Our most valuable asset isn’t a machine.
It’s a library: genuine signatures we can trust, each filed by writer, date, pen and situation. Every signature we examine is compared against it, and it grows with every careful addition.
A hand changes over a lifetime
Signatures drift with age, speed and circumstance. So exemplars are filed by date, and a signature is judged against the writer’s hand in the years it was supposedly written, not against the version most often shown online.
30 genuine examples within two years of 1998. Enough to see her range for those years.
- Writer
- Eleanor Hale (made up)
- Date
- 1998
- Pen
- Fountain pen
- Situation
- Signing table
- Status
- Reference: validated
An illustration with a made-up writer and made-up counts.
What gets inA strict door
The market is full of forgeries copied from forgeries. One bad example in the library would make every later comparison weaker, so only signatures of undisputed origin, or independently validated, become references.
Reference
Context only
Not admitted
Four candidates, three possible outcomes.
Why size mattersMore examples, sharper answers
With a handful of examples, almost anything could be the writer’s natural variation. With many, her range comes into focus, and a signature that falls outside it can’t hide.
Inside the range we can draw from so few examples. It might be her natural variation; we can’t yet tell.
What every new reference adds
Natural variation
A clearer picture of how far the writer’s genuine signatures wander from one another.
Chronology
A finer record of how the hand developed, year by year.
Forgery patterns
A better sense of the mistakes forgers of this writer tend to make.
It can’t be bought.
Anyone can buy a microscope or an off-the-shelf AI model. Nobody can buy years of carefully validated signatures. That library is what our comparisons stand on.
How the comparison works