diff --git a/README.md b/README.md
index ef3cdf8d..2b908283 100644
--- a/README.md
+++ b/README.md
@@ -9,6 +9,8 @@ Pytorch Implementation of Real Space Quantum Monte Carlo Simulations of Molecula
 ## Documentation 
 https://qmctorch.readthedocs.io/en/latest/intro.html
 
+
+<!-- 
 ## Introduction
 
 QMCTorch allows to leverage deep learning to optimize QMC wave functions. The package offers solutions to optimize particle-in-a-box model as well as molecular systems. It uses `pytorch` as a deep learning framework and `pyscf` or `adf` to obtain the basis set information and the first guess of the molecular orbitals.
@@ -128,4 +130,4 @@ Note that comfiguring the solver to perform a geometry optimization is done in o
 
 <p align="center">
 <img src="./pics/h2o.gif" title="Wave function Ooptimization of a H2 molecule">
-</p>
+</p> -->
diff --git a/qmctorch/sampler/walkers.py b/qmctorch/sampler/walkers.py
index 60743a9e..7c3bb558 100644
--- a/qmctorch/sampler/walkers.py
+++ b/qmctorch/sampler/walkers.py
@@ -66,7 +66,7 @@ def _init_center(self):
         Returns:
             torch.tensor: positions of the walkers
         """
-        eps = 1E-6
+        eps = 1E-3
         pos = -eps + 2 * eps * \
             torch.rand(self.nwalkers, self.nelec * self.ndim)
         return pos.type(
@@ -131,7 +131,7 @@ def _init_atomic(self):
                 elif nelec_placed[_idx] < 5:
                     s = 2. / (self.init_domain['atom_num'][_idx] - 2)
                 else:
-                    s = 3. / (self.init_domain['atom_num'][_idx] - 10)
+                    s = 3. / (self.init_domain['atom_num'][_idx] - 3)
                 xyz[ielec,
                     :] += np.random.normal(scale=s, size=(1, 3))
                 nelec_placed[_idx] += 1
diff --git a/qmctorch/utils/plot_data.py b/qmctorch/utils/plot_data.py
index 496fd718..0507e810 100644
--- a/qmctorch/utils/plot_data.py
+++ b/qmctorch/utils/plot_data.py
@@ -74,12 +74,12 @@ def plot_data(observable, obsname):
     plt.show()
 
 
-def plot_walkers_traj(eloc, traj_index='all'):
+def plot_walkers_traj(eloc, walkers='mean'):
     """Plot the trajectory of all the individual walkers
 
     Args:
         obs (SimpleNamespace): Namespace of the observables
-        traj_index (int, str, optional): all or index of a given walker Defaults to 'all'
+        walkers (int, str, optional): all, mean or index of a given walker Defaults to 'all'
 
     Returns:
         float :  Decorelation time
@@ -97,21 +97,28 @@ def plot_walkers_traj(eloc, traj_index='all'):
 
     Tc = (var_decor / var)**2
 
-    if traj_index is not None:
-        plt.subplot(1, 2, 1)
-
-        if traj_index == 'all':
+    if walkers is not None:
+        # plt.subplot(1, 2, 1)
 
+        if walkers == 'all':
             plt.plot(eloc, 'o', alpha=1 / nwalkers, c='grey')
             cmap = cm.hot(np.linspace(0, 1, nwalkers))
             for i in range(nwalkers):
                 plt.plot(celoc.T[:, i], color=cmap[i])
+
+        elif walkers == 'mean':
+            plt.plot(eloc, 'o', alpha=1 / nwalkers, c='grey')
+            plt.plot(np.mean(celoc.T, axis=1), linewidth=5)
+
         else:
-            plt.plot(eloc[traj_index, :], 'o',
+            plt.plot(eloc[walkers, :], 'o',
                      alpha=1 / nwalkers, c='grey')
             plt.plot(celoc.T[traj_index, :])
-        plt.subplot(1, 2, 2)
-        plt.hist(Tc)
+        plt.grid()
+        plt.xlabel('Monte Carlo Steps')
+        plt.ylabel('Energy (Hartree)')
+        # plt.subplot(1, 2, 2)
+        # plt.hist(Tc)
 
     plt.show()