Support #7804

Issues with decaying LHE files with p8/p8 & off-shell decays

Added by Stephen Mrenna about 6 years ago. Updated almost 6 years ago.

Start date:
Due date:
% Done:


Estimated time:


Set 24:mMin = xxx.xx to set the minimum value for W decays. xxx = 10 by default.

Hi Steve
Just a kind reminder of this pending issue..
hide quoted text -

On 12 Dec 2014, at 15:25, Artur Lobanov <> wrote:

Hi Steve,

Since we managed to get Pythia 8 to do off-shell decays, I decided to compare it to Pythia 6 with the same SLHA model.
I noticed some strange thing about the W mass distribution (which is produced from a chargino decay -> W LSP, where M_Chargino — M_LSP = 15GeV).
On this plot I just put the pure M_W for P6/P8:
One can see, that both distributions end at 15GeV, which is the deltaM from the decay, but then the P8 curve also has a second edge at 10 GeV.
This seems off, since there was no parameter restricting the off-shell W.

If we now normalise the histograms in the M_W > 10 GeV range, then we get an exact agreement of P6 and P8:

This would mean, that the M_W shape in P8 is produced like in P6, but then is restricted by 10 GeV from below.

Do you have any idea, what could be the reason for such an odd behaviour of Pythia 8?

Thanks and regards,

On 12 Nov 2014, at 19:02, Artur Lobanov <> wrote:

Dear Steve,

I tried out the trick to do off-shell 3-body decays in P8, which you mentioned.
I put the following decay table into the SLHA file:

DECAY 1000021 1.0
0.25 3 1000024 1 2 # ~g -> ~chi_1+ d ubar
0.25 3 -1000024 2 -1 # ~g -> ~chi_1
u dbar
0.25 3 1000024 3 4 # ~g -> ~chi_1+ s cbar
0.25 3 -1000024 4 -3 # ~g -> ~chi_1
c sbar
DECAY 1000024 0.1
0.0 3 1000022 -11 12 # this is the hack!
1.0 2 1000022 24 # ~chi_1+ -> ~chi_10 W
DECAY 1000022 0.0

It worked fine, producing all the desired decays, but still complained about the missing GAUGE block. Should we also ignore this warning?

PYTHIA Info from SLHAinterface::initSLHA: importing SLHA decay table(s)
PYTHIA Warning in SLHAinterface::initSLHA: adding off shell DECAY mode 1000024 -> 1000022 24
PYTHIA Warning in CoupSUSY::initSUSY: Block GAUGE not found or incomplete; using sin(thetaW) at mZ
PYTHIA Warning in ParticleDataEntry::initBWmass: switching off width for id = 36
PYTHIA Warning in ParticleDataEntry::initBWmass: switching off width for id = 1000024
PYTHIA Warning in ParticleDataEntry::initBWmass: switching off width for id = 1000024
PYTHIA Warning in ParticleDataEntry::initBWmass: switching off width for id = 36


On 10 Nov 2014, at 23:29, Stephen Mrenna <> wrote:

On 10/31/2014 07:23 PM, Alexis Kalogeropoulos wrote:

Hi Steve

One more question - to your opinion, can we still use the “decayed” LHE we got with p6 even with the “far-decayed from the central” events, or we should redo everything from scratch ?

I hate to give you more work, but I would check if throwing away those "far away" events in p6
matches what you get with p8.

Further, even with the p8 decays, as you might remember last time at cern we meet, I told you that the clustering scales line from the MG is replace with a new hashtag with something like # pdf x1 x2 clustering_scales. Do you know now what is what ? Do we need to keep the format at it is produced from p8 decays ?

I'm trying to get a clarification on this.




On 31 Oct 2014, at 22:02, Stephen Mrenna <> wrote:

Hi, Alexis, Artur,

I first investigated what was happening in Pythia8.
I find that, if I add a dummy "on-shell" decay, then the
off-shell decay can occur.

DECAY 1000002 2.0
1.0 2 1000024 1
DECAY 1000024 0.001
0.0 3 1000022 -11 12
1.0 2 1000022 24

However, Pythia8 does not allow the chargino to have a
BW distribution in this case. On reflection, this is probably right.
If I play with MSSM parameters for such a compressed spectrum,
I get a width that is 1e-4 - 1e-6 GeV, so even .1 GeV is a gross

In fact, if I use the width estimate of 1e-3 as above, I find that Pythia6 also
puts the chargino on the mass shell (that is probably below some internal

So, all in all, I think a smaller width is in order for these off-shell decays, and that
your intuition that the chargino should not go far off-shell is correct.


On 10/31/2014 07:35 AM, Alexis Kalogeropoulos wrote:

Hi Steve

On 29 Oct 2014, at 16:45, Stephen Mrenna <> wrote:

My responsibilities within CMS require me to continue to support Py6,
so no problems there.
On 10/29/2014 06:20 AM, Alexis Kalogeropoulos wrote:

So, the first question would be if this behaviour is ok i.e. if pythia should generate a mass way too far from the central one. I mean, for a width of 0.1 GeV, what having a mass of +200-300 GeV is too much I thing. I also append such an event as attachment [1]

The Breit-Wigner is a funny distribution. Those distributions out on the tails don't look to be at odds with
the BW shape. Did you try and fit a BW to them? Honestly, I don't have a good eye for fitting a distribution to a small number of counts on the tails.

Nop, we did not. But on the other hand, the other issue that I don’t understand if this is BW thing then there must be also “-“ values around the central one, not only “+” (i.e. the distribution falls only to the right…). Do you agree ?

As to the correctness: Pythia is sampling trying to account for both Breit-Wigners -- the one for the chargino and the one for the W. There is some small chance that the chargino mass goes up to undo the suppression from the W propagator.

so, you say that in principle it goes up until you get a on-shell W ? In that case, I can understand why M_ch1 moves only to “upward” direction as I said in my last point. n such a case, is this something we can calculate analytically in order to xcheck what is the probability of this to happen ?

Note that sometimes we expect the Higgs mass to fluctuate up to produce two on-shell Z bosons -- hence
the excitement about measuring the width of the Higgs by comparing the off-shell and on-shell production.

[*] Now, we tried to also port our workflow and use p8 for the decay of the LHE files, but it complains that it does not like off-shell decays ( event if we specify in the slha a 3-body decay of the ch1) . This is what we get from p8 ...

and nothing happens, i.e. it does not decay the on-shell s-particles

Okay, I will look at this. There are (at least) two possibilities: 1. There is some flag that is set wrong, so that these decays are not happening, 2. We did not port the Py6 capability to handle two off-shell decays to Py8.

Artur, can you paste here the full recipe/ card you used for p8 which did not decay as we say ?



Keep bugging me!


Kind regards


<Mail Attachment.png>


#1 Updated by Stephen Mrenna almost 6 years ago

  • Status changed from Resolved to Closed

Also available in: Atom PDF