Daniel Schoenig

For a GAM, any kind of pattern in the residual plots may be the result of choosing the basis dimension for a smooth term too low. In the original call...

> I should add though that I wasn't able to re-create the problem described in #43 when simulating with a gam, possible predict.gam excludes the REs? > That's interesting! I...

Hi Florian, I think using the `ff` package would be the best way to implement this feature, as it allows a large simulation matrix to be stored on disk and...

For testing, this repurposes the example in the vignette to produce a simulation matrix with 1 billion elements. ``` testData = createData(sampleSize = 5e5, overdispersion = 1.5, family = poisson())...

Hi Florian, No worries, I needed the ff implementation for a project so it wasn't too much work to put some of this into DHARMa. I completely agree with you...