Probability and confidence documentation records not only what an authenticator currently believes, but how strongly the evidence supports each claim, how dependable the assessment is, where its limits lie and what future evidence could change it. It replaces a bare verdict such as "authentic" with a reviewable account of what is supported, how strongly it is supported and why.
This distinction matters because authentication conclusions tend to harden as they travel. A cautious research note can become a definite catalogue entry, then a simplified sale description, then an inherited truth for the next owner. Good documentation prevents that drift by keeping the precise claim, evidential strength, examination scope, uncertainty and revision history attached to one another.
Collector scenario
Two records reach the same view, but only one can be audited
One collector writes: "Authentic - checked against examples." Another writes: "High confidence on period and production method; moderate confidence on exact issue; low confidence on the named-owner story. Construction and printing align with secure references, but ownership begins twelve years after manufacture and the label may be replaced."
Both may believe the object is probably genuine. Only the second record shows which claims are strong, which remain provisional and what evidence would matter next. That is the purpose of probability and confidence documentation: it makes judgement reusable rather than merely persuasive.
Foundations
Probability and confidence are related, but they are not the same
A durable authentication record should distinguish the plausibility of the claim from the dependability of the assessment process.
Probability
How plausible is the claim?
Probability concerns the proposition being tested: for example, whether a signature was written by the named person, whether a card belongs to the claimed print run, or whether a prop was used in the named production.
In most collectible fields, this is a reasoned judgement rather than a statistically measured percentage. The record should therefore state the claim precisely and describe the direction and strength of the evidence without pretending to mathematical certainty.
Confidence
How much trust should be placed in that judgement?
Confidence concerns the quality and stability of the assessment. It depends on the reliability of the evidence, the completeness of the examination, the quality of reference material, the competence and independence of the examiner, and the possibility that relevant alternatives were missed.
A conclusion can be strongly favourable but only moderately confident when the object was seen through photographs, critical areas were inaccessible or the comparison population is weak.
Documentation rule
Record both, because they answer different questions
A high probability judgement does not automatically justify high confidence. Conversely, an excellent examination may produce high confidence that the evidence genuinely remains balanced and the result is inconclusive.
The record should therefore separate the current view of the claim from the dependability of the process that produced it.
Claim definition
Document a precise proposition, not a vague verdict
The question "Is it authentic?" is usually too broad to support careful probability language. An object can be genuine in one respect and misdescribed in another.
Break compound descriptions into separate claims
A description such as "original production item, used by the lead actor, retained by a crew member and signed during filming" contains several propositions. Each may have a different evidential strength and confidence ceiling.
manufactured during the claimed period
made by the claimed manufacturer or studio
belonging to the exact issue or production run
retaining original components
associated with the named person or event
supported by the stated ownership history
The record should state the highest level actually supported. Strong evidence that an object is period-made must not be allowed to imply exact issue, named use, untouched condition or a complete provenance chain.
Evidence judgement
Record evidence by reliability, relevance and direction
Evidence should not merely be listed. The record should show whether it is trustworthy, whether it distinguishes the claim from alternatives and whether it supports, contradicts or leaves the question unchanged.
Strongly discriminating
Evidence that is difficult to explain under realistic alternatives
Examples include production records tied to an unaltered unique number, an archival image matching distinctive damage, or a manufacturing characteristic exclusive to the documented period and process.
Record the source and how it is linked to this object.
Explain which alternative explanations it weakens.
State any assumptions needed before it can carry full weight.
Compatible but weak
Evidence that fits the claim but also fits many alternatives
Correct dimensions, plausible age, general style, common-period materials and a believable seller account may all be consistent with authenticity without distinguishing an original from a competent reproduction or later issue.
Do not call compatibility confirmation.
Do not multiply several copied features into false strength.
Record why the evidence remains non-exclusive.
Contradictory
Evidence more readily explained by another proposition
Examples include a later material, impossible chronology, anachronistic typography, construction inconsistent with the attributed maker or repeated supposedly unique defects across many objects.
Classify the contradiction rather than hiding it.
Test whether repair, replacement or misdescription explains it.
State whether it affects the whole object or only one claim or component.
Neutral
Evidence that does not materially distinguish between propositions
An old frame, professional-looking certificate, prestigious prior owner or packaging that could have been paired later may matter commercially but contribute little to physical authentication.
Record it if relevant to the history of the object.
Do not count it as positive merely because it is consistent with the claim.
Independence
Count evidential pathways, not repeated assertions
Ten observations do not necessarily outweigh one decisive fact, and five sources may still represent one original claim repeated through the market.
Derivative evidence
A seller statement, certificate written by that seller, auction entry based on the certificate and later database record copied from the auction are one evidential lineage, not four independent confirmations.
Document the common source so confidence does not rise simply because the claim has been repeated.
Correlated physical features
Ink character, dot pattern, registration, pressure and paper deformation may all derive from one printing process. They are useful observations, but not automatically five separate proofs.
Strong documentation identifies the shared mechanism and looks for independent support from construction, provenance, archival records or scientific analysis.
Contradictions
Classify anomalies instead of averaging them away
Conflicting evidence should lower confidence until a coherent explanation is established. Not every contradiction is fatal, but none should disappear merely because the overall story is attractive.
Fatal contradiction
The stated proposition cannot be correct
Use only where the contradiction is decisive and securely attached to the object, such as an integral material or technology that did not exist at the claimed date.
Major contradiction
Strongly inconsistent, but a coherent explanation may remain
A replacement component, repair, later annotation or mistaken issue identification may explain the conflict. The record should identify the explanation as tested, plausible or merely speculative.
Minor anomaly
Unexpected but not yet materially damaging
Do not dismiss it as normal variation without evidence. Record the feature, comparison context and reason it is presently treated as minor.
Unresolved irregularity
Insufficient evidence to classify the feature
This wording is useful when access, reference material or method limitations prevent a fair judgement. It preserves the problem for later review without forcing it into a positive or negative category.
Confidence language
Use a verbal scale that describes evidential support
Defined verbal bands are usually more honest than invented percentages. Each term should carry a stable meaning and be attached to the exact proposition it describes.
Confirmed or directly established
The exact proposition is uniquely tied to exceptionally strong evidence and realistic alternatives have been decisively excluded.
Documentation use: Use sparingly and state exactly what is confirmed. Confirmation of date does not automatically confirm maker, use or ownership.
Very strongly supported
The total evidence overwhelmingly favours the claim and realistic alternatives are difficult to reconcile with it.
Documentation use: Residual uncertainty remains, but it is limited and openly stated.
Strongly supported
Several substantial, preferably independent evidence streams favour the claim and no major unresolved contradiction is known.
Documentation use: Use where one or more limitations remain but do not control the conclusion.
Moderately supported
The balance favours the claim, but the case is incomplete, partly non-exclusive or still vulnerable to a plausible alternative.
Documentation use: Appropriate for many well-researched collector judgements that should not be presented as settled fact.
Limited support
Some evidence is compatible with or mildly favours the claim, but it is insufficient for a firm conclusion.
Documentation use: Use where the claim remains possible but should not carry a premium or definitive description without more evidence.
Inconclusive
The available evidence does not meaningfully distinguish between the principal propositions.
Documentation use: This is a valid outcome, not a polite synonym for fake. Explain what prevents resolution.
Evidence weighs against
One or more material observations are more readily explained by an alternative proposition.
Documentation use: State the competing explanation and whether further examination could alter the balance.
Strongly contradicted
The evidence is seriously inconsistent with the claim and strongly favours an alternative.
Documentation use: Use where the claim remains theoretically possible but would require several unsupported exceptions.
Demonstrably false
A decisive contradiction establishes that the exact stated claim cannot be correct.
Documentation use: Specify the decisive contradiction and confirm that the feature is original to, or securely associated with, the object.
Examination scope
Every method creates a confidence ceiling
More persuasive wording cannot overcome the limits imposed by photograph-only review, inaccessible areas, weak references, restricted testing or incomplete provenance.
Listing-level review
Screening from seller descriptions and selected images
Useful for deciding whether to investigate further, but too limited for many definitive conclusions. Record that the opinion is preliminary and dependent on seller-supplied information.
Photograph-based opinion
Visible features can be assessed, hidden or sensory evidence cannot
Record image resolution, lighting, colour uncertainty, absent views, unverified scale and any possibility of manipulation. The conclusion should not exceed what those images can support.
In-hand examination
Direct access expands the evidential base
Weight, surface, tool marks, construction, wear, concealed areas and component relationships may become assessable. Record which areas remained inaccessible or were not dismantled.
Instrumental or multidisciplinary assessment
Technical analysis can answer narrower questions more reliably
Microscopy, imaging, spectroscopy, radiography, paper or ink analysis may strengthen or exclude specific propositions. Record what the method can and cannot establish; old material alone does not prove authorship or association.
Decision context
Authentication probability is not the same as a decision threshold
Different collector decisions can legitimately require different levels of support without changing the underlying evidence.
A collector may retain an uncertain object for study, refuse to pay an attribution premium, insure it only at an unattributed value, avoid naming the alleged maker publicly, or seek further testing before sale. "Not sufficiently established for purchase at this price" does not necessarily mean "probably false."
Document both the evidential conclusion and the decision made from it. This prevents a cautious commercial choice from being mistaken for a negative authentication verdict, or market acceptance from being mistaken for proof.
Collector workflow
A disciplined way to create the record
The following sequence keeps the claim, alternatives, evidence, limitations and revision triggers visible from the beginning.
01
Define the exact claim
Separate identity, date, maker, issue, originality, association, provenance and alteration. Avoid the undefined question "Is it real?"
02
Define realistic alternatives
Include deliberate counterfeit, later official issue, innocent reproduction, composite object, authentic item with false provenance, repaired original and misidentified variant where relevant.
03
Record the starting context
Note production quantity, market frequency, known counterfeit activity, geographical plausibility and any seller or source circumstances that affect the initial plausibility of the claim.
04
List evidence by independent pathway
Separate physical construction, materials, documentary provenance, historical fit, specialist comparison and scientific evidence. Mark derivative sources that repeat the same underlying assertion.
05
Assess reliability and discrimination
For each item, record whether it is trustworthy and how strongly it distinguishes the claim from the most plausible alternative.
06
Investigate contradictions
Classify each conflict and test repair, replacement, contamination, reference error, misdescription or an alternative production history before drawing the final conclusion.
07
Test sensitivity
Ask what happens if the strongest evidence is removed, a disputed assumption changes, or the comparison population proves less secure than expected.
08
State claim-specific conclusions
Give separate conclusions for period, maker, issue, originality, association and provenance rather than allowing the strongest claim to inflate the weakest.
09
State the confidence ceiling
Identify the limits created by inaccessible areas, photograph-only review, weak reference sets, non-destructive testing restrictions or missing records.
10
Record revision triggers
Specify what evidence would raise, lower or resolve confidence, then preserve dated revisions rather than silently overwriting the earlier judgement.
Model collector record
A conclusion that preserves the structure of the evidence
Claim assessed: Factory-produced example from the first 1977 commercial issue, with the accompanying label original to manufacture.
Conclusion: Physical construction, print process and paper characteristics strongly support manufacture during the claimed period. The primary object is therefore strongly supported as a period-original example.
Component qualification: The label differs from secure first-run references and is assessed as a later replacement. This reduces originality but does not, on present evidence, invalidate the underlying object.
Confidence: High regarding period manufacture; moderate regarding exact issue; low regarding the label's originality.
Limitations: Examination was non-destructive, the reverse of the mounted section was inaccessible and the earliest ownership record begins fourteen years after manufacture.
Revision triggers: Confidence would increase if an independently authenticated first-run example with the same construction sequence is documented. It would decrease if the paper or print characteristics are shown to occur commonly in later official reissues.
Failure modes
How confidence records become misleading
Most problems arise not from using cautious language, but from allowing the evidence structure to disappear as the conclusion is copied, sold, inherited or repeated.
Claim drift
A narrow finding becomes a broad commercial statement
Evidence may establish period manufacture but not named ownership, exact use or untouched originality. Keep the supported rung of the claim ladder visible wherever the conclusion is reused.
False independence
One assertion is counted several times
A seller statement, seller-issued certificate, auction description copied from that certificate and later database entry are not four independent sources. Record their shared origin.
Precision theatre
An invented percentage creates authority without data
A statement such as "87% likely authentic" is not more rigorous unless supported by validated models, reliable population data and disclosed assumptions. Carefully defined verbal conclusions are usually safer.
Authority substitution
Market acceptance is mistaken for evidential support
Auction pedigree, encapsulation, publication history and celebrity ownership may affect market confidence. They do not automatically establish the physical or historical claim.
Anomaly laundering
Contradictions are relabelled as variation without evidence
Normal variation must be shown through secure comparison material. An unexplained difference should remain documented as an anomaly until its status is established.
Frozen certainty
A conclusion is preserved but its reasons and limitations disappear
Keep dated evidence, photographs, prior opinions, seller representations and later revisions so future owners can understand why confidence changed.
Documentation checklist
What a durable probability and confidence record should contain
The record can be concise for routine objects, but these fields provide a reliable structure when value, rarity, dispute or future reliance makes the conclusion important.
✓The exact proposition being assessed
✓The principal alternative explanations considered
✓The examination level and date
✓The evidence sources and their independence
✓Reliability and discriminating value of each significant evidence item
✓Contradictions, anomalies and unresolved irregularities
✓Claim-specific conclusion wording
✓Confidence level and reason it does not sit higher or lower
✓The confidence ceiling created by missing access, data or methods
✓What evidence would change the conclusion
✓Revision history, earlier opinions and linked supporting files
Specialist threshold
When the collector should seek specialist help
Escalation is warranted when the consequence of error exceeds what the available evidence and collector expertise can responsibly support.
Escalate the examination
The claim carries a substantial value, insurance or legal consequence.
A major contradiction remains unresolved after ordinary inspection.
The conclusion depends on specialist material, printing, handwriting or manufacturing knowledge.
Testing may be destructive, invasive or capable of damaging the object.
Known counterfeits are sophisticated or rapidly evolving.
Escalate the documentation
The wording will appear in a sale, valuation, exhibition, insurance schedule or formal handover.
Several experts disagree for substantive rather than merely stylistic reasons.
The conclusion depends on a certificate or authority whose scope is unclear.
A previous opinion has been revised and the change must be explained transparently.
The object may enter a public, institutional or scholarly record.
Key takeaways
Document a defined claim, not a vague object-level verdict.
Keep probability of the claim separate from confidence in the assessment.
Judge evidence by reliability and discrimination, not by quantity or compatibility alone.
Record derivative sources and correlated features so confidence is not artificially multiplied.
Give different claims and components different confidence levels where the evidence is uneven.
State contradictions, examination limits and the confidence ceiling openly.
Treat inconclusive as a responsible outcome when the evidence cannot distinguish the principal alternatives.
Preserve dated revisions and state what future evidence would change the judgement.
