cacciari, Rojo, GPS & Soyez 08

122 Towards Jetography, G. Salam (p. 34) Physics with jets Dijet resonances

123 Towards Jetography, G. Salam (p. 35) Physics with jets Boosted heavy particles The dijet mass is a classic jets analysis. But LHC also opens up characteristically new kinematic regions, because s m EW. We can and should make use of this Illustrated in next slides, for Higgs search with m H = 115 GeV, H b b

124 Towards Jetography, G. Salam (p. 36) Physics with jets Boosted heavy particles E.g.: WH/ZH search LHC Signal is W lν, H b b. Backgrounds include Wb b, t t lνb bjj,... Studied e.g. in ATLAS TDR Difficulties, e.g. gg t t has lνb b with same intrinsic mass scale, but much higher partonic luminosity Need exquisite control of bkgd shape Try a long shot? Go to high p t (p th,p tv > 200 GeV) Lose 95% of signal, but more efficient? Maybe kill t t & gain clarity? W e,µ H b b ν

125 Towards Jetography, G. Salam (p. 36) Physics with jets Boosted heavy particles E.g.: WH/ZH search LHC Signal is W lν, H b b. Backgrounds include Wb b, t t lνb bjj,... Studied e.g. in ATLAS TDR pp WH lνb b + bkgds ATLAS TDR Difficulties, e.g. gg t t has lνb b with same intrinsic mass scale, but much higher partonic luminosity Need exquisite control of bkgd shape H b Try a long shot? Go to high p t (p th,p tv > 200 GeV) Lose 95% of signal, but more efficient? Maybe kill t t & gain clarity? W e,µ b ν

126 Towards Jetography, G. Salam (p. 36) Physics with jets Boosted heavy particles E.g.: WH/ZH search LHC Signal is W lν, H b b. Backgrounds include Wb b, t t lνb bjj,... Studied e.g. in ATLAS TDR Difficulties, e.g. gg t t has lνb b with same intrinsic mass scale, but much higher partonic luminosity Need exquisite control of bkgd shape pp WH lνb b + bkgds ATLAS TDR b b H Try a long shot? Go to high p t (p th,p tv > 200 GeV) Lose 95% of signal, but more efficient? Maybe kill t t & gain clarity? e,µ W ν

127 Towards Jetography, G. Salam (p. 36) Physics with jets Boosted heavy particles E.g.: WH/ZH search LHC Signal is W lν, H b b. Backgrounds include Wb b, t t lνb bjj,... Studied e.g. in ATLAS TDR gg t t has lνb b with same intrinsic Question: mass scale, but much higher partonic luminosity What s the best strategy to identify the Need exquisite control of bkgd shape two-pronged structure of the boosted Higgs decay? b b ATLAS TDR pp WH lνb b + bkgds Difficulties, e.g. H Try a long shot? Go to high p t (p th,p tv > 200 GeV) Lose 95% of signal, but more efficient? Maybe kill t t & gain clarity? e,µ W ν

128 Towards Jetography, G. Salam (p. 37) Physics with jets Boosted heavy particles Past methods Use k t alg. s distance measure (rel. trans. mom.) to cut out QCD bkgd: Use k t jet-algorithm s hierarchy to split the jets Y-splitter d kt ij = min(p 2 ti,p2 tj ) R2 ij only partially correlated with mass

129 Towards Jetography, G. Salam (p. 37) Physics with jets Boosted heavy particles Past methods Use k t alg. s distance measure (rel. trans. mom.) to cut out QCD bkgd: Use k t jet-algorithm s hierarchy to split the jets Y-splitter d kt ij = min(p 2 ti,p2 tj ) R2 ij only partially correlated with mass

130 Towards Jetography, G. Salam (p. 38) Physics with jets Boosted heavy particles Our tool The Cambridge/Aachen jet alg. Dokshitzer et al 97 Wengler & Wobisch 98 Work out Rij 2 = y2 ij + φ2 ij between all pairs of objects i,j; Recombine the closest pair; Repeat until all objects separated by R ij > R. [in FastJet] Gives hierarchical view of the event; work through it backwards to analyse jet

131 Towards Jetography, G. Salam (p. 38) Physics with jets Boosted heavy particles Our tool The Cambridge/Aachen jet alg. Dokshitzer et al 97 Wengler & Wobisch 98 Work out Rij 2 = y2 ij + φ2 ij between all pairs of objects i,j; Recombine the closest pair; Repeat until all objects separated by R ij > R. [in FastJet] Gives hierarchical view of the event; work through it backwards to analyse jet k t algorithm Cam/Aachen algorithm Allows you to dial the correct R to keep perturbative radiation, but throw out UE

132 Towards Jetography, G. Salam (p. 39) Physics with jets Boosted heavy particles pp ZH ν νb m H =115GeV Herwig Jimmy FastJet 2.3 SIGNAL Zbb BACKGROUND Cluster event, C/A, R=1.2 Butterworth, Davison, Rubin & GPS 08 arbitrary norm.

133 Towards Jetography, G. Salam (p. 39) Physics with jets Boosted heavy particles pp ZH ν νb m H =115GeV Herwig Jimmy FastJet 2.3 SIGNAL Zbb BACKGROUND Fill it in, show jets more clearly Butterworth, Davison, Rubin & GPS 08 arbitrary norm.

134 Towards Jetography, G. Salam (p. 39) Physics with jets Boosted heavy particles pp ZH ν νb m H =115GeV Herwig Jimmy FastJet SIGNAL 200 < p tz < 250 GeV m H [GeV] Zbb BACKGROUND < p tz < 250 GeV Consider hardest jet, m = 150 GeV Butterworth, Davison, Rubin & GPS m H [GeV] arbitrary norm.

135 Towards Jetography, G. Salam (p. 39) Physics with jets Boosted heavy particles pp ZH ν νb m H =115GeV Herwig Jimmy FastJet SIGNAL 200 < p tz < 250 GeV m H [GeV] Zbb BACKGROUND < p tz < 250 GeV split: m = 150 GeV, max(m 1,m2) m = 0.92 repeat Butterworth, Davison, Rubin & GPS m H [GeV] arbitrary norm.

136 Towards Jetography, G. Salam (p. 39) Physics with jets Boosted heavy particles pp ZH ν νb m H =115GeV Herwig Jimmy FastJet SIGNAL 200 < p tz < 250 GeV m H [GeV] Zbb BACKGROUND < p tz < 250 GeV split: m = 139 GeV, max(m 1,m2) m = 0.37 mass drop Butterworth, Davison, Rubin & GPS m H [GeV] arbitrary norm.

137 Towards Jetography, G. Salam (p. 39) Physics with jets Boosted heavy particles pp ZH ν νb m H =115GeV Herwig Jimmy FastJet SIGNAL 200 < p tz < 250 GeV m H [GeV] Zbb BACKGROUND < p tz < 250 GeV check: y 12 p t2 p t1 0.7 OK + 2 b-tags (anti-qcd) Butterworth, Davison, Rubin & GPS m H [GeV] arbitrary norm.

138 Towards Jetography, G. Salam (p. 39) Physics with jets Boosted heavy particles pp ZH ν νb m H =115GeV Herwig Jimmy FastJet SIGNAL 200 < p tz < 250 GeV m H [GeV] Zbb BACKGROUND < p tz < 250 GeV R filt = 0.3 Butterworth, Davison, Rubin & GPS m H [GeV] arbitrary norm.

139 Towards Jetography, G. Salam (p. 39) Physics with jets Boosted heavy particles pp ZH ν νb m H =115GeV Herwig Jimmy FastJet SIGNAL 200 < p tz < 250 GeV m H [GeV] Zbb BACKGROUND < p tz < 250 GeV R filt = 0.3: take 3 hardest, m = 117 GeV Butterworth, Davison, Rubin & GPS m H [GeV] arbitrary norm.

140 Towards Jetography, G. Salam (p. 40) Physics with jets Boosted heavy particles combine HZ and HW, p t > 200 GeV Take Z l + l, Z ν ν, W lν l = e, µ p tv,p th > 200 GeV η V, η H < 2.5 Assume real/fake b-tag rates of 0.6/0.02. Some extra cuts in HW channels to reject t t. Assume m H = 115 GeV. At 5σ for 30 fb 1 this looks like a competitive channel for light Higgs discovery. A powerful method! Currently under study in the LHC experiments

141 Towards Jetography, G. Salam (p. 41) Physics with jets Boosted heavy particles ATLAS results Leptonic channel What changes compared to particle-level analysis? 1.5σ as compared to 2.1σ Expected given larger mass window

142 Towards Jetography, G. Salam (p. 41) Physics with jets Boosted heavy particles ATLAS results Missing E T channel What changes compared to particle-level analysis? 1.5σ as compared to 3σ Suffers: some events redistributed to semi-leptonic channel

143 Towards Jetography, G. Salam (p. 41) Physics with jets Boosted heavy particles ATLAS results Semi-leptonic channel What changes compared to particle-level analysis? 3σ as compared to 3σ Benefits: some events redistributed from missing E T channel

144 Towards Jetography, G. Salam (p. 42) Physics with jets Boosted heavy particles ATLAS combined results Likelihood-based analysis of all three channels together gives signal significance of 3.7σ for 30 fb 1 To be compared with 4.2σ in hadron-level analysis for m H = 120 GeV K-factors not included: don t affect significance ( 1.5 for VH, for Vbb) With 5% (20%) background uncertainty, ATLAS result becomes 3.5σ (2.8σ) Comparison to other channels at ATLAS (m H = 120, 30 fb 1 ): gg H γγ WW H ττ gg H ZZ 4.2σ 4.9σ 2.6σ Extracted from

145 Towards Jetography, G. Salam (p. 43) Physics with jets Boosted heavy particles The potential of better jet finding X gg VH,H b b tth,h b b RPV χ 0 1 qqq 1/N dn/dbin / gg, M = 2000 GeV k t, R=0.4 Q w f=0.13 = 190 GeV arxiv: m 1 /(100GeV) dn/dbin per fb signal + background background (just dijets) signal Cam/Aachen R=0.7 p t1 > 500 GeV dijet mass [GeV] Herwig Jimmy m 1 [GeV] 1/N dn/dbin / gg, M = 2000 GeV C/A-filt, R=1.3 Q w f=0.13 = 53 GeV (subtr.) arxiv: m 1 /(100GeV) dn/dbin per fb signal + background background (just dijets) signal Cam/Aachen + filt, R= dijet mass [GeV] p t1 > 500 GeV, z split > 0.15, m bc > 0.25 m abc p t3, p t4 > {100,80} GeV, y 13, y 14 < m 1 [GeV] 1) Cacciari, Rojo, GPS & Soyez 08; 2) Butterworth, Davison, Rubin & GPS 08; 3) Plehn, GPS & Spannowsky 09; 4) Butterworth, Ellis, Raklev & GPS 09.

146 Towards Jetography, G. Salam (p. 44) Conclusions Conclusions

147 Towards Jetography, G. Salam (p. 45) Conclusions Conclusions There are no longer any valid reasons for using jet algorithms that are incompatible with the Snowmass criteria. LHC experiments are adopting the new tools Individual analyses need to follow suit It s time to move forwards with the question of how best to use jets in searches Examples here show two things: Good jet-finding brings significant gains There s room for serious QCD theory input into optimising jet use Not the only way of doing things But brings more insight than trial & error MC This opens the road towards Jetography, QCD-based autofocus for jets

148 Towards Jetography, G. Salam (p. 46) Extras EXTRAS

149 Towards Jetography, G. Salam (p. 47) Extras speed k t and geometry There are N(N 1)/2 distances d ij surely we have to calculate them all in order to find smallest? k t distance measure is partly geometrical: min i,j d ij min(min{kti,k 2 tj} R 2 ij) 2 i,j = min(kti R 2 ij) 2 i,j = min(kti 2 min i j ւ 2D dist. on rap., φ cylinder ) R 2 ij In words: for each i look only at the k t distance to its 2D geometrical nearest neighbour (GNN). k t distance need only be calculated between GNNs Each point has 1 GNN need only calculate N d ij s Cacciari & GPS, 05

150 Towards Jetography, G. Salam (p. 47) Extras speed k t and geometry There are N(N 1)/2 distances d ij surely we have to calculate them all in order to find smallest? k t distance measure is partly geometrical: min i,j d ij min(min{kti,k 2 tj} R 2 ij) 2 i,j = min(kti R 2 ij) 2 i,j = min(kti 2 min i j ւ 2D dist. on rap., φ cylinder ) R 2 ij In words: for each i look only at the k t distance to its 2D geometrical nearest neighbour (GNN). k t distance need only be calculated between GNNs Each point has 1 GNN need only calculate N d ij s Cacciari & GPS, 05

151 Towards Jetography, G. Salam (p. 47) Extras speed k t and geometry There are N(N 1)/2 distances d ij surely we have to calculate them all in order to find smallest? k t distance measure is partly geometrical: min i,j d ij min(min{kti,k 2 tj} R 2 ij) 2 i,j = min(kti R 2 ij) 2 i,j = min(kti 2 min i j ւ 2D dist. on rap., φ cylinder ) R 2 ij In words: for each i look only at the k t distance to its 2D geometrical nearest neighbour (GNN). k t distance need only be calculated between GNNs Each point has 1 GNN need only calculate N d ij s Cacciari & GPS, 05

152 Towards Jetography, G. Salam (p. 47) Extras speed k t and geometry There are N(N 1)/2 distances d ij surely we have to calculate them all in order to find smallest? k t distance measure is partly geometrical: min i,j d ij min(min{kti,k 2 tj} R 2 ij) 2 i,j = min(kti R 2 ij) 2 i,j = min(kti 2 min i j ւ 2D dist. on rap., φ cylinder ) R 2 ij In words: for each i look only at the k t distance to its 2D geometrical nearest neighbour (GNN). k t distance need only be calculated between GNNs Each point has 1 GNN need only calculate N d ij s Cacciari & GPS, 05

153 Towards Jetography, G. Salam (p. 47) Extras speed k t and geometry There are N(N 1)/2 distances d ij surely we have to calculate them all in order to find smallest? k t distance measure is partly geometrical: min i,j d ij min(min{kti,k 2 tj} R 2 ij) 2 i,j = min(kti R 2 ij) 2 i,j = min(kti 2 min i j ւ 2D dist. on rap., φ cylinder ) R 2 ij In words: for each i look only at the k t distance to its 2D geometrical nearest neighbour (GNN). k t distance need only be calculated between GNNs Each point has 1 GNN need only calculate N d ij s Cacciari & GPS, 05

154 Towards Jetography, G. Salam (p. 48) Extras speed 2d nearest-neighbours Given a set of vertices on plane (1...10) a Voronoi diagram partitions plane into cells containing all points closest to each vertex Dirichlet 1850, Voronoi 1908 A vertex s nearest other vertex is al How 6 does use of GNN help? Aren t 9 there still N2 N 2 2 R2 ways ij 2 to check...? in an adjacent cell. Geometrical nearest neighbour finding is a classic problem in the field of Computational Geometry E.g. GNN of point 7 must be among 1,4,2,8,3 (it is 3) Construction of Voronoi diagram for N points: N ln N time Fortune 88 Update of 1 point in Voronoi diagram: expected lnn time Devillers 99 [+ related work by other authors] Convenient C++ package available: CGAL, with help of CGAL, k t clustering can be done in N ln N time Coded in the FastJet package (v1), Cacciari & GPS 06

155 Towards Jetography, G. Salam (p. 48) Extras speed 2d nearest-neighbours Given a set of vertices on plane (1...10) a Voronoi diagram partitions plane into cells containing all points closest to each vertex Dirichlet 1850, Voronoi 1908 A vertex s nearest other vertex is always in an adjacent cell. E.g. GNN of point 7 must be among 1,4,2,8,3 (it is 3) Construction of Voronoi diagram for N points: N ln N time Fortune 88 Update of 1 point in Voronoi diagram: expected lnn time Devillers 99 [+ related work by other authors] Convenient C++ package available: CGAL, with help of CGAL, k t clustering can be done in N ln N time Coded in the FastJet package (v1), Cacciari & GPS 06

156 Towards Jetography, G. Salam (p. 48) Extras speed 2d nearest-neighbours Given a set of vertices on plane (1...10) a Voronoi diagram partitions plane into cells containing all points closest to each vertex Dirichlet 1850, Voronoi 1908 A vertex s nearest other vertex is always in an adjacent cell. E.g. GNN of point 7 must be among 1,4,2,8,3 (it is 3) Construction of Voronoi diagram for N points: N ln N time Fortune 88 Update of 1 point in Voronoi diagram: expected lnn time Devillers 99 [+ related work by other authors] Convenient C++ package available: CGAL, with help of CGAL, k t clustering can be done in N ln N time Coded in the FastJet package (v1), Cacciari & GPS 06

157 Towards Jetography, G. Salam (p. 48) Extras speed 2d nearest-neighbours Given a set of vertices on plane (1...10) a Voronoi diagram partitions plane into cells containing all points closest to each vertex Dirichlet 1850, Voronoi 1908 A vertex s nearest other vertex is always in an adjacent cell. E.g. GNN of point 7 must be among 1,4,2,8,3 (it is 3) Construction of Voronoi diagram for N points: N ln N time Fortune 88 Update of 1 point in Voronoi diagram: expected lnn time Devillers 99 [+ related work by other authors] Convenient C++ package available: CGAL, with help of CGAL, k t clustering can be done in N ln N time Coded in the FastJet package (v1), Cacciari & GPS 06

158 Towards Jetography, G. Salam (p. 49) Extras speed Speeds in 2005 t / s R=0.7 KtJet k t CDF MidPoint (seeds > 0 GeV) CDF MidPoint (seeds > 1 GeV) CDF JetClu (very unsafe) 3.4 GHz P4, 2 GB 10-4 Tevatron LHC lo-lumi LHC hi-lumi LHC Pb-Pb N FastJet (v2.x), codes all developments, natively (k t, Cam/Aachen, anti-k t ) or as plugins (SISCone): Cacciari, GPS & Soyez

159 Towards Jetography, G. Salam (p. 49) Extras speed Speeds in 2009 t / s R=0.7 KtJet k t CDF MidPoint (seeds > 0 GeV) CDF MidPoint (seeds > 1 GeV) CDF JetClu (very unsafe) Seedless IR Safe Cone (SISCone) FastJet anti-k t Cam/Aachen k t 10-4 LHC lo-lumi LHC hi-lumi LHC Pb-Pb N FastJet (v2.x), codes all developments, natively (k t, Cam/Aachen, anti-k t ) or as plugins (SISCone): Cacciari, GPS & Soyez

160 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

161 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

162 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

163 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

164 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

165 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

166 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

167 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

168 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

169 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

170 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

171 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat

172 Towards Jetography, G. Salam (p. 50) Extras ICPR unsafety Iterative Cone [with progressive removal] Procedure: Find one stable cone By iterating from hardest seed particle Call it a jet; remove its particles from the event; repeat Iterative Cone with Progressive Removal (IC-PR) e.g. CMS it. cone, [Pythia Cone, GetJet],... NB: not same type of algorithm as Atlas Cone, MidPoint, SISCone

173 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

174 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

175 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

176 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

177 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

178 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

179 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

180 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

181 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

182 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

183 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

184 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

185 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

186 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

187 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

188 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

189 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity jet 2 Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

190 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity jet 2 Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

191 Towards Jetography, G. Salam (p. 51) Extras ICPR unsafety ICPR iteration issue 500 cone iteration cone axis cone p T (GeV/c) jet 1 rapidity jet 2 Collinear splitting can modify the hard jets: ICPR algorithms are collinear unsafe = perturbative calculations give

192 Towards Jetography, G. Salam (p. 52) Extras ICPR unsafety Consequences of collinear unsafety Collinear Safe Collinear Unsafe jet 1 jet 1 jet 1 n α s x ( ) n α s x (+ ) Infinities cancel n α s x ( ) jet 1 n jet 2 α s x (+ ) Infinities do not cancel Invalidates perturbation theory

193 Towards Jetography, G. Salam (p. 52) Extras ICPR unsafety Consequences of collinear unsafety Collinear Safe Collinear Unsafe jet 1 jet 1 jet 1 n α s x ( ) n α s x (+ ) Infinities cancel n α s x ( ) jet 1 n jet 2 α s x (+ ) Infinities do not cancel Invalidates perturbation theory

194 Towards Jetography, G. Salam (p. 53) Extras IRC unsafety impact Impact of IRC issues in W+3j CDF have measured W+3jet X-section with JetClu (IR 2+1 unsafe). NLO calculation with JetClu would diverge [for zero seed threshold] Strategy for theory: use 2 algs for theory prediction, SISCone & anti-k t ; difference between them is IRC unsafety systematic. With CDF cuts and R choice, difference is O (20%) NLO: Ellis, Melnikov & Zanderighi 09 20% exp. systematics

195 Towards Jetography, G. Salam (p. 53) Extras IRC unsafety impact Impact of IRC issues in W+3j CDF have measured W+3jet X-section with JetClu (IR 2+1 unsafe). NLO calculation with JetClu would diverge [for zero seed threshold] Strategy for theory: use 2 algs for theory prediction, SISCone & anti-k t ; difference between them is IRC unsafety systematic. With CDF cuts and R choice, difference is O (20%) NLO: Ellis, Melnikov & Zanderighi 09 20% exp. systematics dσ anti-kt /dp t3 / dσ SISCone /dp t3 2 anti-k t / SISCone Tevatron W + LO (MCFM 5.2) η jets < 2, R= p t3 [GeV]

196 Towards Jetography, G. Salam (p. 53) Extras IRC unsafety impact Impact of IRC issues in W+3j CDF have measured W+3jet X-section with JetClu (IR 2+1 unsafe). NLO calculation with JetClu would diverge [for zero seed threshold] Strategy for theory: use 2 algs for theory prediction, SISCone & anti-k t ; difference between them is IRC unsafety systematic. With CDF cuts and R choice, difference is O (20%) NLO: Ellis, Melnikov & Zanderighi 09 20% exp. systematics With other cuts and R choice, IRC systematic can be up to 75% Future measurements deserve to be done with IRC safe algs... dσ anti-kt /dp t3 / dσ SISCone /dp t3 2 anti-k t / SISCone Tevatron W + LO (MCFM 5.2) η jets < 1, R= η jets < 2, R=0.7 η jets < 1, R=0.4 η jets < 2, R= p t3 [GeV]

197 Towards Jetography, G. Salam (p. 54) Extras IRC unsafety impact Does lack of IRC safety matter? I do searches, not QCD. Why should I care about IRC safety? Are you looking for a mass-peak? you needn t care much Are you looking for an excess over bkgd? you need control samples, validated against QCD W+1,2,3 jets }{{} NLO v. data W+n jets }{{} LO, LO+MC v. data new-physics search }{{} LO+MC v. data

198 Towards Jetography, G. Salam (p. 54) Extras IRC unsafety impact Does lack of IRC safety matter? I do searches, not QCD. Why should I care about IRC safety? Are you looking for a mass-peak? you needn t care much Are you looking for an excess over bkgd? you need control samples, validated against QCD W+1,2,3 jets }{{} NLO v. data W+n jets }{{} LO, LO+MC v. data new-physics search }{{} LO+MC v. data

199 Towards Jetography, G. Salam (p. 54) Extras IRC unsafety impact Does lack of IRC safety matter? I do searches, not QCD. Why should I care about IRC safety? Are you looking for a mass-peak? you needn t care much Are you looking for an excess over bkgd? you need control samples, validated against QCD W+1,2,3 jets }{{} W+n jets }{{} new-physics search }{{} NLO v. data LO, LO+MC v. data LO+MC v. data IR safe alg. IR safe alg. IR safe alg.