+Since the snippet is relatively small and it can be discontinuous part of the text, the \r
+text alignment methods described in section~\ref{text_alignment} were insufficient \r
+in decision making over document download. Therefore we chose to compare existence of snippet word tuples\r
+in the suspicious document. For 1-tuples the measure means how many words from the snippet\r
+also exist in the suspicious document. If the snippet contains many common words they may\r
+also occur in many documents. In this case the 1-tuple measurement is little decisive. \r
+\r
+We used 2-tuples measurement, which indicates how many neighbouring word pairs coexist in the snippet and in the suspicious document.\r
+We decided according to this value whether to download the source or not. For the deduction \r
+ of the threshold value we used 4413 search results from various queries according to documents \r
+ in the training corpus. Each resulting document was textually aligned to its corresponding suspicious document.\r
+One similarity represents continuous passage of text alignment similarity as is described in the following section~\ref{text_alignment}.\r
+In this way we obtained 248 similarities in total after downloading all of the 4431 documents.\r
+\r
+The 2-tuples similarity performance is depicted in Figure~\ref{fig:snippet_graph}.\r
+Horizontal axis represents threshold of the 2-tuples similarity percentage between the snippet and the suspicious document.\r
+The graph curves represent obtain resource percentage according to the snippet similarity threshold.\r
+A profitable threshold is the one with the largest distance between those two curves.\r
+We chose threshold of the snippet similarity to 20\%, which in the graph corresponds to 20\% of all\r
+downloads and simultaneously with 70\% discovered similarities.\r
+ \r
+\subsection{Source Retrieval Results}\r
+In PAN 2013 Source Retrieval subtask we competed with other 8 teams. \r
+There can not be selected the best approach because there were several independent\r
+performance measures. Possibly each approach has its pros and cons and many approaches\r
+are usable in different situations. \r
+\r
+We believe that in the realistic plagiarism detection the most important is keeping the number of\r
+queries low and simultaneously maximizing recall. \r
+% It is often some tradeoff between cost and efectivness.\r
+It is also advisable to keep the number of downloads down, but on the other hand,\r
+it is relatively cheep and easily scalable operation.\r
+\r
+Our approach had the second best ration of recall to the number of used queries, which\r
+tells about the query efficacy. The approach with the best ratio used few queries (4.9 queries per document which\r
+was 0.4 of the amount we used), but also obtained the lowest recall (0.65 of our recall).\r
+The approach with highest recall (and also lowest precision) achieved 2.8 times higher recall with 3.9 times more queries compared to ours.\r
+\r
+Our approach achieved also low precision, which means we reported many more results and they\r
+were not considered as correct hits. On the other hand each reported result contained some\r
+textual similarity according to text alignment subtask score, which we believe is still worthwhile to report.\r