Weird differences between Single Simulation and Batch simulation

Sienss · July 30, 2020, 4:57pm

Hi !

So i have almost 10k differences between single sim and batch sim.
Can anyone explain me why ? And is it possible to reduce it ?

batch
Single

Any idea ?

yellowfive · July 30, 2020, 5:44pm

There must be something set up differently between the two runs. The easiest way to check that would be to re-run your batch simulation with the Batch Reporting Type set to Log (under the Other section when creating your setup), and also run your single simulation with report type set to Log.

With that, you can click on the row in your batch simulation result and see the full report for it and compare it directly to the single simulation that you ran.

(Note: use the Batch Reporting Type option sparingly and only on very small batches, it will significantly slow down your simulations.)

Sienss · July 31, 2020, 1:12am

Thanks for the instruction !
I’ve done them with logs this time:

Single sim set with logs
Batch done
Sims extracted from the batch

So. I’ve found this :

So i retried to do the batch and still get ‘override’ stats. Is it due to the corruption sytem ?

After sleeping on it, it can’t be corruption because their is a tendancy when i put haste corruption compare to crit corruption. So i have no idea now :x

yellowfive · July 31, 2020, 7:00pm

I’ll take a look and see why the stats aren’t matching up… a batch will always “override” your stats, but it should simply override them to the stats that you have equipped.

yellowfive · July 31, 2020, 9:12pm

The difference in stats seems to be an issue with the Heart of Darkness trait – I think the override is disabling it. I’ll fix that in the next update. It’s kind of an annoying case… in some cases it is a bit ambiguous whether you want the override to include that trait’s stats or not. In this case, I think it’s clear that you want the override to include those stats, so I have to write some stuff so the simulator can tell when it wants to include that and when it doesn’t.

Sienss · July 31, 2020, 10:08pm

Okay !
Personnaly i dont mind the difference, because the point isn’t the number but how corruption are set between the other.
Like knowing that Haste is generally on top in opposition of having crit for example.
Thanks for digging into it !
The result are good then but offset by a bit so i dont mind ! =D
Thanks again !

yellowfive · July 31, 2020, 10:28pm

I just posted an update – issue should be fixed now.