complexity of config files

Just a question: what is the reason for not including all parameters in the search space of the config files?

For example, when looking at this entry in `regressor.py`:
```python
'sklearn.ensemble.ExtraTreesRegressor': {
'n_estimators': [100],
'max_features': np.arange(0.05, 1.01, 0.05),
'min_samples_split': range(2, 21),
'min_samples_leaf': range(1, 21),
'bootstrap': [True, False]
},
```
and comparing it to the [parameters listed in its docs](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.ExtraTreesRegressor.html), one could easily come up with a more elaborate config for it (increasing/decreasing `n_estimators`, using different `criterion`, all the `min` and `max` settings could be given a range of floats, etc.).

Is this an arbitrary decision made for each parameter, based on expected performance vs. computational requirements? **And if so, isn't it better to let the genetic algorithm take care of this decision?**