tweaks the computation of total_inlinks_for_doc, a=chris

diff --git a/lib/index_bundle_iterators/group_iterator.php b/lib/index_bundle_iterators/group_iterator.php
index d2dc756af..11c1ed604 100644
--- a/lib/index_bundle_iterators/group_iterator.php
+++ b/lib/index_bundle_iterators/group_iterator.php
@@ -376,9 +376,23 @@ class GroupIterator extends IndexBundleIterator
                 /*
                     An attempt to approximate the total number of inlinks
                     to a document which will have the terms in question.
+
+                    A result after grouping consists of a document and inlinks
+                    which contain the terms of the base iterator.
+
+                    $hash_count/($num_docs_seen *$num_inlinks)
+                    approximates the probability that an inlink for
+                    a particular document happens to
+                    contain the terms of the iterator. After $num_docs_seen
+                    many documents, there are $num_inlinks - $hash_count
+                    many inlinks which might appear in the remainder of
+                    the iterators document list, giving a value
+                    for the $num_inlinks_not_seen as per the equation below:
                  */
-                $total_inlinks_for_doc = min($num_inlinks,
-                    $hash_count * $this->num_docs/$num_docs_seen);
+                $num_inlinks_not_seen =
+                    ($num_inlinks - $hash_count)*$hash_count/
+                    ($num_docs_seen * $num_inlinks);
+                $total_inlinks_for_doc = $hash_count + $num_inlinks_not_seen;

                 /*
                      we score[x] of the xth inlink for this document