diff --git a/tests/html/graham.html b/tests/html/graham.html
index debc47b..85e8a48 100644
--- a/tests/html/graham.html
+++ b/tests/html/graham.html
@@ -19,6 +19,10 @@
             <li><a href="quickhull.html">QuickHull Test Results</a></li>
         </ul>
         <br />
-        <h1>Convex Hull Algorithms - Test Results Comparison</h1>
+        <h1>Graham's Scan test results</h1>
+        <p>
+            Graham's Scan's results were purely dependent on input size as it is
+            test/output independent and takes.
+        </p>
     </body>
 </html>
diff --git a/tests/html/index.html b/tests/html/index.html
index ab899e8..785b838 100644
--- a/tests/html/index.html
+++ b/tests/html/index.html
@@ -29,7 +29,8 @@ <h3>Test5</h3>
         <h3>Test6</h3>
         <p>
             A set of 1000 uniformly spaced (uniform angle spacing) points
-            slightly displaced away or towards the center
+            slightly displaced away or towards the center. 548 points were on
+            the hull
         </p>
         <h3>Test7 (Not used because of unreliable results)</h3>
         <p>
diff --git a/tests/html/jarvis.html b/tests/html/jarvis.html
index 7c8faee..a07758c 100644
--- a/tests/html/jarvis.html
+++ b/tests/html/jarvis.html
@@ -19,6 +19,12 @@
             <li><a href="quickhull.html">QuickHull Test Results</a></li>
         </ul>
         <br />
-        <h1>Convex Hull Algorithms - Test Results Comparison</h1>
+        <h1>Jarvis March test results</h1>
+        <p>
+            Jarvis March worked very efficiently on the random testcases.
+            However, it's results were bad for the 2nd testcase and the 6th
+            testcase. This is because it is and output dependent algorith with
+            complexity O(n*h).
+        </p>
     </body>
 </html>
diff --git a/tests/html/quickhull.html b/tests/html/quickhull.html
index eb8e52c..49a08a1 100644
--- a/tests/html/quickhull.html
+++ b/tests/html/quickhull.html
@@ -19,6 +19,15 @@
             </li>
         </ul>
         <br />
-        <h1>Convex Hull Algorithms - Test Results Comparison</h1>
+        <h1>Quickhull test results</h1>
+        <p>
+            Quickhull had the best overall results. Even though the worse case
+            complexity of quickhull is O(n<sup>2</sup>) it works very fast in
+            average case. It is impossible to create such a case which is big
+            enough on a computer with enough precision. QuickHull performed
+            worse than Graham's Scan in the 2nd testcase only. This was probably
+            because of the skewed nature of a parabola which makes the division
+            of points not equal on both sides of a line
+        </p>
     </body>
 </html>