Florian Hartig

@bbolker provides some code to exemplify the (??possible??) problem, see below. ``` library(glmmTMB) library(DHARMa) ## version 0.3.3.0 library(bbmle) ## for AICtab ## data from Dobson and Barnett dd

OK, so if I fit the same data with a gam, residuals look fine. Note that simulations from family = nb are not yet implemented in mgcv, so we can...

p.s.: as we fit against time, probably good to test for temporal autocorrelation as well, but this was negative `testTemporalAutocorrelation(res)` 

Hi Ben, do you mean in qgam (for the residuals) or in the gam?

OK, I have made some further simulations (see below). Regardless of the previous comments (which is that I think it's not unreasonable that a gam fits better), it is true...

wups, @Istalan, thanks for that, classic error from me (classic in the sense I from questions I realised that users also forget to specify the time all the time). OK,...

@bbolker, some additional, more fundamental thoughts: regardless of the specific (possible) type I error and smoothing settings in qgam, I have been contemplating for quite some time about how much...

Note: gamlss seems to suggest a similar procedure in https://www.rdocumentation.org/packages/gamlss/versions/5.1-4/topics/rqres.plot

D&S comment that they recommend several plots. I think what's easier to do is to do several simulations, and then plot with densities, and adjust weights in all tests or...

As the error message says, there seems to be something going wrong in the glmer simulations, so this is an lme4, not a DHARMa problem. It's not strange that it...