(calibration_markov_chain_results_residuals)=
# Drawing the residuals

The residuals can be visualised with the standard `drawResiduals` method of any `TCalibration` object. 
Its arguments and options have already been detailed in 
[](#calibration_classes_functions_observations_calibration_classes_draw_parameter), and its prototype 
is recalled here for convenience:

````{only} cpp
```cpp
void drawResiduals(TString sTitre, const char *variable = "*", const char *select = "1>0", Option_t * option = "");
```
````

````{only} py
```python
drawResiduals(sTitre, variable = "*", select = "1>0", option = "")
```
````

Since no specific implementation has been added for MCMC calibration, the main point to clarify is 
how the *a priori* and *a posteriori* configurations are defined:

- ***A priori* configuration:** obtained from the parameter initialisation. This can either be 
provided explicitly by the user through the method described in 
[](#calibration_markov_chain_mcmc_properties_init_process), or, if not specified, generated 
from a random draw according to the given *a priori* probability density.

- ***A posteriori* configuration:** by default, the mean of the posterior distribution is used 
(as discussed in the constructor section of this class, see [](#calibration_markov_chain_tmcmc)). 
If the option **"mode"** is specified when constructing the `TMCMC` object, then the posterior mode 
is estimated and used instead.

An example of this function is shown in {numref}`usecases_calibrationMCMCLinReg_Res`.