FFugitiveAI
Menu

CASE 020 · Independent Replication Analysis

Independent Replication

Lock the method before the held-out packet. Verify the initial control, decoder, and intervention, then apply the identical rule to the separately authored packet before claiming replication.

Advanced 525 pts Hint penalty: 75 pts

Artifact

Inspect this carefully

Replication exercise: use one method chosen before the held-out evidence is interpreted. The Initial packet is enough to establish a convincing channel and intervention result, but it is not enough to claim REPLICATED.

Locked method — declare this before evaluating either packet
  1. Carrier positions: inspect only the single character immediately after the colon at numbered positions 1 through 8 in each record.
  2. Bit assignment: U+0020 SPACE = 0 and U+00A0 NO-BREAK SPACE = 1. Any other code point invalidates that record.
  3. Grouping: read positions 1→8 as one 8-bit byte, most-significant bit first, and accept only uppercase ASCII letters A–Z.
  4. Intervention: replace only U+00A0 with U+0020. A candidate must change from SIGNAL to BASELINE; its packet control must remain BASELINE before and after.
  5. Packet acceptance: decode the five candidate records in listed order. Do not change positions, bit values, byte order, normalization, or acceptance criteria after opening the held-out packet.
  6. Final answer: concatenate the five-letter Initial result followed by the five-letter Held-out result and submit the ten-letter string.

Initial packet

This packet was authored first and is the evidence used to establish the method. Evaluate its matched control and five candidates with the locked rule.

Initial packet matched control1: north | 2: east | 3: south | 4: west | 5: amber | 6: cedar | 7: river | 8: stone
Initial packet candidate 11: north | 2: east | 3: south | 4: west | 5: amber | 6: cedar | 7: river | 8: stone
Initial packet candidate 21: north | 2: east | 3: south | 4: west | 5: amber | 6: cedar | 7: river | 8: stone
Initial packet candidate 31: north | 2: east | 3: south | 4: west | 5: amber | 6: cedar | 7: river | 8: stone
Initial packet candidate 41: north | 2: east | 3: south | 4: west | 5: amber | 6: cedar | 7: river | 8: stone
Initial packet candidate 51: north | 2: east | 3: south | 4: west | 5: amber | 6: cedar | 7: river | 8: stone
RecordOriginalAfter U+00A0 → U+0020
Initial packet matched controlBASELINEBASELINE
Initial packet candidate 1SIGNALBASELINE
Initial packet candidate 2SIGNALBASELINE
Initial packet candidate 3SIGNALBASELINE
Initial packet candidate 4SIGNALBASELINE
Initial packet candidate 5SIGNALBASELINE
Open the separately authored held-out packet

Held-out packet

This packet uses different surface words but the same eight carrier positions, decoder, bit assignment, grouping, intervention, and acceptance criteria. No method changes are allowed after opening it.

Held-out packet matched control1: maple | 2: harbor | 3: quartz | 4: linen | 5: copper | 6: meadow | 7: violet | 8: ember
Held-out packet candidate 11: maple | 2: harbor | 3: quartz | 4: linen | 5: copper | 6: meadow | 7: violet | 8: ember
Held-out packet candidate 21: maple | 2: harbor | 3: quartz | 4: linen | 5: copper | 6: meadow | 7: violet | 8: ember
Held-out packet candidate 31: maple | 2: harbor | 3: quartz | 4: linen | 5: copper | 6: meadow | 7: violet | 8: ember
Held-out packet candidate 41: maple | 2: harbor | 3: quartz | 4: linen | 5: copper | 6: meadow | 7: violet | 8: ember
Held-out packet candidate 51: maple | 2: harbor | 3: quartz | 4: linen | 5: copper | 6: meadow | 7: violet | 8: ember
RecordOriginalAfter U+00A0 → U+0020
Held-out packet matched controlBASELINEBASELINE
Held-out packet candidate 1SIGNALBASELINE
Held-out packet candidate 2SIGNALBASELINE
Held-out packet candidate 3SIGNALBASELINE
Held-out packet candidate 4SIGNALBASELINE
Held-out packet candidate 5SIGNALBASELINE

Replication criterion: only after both packet controls behave normally and both five-record candidate sets pass the same locked rule may the evidence be promoted to REPLICATED. This is a repository-owned educational exercise, not a live detector or external system test.

Machine View Inspect alternate representations

Compare the rendered artifact above with representations derived from the same browser-delivered artifact. Machine View does not decode, score, or identify the hidden message.

Opening a representation may expose clues that are already present in the page source or DOM.

Source Browser serialization 

            

            

                Text DOM textContent
                

            

            

                Characters Unicode code points
                

            

        


        

            
Submit recovered message
Server-side validation

            

                
                

                    
                    
                

                

                                    

            

            

                
                0 hints used
            

            
    
        
    

        

        
    
            
    About the technique
    
            
    Independent Replication Analysis
    
            
    Independent replication is the final rung in FugitiveAI’s educational Evidence Ladder. A result becomes stronger when a method selected on initial evidence survives a separately authored held-out packet without changing the extraction rule after seeing the new data. This exercise keeps the content harmless and deterministic while demonstrating why a convincing first sample, even with causal intervention evidence, should not automatically be labeled replicated.
    
            
    Read the related field guide →
    
        
    

        
    
            
    Questions
    
            
    Frequently asked
    
                            
    
                    Why is the Initial packet not enough for REPLICATED?
                    
    Because the extraction method was established on that evidence. Replication asks whether the same predeclared method survives new held-out evidence rather than only the sample that motivated or confirmed the original finding.
    
                
    
                            
    
                    Can I change the bit mapping for the Held-out packet if the first attempt looks wrong?
                    
    No. Changing carrier positions, bit values, byte order, normalization, or acceptance criteria after seeing held-out evidence would turn replication into a new post-hoc search.
    
                
    
                            
    
                    Does replication prove malicious intent?
                    
    No. In FugitiveAI, REPLICATED describes evidentiary strength for the demonstrated machine-readable relationship. It does not establish malicious intent, external receiver use, or real-world attribution.