The Third Umpire · Lesson 7 — Cutting the Tape ← Course

Cutting the tape — chunking.

You don't hand the umpire the whole match. You cut it into clips, and pull only the one that matters.

The archive is not made of whole documents. It is made of small passages called chunks — every document is sliced into these up front, before anything is ever searched.

Why? A full match report might run to hundreds of lines. If the search could only return the entire report, you would blow the context budget (Lesson 2) and bury the one relevant sentence in noise. Cut it into chunks, and the search can hand back just the paragraph you need.

You would not replay the whole day's play to rule on one delivery. You cut the tape into clips — one per moment — and call up only the clip that matters. Same idea.

Chunk size is a judgement call: too big and each piece is unfocused and wasteful; too small and it loses the context around it. A paragraph, or a few sentences, is the usual sweet spot.

Pick a question and watch the search pull just the one clip it needs out of the whole report — and see how much it left behind.

One report, cut into five clips. Pick a question and see which single clip is pulled.
the match report · cut into chunks

Cut once, retrieve the piece.

Splitting happens up front, when the document is added. At question time the search returns just the matching chunk — not the whole file. 


      

    


    

      
Small clips, not whole tapes.

      

        Documents are split into chunks so the search can return just the passage that
        answers the question — focused, and small enough to fit the window. You retrieve the
        paragraph, not the book.
      

      

        Chunk size is a trade-off: big enough to keep the surrounding context, small enough
        to stay focused and fit the budget. Get it right and every retrieval is a clean,
        relevant clip.
      

      

        Next: the wrong footage →
        
      

    


  


  

    

      ▸ Go deeper — getting chunking right optional
      

        
Why not just retrieve the whole document

        

          Two reasons. It costs budget — a giant document eats the window (Lesson 2). And it
          buries the signal — the one relevant line sits in a sea of irrelevant ones, and the
          model attends worse to a big blob than to a tight, on-point passage.
        

        
Size and overlap

        

          Too big and a chunk is unfocused; too small and it loses the context a fact needs to
          make sense. Chunks often overlap a little at their edges, so a sentence split
          across a boundary is not lost from both halves.
        

        
Cut on natural boundaries

        

          Splitting blindly every N characters can slice a sentence in half. Better to cut on
          structure — paragraphs, sections, headings — so each chunk is one coherent thought
          that stands on its own.
        

        
Tag each chunk

        

          Store metadata alongside each chunk — which document it came from, the section, the
          date. That is what lets the system cite a source, filter by recency, and tell you
          exactly where an answer came from.