diff --git a/R/LearnerClassifXgboost.R b/R/LearnerClassifXgboost.R
index 18a545a9..153f4fff 100644
--- a/R/LearnerClassifXgboost.R
+++ b/R/LearnerClassifXgboost.R
@@ -32,10 +32,11 @@
 #'
 #' @section Early Stopping and Validation:
 #' In order to monitor the validation performance during the training, you can set the `$validate` field of the Learner.
-#' For information on how to configure the valdiation set, see the *Validation* section of [`mlr3::Learner`].
+#' For information on how to configure the valdiation set, see the *Validation* section of [mlr3::Learner].
 #' This validation data can also be used for early stopping, which can be enabled by setting the `early_stopping_rounds` parameter.
 #' The final (or in the case of early stopping best) validation scores can be accessed via `$internal_valid_scores`, and the optimal `nrounds` via `$internal_tuned_values`.
 #' The internal validation measure can be set via the `eval_metric` parameter that can be a [mlr3::Measure], a function, or a character string for the internal xgboost measures.
+#' Using an [mlr3::Measure] is slower than the internal xgboost measures, but allows to use the same measure for tuning and validation.
 #'
 #' @templateVar id classif.xgboost
 #' @template learner
diff --git a/man/mlr_learners_classif.xgboost.Rd b/man/mlr_learners_classif.xgboost.Rd
index faa2c7ec..76820437 100644
--- a/man/mlr_learners_classif.xgboost.Rd
+++ b/man/mlr_learners_classif.xgboost.Rd
@@ -27,10 +27,9 @@ See \url{https://xgboost.readthedocs.io/en/stable/build.html#building-with-gpu-s
 \item \code{nrounds}:
 \itemize{
 \item Actual default: no default.
-\item Adjusted default: 1.
-\item Reason for change: Without a default construction of the learner
-would error. Just setting a nonsense default to workaround this.
-\code{nrounds} needs to be tuned by the user.
+\item Adjusted default: 1000.
+\item Reason for change: Without a default construction of the learner would error.
+The lightgbm learner has a default of 1000, so we use the same here.
 }
 \item \code{nthread}:
 \itemize{
@@ -50,10 +49,11 @@ would error. Just setting a nonsense default to workaround this.
 \section{Early Stopping and Validation}{
 
 In order to monitor the validation performance during the training, you can set the \verb{$validate} field of the Learner.
-For information on how to configure the valdiation set, see the \emph{Validation} section of \code{\link[mlr3:Learner]{mlr3::Learner}}.
+For information on how to configure the valdiation set, see the \emph{Validation} section of \link[mlr3:Learner]{mlr3::Learner}.
 This validation data can also be used for early stopping, which can be enabled by setting the \code{early_stopping_rounds} parameter.
 The final (or in the case of early stopping best) validation scores can be accessed via \verb{$internal_valid_scores}, and the optimal \code{nrounds} via \verb{$internal_tuned_values}.
 The internal validation measure can be set via the \code{eval_metric} parameter that can be a \link[mlr3:Measure]{mlr3::Measure}, a function, or a character string for the internal xgboost measures.
+Using an \link[mlr3:Measure]{mlr3::Measure} is slower than the internal xgboost measures, but allows to use the same measure for tuning and validation.
 }
 
 \section{Dictionary}{
diff --git a/man/mlr_learners_regr.xgboost.Rd b/man/mlr_learners_regr.xgboost.Rd
index b6aca597..0d7ed7e6 100644
--- a/man/mlr_learners_regr.xgboost.Rd
+++ b/man/mlr_learners_regr.xgboost.Rd
@@ -56,7 +56,6 @@ lrn("regr.xgboost")
    eta \tab numeric \tab 0.3 \tab  \tab \eqn{[0, 1]}{[0, 1]} \cr
    eval_metric \tab untyped \tab "rmse" \tab  \tab - \cr
    feature_selector \tab character \tab cyclic \tab cyclic, shuffle, random, greedy, thrifty \tab - \cr
-   feval \tab untyped \tab NULL \tab  \tab - \cr
    gamma \tab numeric \tab 0 \tab  \tab \eqn{[0, \infty)}{[0, Inf)} \cr
    grow_policy \tab character \tab depthwise \tab depthwise, lossguide \tab - \cr
    interaction_constraints \tab untyped \tab - \tab  \tab - \cr
@@ -110,10 +109,11 @@ lrn("regr.xgboost")
 \section{Early Stopping and Validation}{
 
 In order to monitor the validation performance during the training, you can set the \verb{$validate} field of the Learner.
-For information on how to configure the valdiation set, see the \emph{Validation} section of \code{\link[mlr3:Learner]{mlr3::Learner}}.
+For information on how to configure the valdiation set, see the \emph{Validation} section of \link[mlr3:Learner]{mlr3::Learner}.
 This validation data can also be used for early stopping, which can be enabled by setting the \code{early_stopping_rounds} parameter.
 The final (or in the case of early stopping best) validation scores can be accessed via \verb{$internal_valid_scores}, and the optimal \code{nrounds} via \verb{$internal_tuned_values}.
 The internal validation measure can be set via the \code{eval_metric} parameter that can be a \link[mlr3:Measure]{mlr3::Measure}, a function, or a character string for the internal xgboost measures.
+Using an \link[mlr3:Measure]{mlr3::Measure} is slower than the internal xgboost measures, but allows to use the same measure for tuning and validation.
 }
 
 \section{Initial parameter values}{
@@ -122,10 +122,9 @@ The internal validation measure can be set via the \code{eval_metric} parameter
 \item \code{nrounds}:
 \itemize{
 \item Actual default: no default.
-\item Adjusted default: 1.
-\item Reason for change: Without a default construction of the learner
-would error. Just setting a nonsense default to workaround this.
-\code{nrounds} needs to be tuned by the user.
+\item Adjusted default: 1000.
+\item Reason for change: Without a default construction of the learner would error.
+The lightgbm learner has a default of 1000, so we use the same here.
 }
 \item \code{nthread}:
 \itemize{