semileptonic decays in madspin

Hi,

This is indeed not a valid syntax.

Each decay line should be associated to a dedicated particle so in your case, you need (at least) one line for the top and (at least one line for the anti-top.
Your syntax is interpreted for the full matrix- element as
p p > t t~, (t > w+ b, w+ > c b~, t~ > w- b~, w- > ell- vl~), (t > w+ b, w+ > c b~, t~ > w- b~, w- > ell- vl~)
note the paranthesis that specify that "w+", "t~" and "w-" should only apply to the decay of the top given by "t > w+ b"
But since you do not have "t~" and "w-" generate those are ignored and your syntax is just:
p p > t t~, (t > w+ b, w+ > c b~), (t > w+ b, w+ > c b~)
(madspin does not raise warning for that because you would have a lot of false alarm if MS was doing it).

So far we do not have a way to do semi-leptonic automatically. The official solution is to split the sample in two (with some poisson distribution) and then run each mode.

But for the fun, I have tried the following patch:
https://github.com/mg5amcnlo/mg5amcnlo/tree/3.5.0_ms

which allow the following MG5aMC script (with the command to edit the madspin_card automatically (hence the --add which allows to not overwrite the decay line for the t since one is already existing):
generate p p > t t~
output
launch
madspin=on
set fixed_scale 400
decay t > w+ b, w+ > e+ ve @1
decay t > w+ b, w+ > j j @2 --add
decay t~ > w- b~, w- > j j @1
decay t~ > w- b~, w- > e- ve~ @2 --add

those command does generate the madspin_card as
decay w+ > all all
decay w- > all all
decay z > all all
# running the actual code

decay t > w+ b, w+ > e+ ve @1
decay t > w+ b, w+ > j j @2
decay t~ > w- b~, w- > j j @1
decay t~ > w- b~, w- > e- ve~ @2
launch

As you can guess the @1/@2 in the decay syntax allows to correlate those decay together to reach the effect that you want.

Now, this seems to work a bit too easily to my point of view, cross-section seems correct (with no modification which is surprising).
and I have not found one event with two lepton (which is expected with the above patch).

So in other word, do not use this blindly. It is certainly not a feature that I would consider as validated even for tt~ production at LO.
Adding MLM (which is your example case) in top of it might break some symmetry that make it work in the above example/...
When I have time, I will start a proper validation procedure but this will certainly not be part of 3.5.0 (since first alpha release is schedule for this or next week).

If you try it and have success (or issue) do not hesitate to comment here or on github.

Cheers,

Olivier

Thank You for your quick response.
Unfortunately, I will not be able to test it right now (I am forced to use the madgraph/madspin version embedded in CMSSW), but I will try it as soon as I can.