MadSpinError : Found too many weight larger than the computed max_weight

Hi,

I never test such type of polarised production (a particle polarised within MadSpin).
It is surprising that this does not work out of the bo which might means that they are something interesting to learn in term of physics.

I have tried to run the same type of computation with parameter to increase the time used to estimate the maximum of the distribution:

> set Nevents_for_max_weight 750 # number of events for the estimate of the max. weight
> # set BW_cut 15 # cut on how far the particle can be off-shell
> # set spinmode onshell # Use one of the madspin special mode
> set max_weight_ps_point 1000 # number of PS to estimate the maximum for each event

(so it takes ~20 times longer to evaluate the maximum weight)

Looks like the maximum is now extremely high since the code requires a lot of attempt to successfully generates a valid decay.

Additionally, the maximum is still not high enough since I do have the following error printed:

> ERROR: Found a weight MUCH larger than the computed max_weight (ratio: 11.828160948787199).
> This usually means that the Narrow width approximation reaches it's limit on part of the Phase-Space.
> Do not trust too much the tale of the distribution and/or relaunch the code with smaller BW_cut.
> This is for channel ('t_bwp0_wp_dxu', 'tx_wm0bx_wm_vexem') with current BW_value at : 14.1572'

Since the issue seems be related to quite offshell decay (here 14 times the width).

I have tried the following setup:

> set Nevents_for_max_weight 750 # number of events for the estimate of the max. weight
> set BW_cut 1 # cut on how far the particle can be off-shell
> # set spinmode onshell # Use one of the madspin special mode
> set max_weight_ps_point 1000 # number of PS to estimate the maximum for each event

This one was more "reasonable" in term of speed to generate the decay (still ~2s per event)
but still fail with

> Command "generate_events -f" interrupted with error:
> MadSpinError : Found too many weight larger than the computed max_weight (4/205 = 1.951219512195122%).
> Please relaunch MS with more events/PS point by event in the
> computation of the maximum_weight.
>
> Please report this bug on https://bugs.launchpad.net/mg5amcnlo
> More information is found in '/Users/omattelaer/Documents/workspace/2.9.4/PROC_sm_53/run_04_tag_1_debug.log'.
> Please attach this file to your report.

My suggestion would be to generate
generate mu+ mu- > t t~ > vm vm~ w+{0} w-{0} b b~
and then the decay in madspin is less problematic.
(which might hint more towards a bug than a physical issue).

Do not forget to specify in which frame in which you want to evaluate your matrix-element since polarization break the lorentz invariance of the computation.

Cheers,

Olivier

> On 19 Sep 2021, at 10:41, Chang-Yuan Yao <email address hidden> wrote:
>
> decay t~ > w-{0} b~, w- > e- ve~