Ch5: Fix python code in Priceline showdown example

Name the "sp3" plot so it may be referenced in the label.
Fix indexing assignment error in showdown_loss() by creating a "close_mask"
intermediate boolean indexing variable.
Fix risk assignment -- values over by less than $250 were not being assigned the risk value. Ensure this happens by assigning loss[ix] = risk last.
Outputs and plots are not changed.

Author: mbauman

Chapter5_LossFunctions/LossFunctions.ipynb

@@ -208,7 +208,7 @@
       "\n",
       "subplot(313)\n",
       "x = np.linspace(0, 25000, 200)\n",
-      "plt.fill_between(x , 0, norm_pdf( x, 12000, 3000), \n",
+      "sp3 = plt.fill_between(x , 0, norm_pdf( x, 12000, 3000), \n",
       "                color = \"#7A68A6\", lw = 3, alpha = 0.6,\n",
       "                 label = \"Trip price guess\")\n",
       "plt.autoscale( tight=True)\n",
@@ -357,9 +357,10 @@
       "def showdown_loss( guess, true_price, risk = 80000):\n",
       "        loss = np.zeros_like( true_price )\n",
       "        ix = true_price < guess\n",
-      "        loss[ix] = risk\n",
       "        loss[~ix] = np.abs( guess - true_price[~ix] )\n",
-      "        loss[ abs( true_price - guess) <= 250 ] = -2*true_price\n",
+      "        close_mask = [ abs( true_price - guess) <= 250 ]\n",
+      "        loss[close_mask] = -2*true_price[close_mask]\n",
+      "        loss[ix] = risk\n",
       "        return loss\n",
       "\n",
       "\n",