diff --git a/CHANGELOG.md b/CHANGELOG.md
index 4a68470899..dfde763ac2 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -15,7 +15,7 @@ but cannot always guarantee backwards compatibility. Changes that may **break co
   - 🚀🚀 Optimized `historical_forecasts()` for pre-trained `TorchForecastingModel` running up to 20 times faster than before!. [#2013](https://github.com/unit8co/darts/pull/2013) by [Dennis Bader](https://github.com/dennisbader).
   - Added callback `darts.utils.callbacks.TFMProgressBar` to customize at which model stages to display the progress bar. [#2020](https://github.com/unit8co/darts/pull/2020) by [Dennis Bader](https://github.com/dennisbader).
 - Improvements to documentation:
-  - Adapted the example notebooks to properly apply data transformers and avoid look-ahead bias. [#2020](https://github.com/unit8co/darts/pull/2020) by [Samriddhi Singh](https://github.com/SimTheGreat).
+  - Adapted the example notebooks to properly apply data transformers and avoid look-ahead bias. [#2020](https://github.com/unit8co/darts/pull/2020) by [Samriddhi Singh](https://github.com/SimTheGreat). 
 - Improvements to Regression Models:
   - `XGBModel` now leverages XGBoost's native Quantile Regression support that was released in version 2.0.0 for improved probabilistic forecasts. [#2051](https://github.com/unit8co/darts/pull/2051) by [Dennis Bader](https://github.com/dennisbader).
 - Other improvements:
@@ -24,6 +24,7 @@ but cannot always guarantee backwards compatibility. Changes that may **break co
 **Fixed**
 - Fixed a bug when calling optimized `historical_forecasts()` for a `RegressionModel` trained with unequal component-specific lags. [#2040](https://github.com/unit8co/darts/pull/2040) by [Antoine Madrona](https://github.com/madtoinou).
 - Fixed a bug when using encoders with `RegressionModel` and series with a non-evenly spaced frequency (e.g. Month Begin). This raised an error during lagged data creation when trying to divide a pd.Timedelta by the ambiguous frequency. [#2034](https://github.com/unit8co/darts/pull/2034) by [Antoine Madrona](https://github.com/madtoinou).
+- Fixed an issue where `TCNModel` training included the last (input_chunk_length - output_chunk_length) target points in the loss computation. [#2006](https://github.com/unit8co/darts/pull/2006) by [Dennis Bader](https://github.com/dennisbader).
 
 ### For developers of the library:
 
diff --git a/darts/models/forecasting/pl_forecasting_module.py b/darts/models/forecasting/pl_forecasting_module.py
index 79c1902a81..565031c85b 100644
--- a/darts/models/forecasting/pl_forecasting_module.py
+++ b/darts/models/forecasting/pl_forecasting_module.py
@@ -186,13 +186,6 @@ def __init__(
         self.pred_n_jobs: Optional[int] = None
         self.predict_likelihood_parameters: Optional[bool] = None
 
-    @property
-    def first_prediction_index(self) -> int:
-        """
-        Returns the index of the first predicted within the output of self.model.
-        """
-        return 0
-
     @abstractmethod
     def forward(self, *args, **kwargs) -> Any:
         super().forward(*args, **kwargs)
@@ -600,9 +593,7 @@ def _get_batch_prediction(
             dim=dim_component,
         )
 
-        out = self._produce_predict_output(x=(input_past, static_covariates))[
-            :, self.first_prediction_index :, :
-        ]
+        out = self._produce_predict_output(x=(input_past, static_covariates))
 
         batch_prediction = [out[:, :roll_size, :]]
         prediction_length = roll_size
@@ -649,9 +640,7 @@ def _get_batch_prediction(
                 ] = future_past_covariates[:, left_past:right_past, :]
 
             # take only last part of the output sequence where needed
-            out = self._produce_predict_output(x=(input_past, static_covariates))[
-                :, self.first_prediction_index :, :
-            ]
+            out = self._produce_predict_output(x=(input_past, static_covariates))
 
             batch_prediction.append(out)
             prediction_length += self.output_chunk_length
@@ -783,9 +772,7 @@ def _get_batch_prediction(
             )
         )
 
-        out = self._produce_predict_output(x=(input_past, input_future, input_static))[
-            :, self.first_prediction_index :, :
-        ]
+        out = self._produce_predict_output(x=(input_past, input_future, input_static))
 
         batch_prediction = [out[:, :roll_size, :]]
         prediction_length = roll_size
@@ -853,7 +840,7 @@ def _get_batch_prediction(
             # take only last part of the output sequence where needed
             out = self._produce_predict_output(
                 x=(input_past, input_future, input_static)
-            )[:, self.first_prediction_index :, :]
+            )
 
             batch_prediction.append(out)
             prediction_length += self.output_chunk_length
diff --git a/darts/models/forecasting/tcn_model.py b/darts/models/forecasting/tcn_model.py
index 076fd939df..91a9180425 100644
--- a/darts/models/forecasting/tcn_model.py
+++ b/darts/models/forecasting/tcn_model.py
@@ -4,7 +4,7 @@
 """
 
 import math
-from typing import Optional, Sequence, Tuple
+from typing import Optional, Tuple
 
 import torch
 import torch.nn as nn
@@ -16,8 +16,6 @@
     io_processor,
 )
 from darts.models.forecasting.torch_forecasting_model import PastCovariatesTorchModel
-from darts.timeseries import TimeSeries
-from darts.utils.data import PastCovariatesShiftedDataset
 from darts.utils.torch import MonteCarloDropout
 
 logger = get_logger(__name__)
@@ -139,14 +137,12 @@ def __init__(
         weight_norm: bool,
         target_size: int,
         nr_params: int,
-        target_length: int,
         dropout: float,
         **kwargs
     ):
 
         """PyTorch module implementing a dilated TCN module used in `TCNModel`.
 
-
         Parameters
         ----------
         input_size
@@ -155,8 +151,6 @@ def __init__(
             The dimensionality of the output time series.
         nr_params
             The number of parameters of the likelihood (or 1 if no likelihood is used).
-        target_length
-            Number of time steps the torch module will predict into the future at once.
         kernel_size
             The size of every kernel in a convolutional layer.
         num_filters
@@ -179,10 +173,8 @@ def __init__(
 
         Outputs
         -------
-        y of shape `(batch_size, input_chunk_length, target_size, nr_params)`
-            Tensor containing the predictions of the next 'output_chunk_length' points in the last
-            'output_chunk_length' entries of the tensor. The entries before contain the data points
-            leading up to the first prediction, all in chronological order.
+        y of shape `(batch_size, output_chunk_length, target_size, nr_params)`
+            Tensor containing the predictions of the next 'output_chunk_length' points.
         """
 
         super().__init__(**kwargs)
@@ -191,7 +183,6 @@ def __init__(
         self.input_size = input_size
         self.n_filters = num_filters
         self.kernel_size = kernel_size
-        self.target_length = target_length
         self.target_size = target_size
         self.nr_params = nr_params
         self.dilation_base = dilation_base
@@ -249,11 +240,7 @@ def forward(self, x_in: Tuple):
             batch_size, self.input_chunk_length, self.target_size, self.nr_params
         )
 
-        return x
-
-    @property
-    def first_prediction_index(self) -> int:
-        return -self.output_chunk_length
+        return x[:, -self.output_chunk_length :, :, :]
 
 
 class TCNModel(PastCovariatesTorchModel):
@@ -511,25 +498,7 @@ def _create_model(self, train_sample: Tuple[torch.Tensor]) -> torch.nn.Module:
             num_filters=self.num_filters,
             num_layers=self.num_layers,
             dilation_base=self.dilation_base,
-            target_length=self.output_chunk_length,
             dropout=self.dropout,
             weight_norm=self.weight_norm,
             **self.pl_module_params,
         )
-
-    def _build_train_dataset(
-        self,
-        target: Sequence[TimeSeries],
-        past_covariates: Optional[Sequence[TimeSeries]],
-        future_covariates: Optional[Sequence[TimeSeries]],
-        max_samples_per_ts: Optional[int],
-    ) -> PastCovariatesShiftedDataset:
-
-        return PastCovariatesShiftedDataset(
-            target_series=target,
-            covariates=past_covariates,
-            length=self.input_chunk_length,
-            shift=self.output_chunk_length,
-            max_samples_per_ts=max_samples_per_ts,
-            use_static_covariates=self.uses_static_covariates,
-        )
diff --git a/darts/models/forecasting/torch_forecasting_model.py b/darts/models/forecasting/torch_forecasting_model.py
index 8312808689..18617615e3 100644
--- a/darts/models/forecasting/torch_forecasting_model.py
+++ b/darts/models/forecasting/torch_forecasting_model.py
@@ -1505,13 +1505,6 @@ def predict_from_dataset(
         # flatten and return
         return [ts for batch in predictions for ts in batch]
 
-    @property
-    def first_prediction_index(self) -> int:
-        """
-        Returns the index of the first predicted within the output of self.model.
-        """
-        return 0
-
     @property
     def min_train_series_length(self) -> int:
         """
diff --git a/docs/source/conf.py b/docs/source/conf.py
index dd4017a787..45ea057f1a 100644
--- a/docs/source/conf.py
+++ b/docs/source/conf.py
@@ -53,7 +53,7 @@
     + "PastCovariatesTorchModel,FutureCovariatesTorchModel,DualCovariatesTorchModel,MixedCovariatesTorchModel,"
     + "SplitCovariatesTorchModel,TorchParametricProbabilisticForecastingModel,"
     + "min_train_series_length,"
-    + "untrained_model,first_prediction_index,future_covariate_series,past_covariate_series,"
+    + "untrained_model,future_covariate_series,past_covariate_series,"
     + "initialize_encoders,register_datapipe_as_function,register_function,functions,"
     + "SplitTimeSeriesSequence,randint,AnomalyModel",
 }