Added a couple of changes from the discussion

Author: Scott Hendrickson

bandit-algorithms-101/bandit-algorithms-101.ipynb

@@ -181,7 +181,7 @@
    "outputs": [],
    "source": [
     "# 1 arm\n",
-    "payoff = [0,1]\n",
+    "payoff = [-0.1,0.5]\n",
     "a = np.bincount(binomial(1,0.5,100))\n",
     "print np.dot(a, payoff)"
    ]
@@ -342,21 +342,13 @@
     "means = [0.1, 0.1, 0.9]\n",
     "reward = np.zeros(trials)\n",
     "for m in means:\n",
+    "    # equal rewards of 1 or 0\n",
     "    reward += binomial(1,m,trials)\n",
     "df = pd.DataFrame({\"reward\":reward, \"fair_reward\":binomial(3,0.5,trials)})\n",
     "df.hist()\n",
     "plt.show()"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Wait what?\n",
-    "\n",
-    "* 4  vs. 3?"
-   ]
-  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -546,6 +538,17 @@
     "ggplot(aes(x=\"T\",y=\"Reward\", color=\"Sim\"), data=a) + geom_line()"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false
+   },
+   "outputs": [],
+   "source": [
+    "ggplot(aes(x=\"T\",y=\"CumulativeReward\", color=\"Sim\"), data=a) + geom_line()"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -621,15 +624,6 @@
    "source": [
     "The end"
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": []
   }
  ],
  "metadata": {