Authenticity is Everything
Has anyone tried this Verimarx Ai tool that's advertised on this site?
Was wondering if anyone here has used it here and had it work well? I can't find much information on this tool, supposedly it detects forgeries, just wanted thoughts on it.
Replies to This Discussion
-
Permalink Reply by Dominick Lisi on
-
You raise a fair question, and I understand why you want clarity on the source material and how the system learns.
To be clear, VeriMarx is not simply pulling random real and fake examples from the internet and treating them all as valid exemplars for direct comparison. When I referred to user-submitted material and open sources, I was not suggesting that every publicly available image is accepted at face value or used as an automatic authentic reference. That would obviously create a major quality-control problem.
A more accurate way to describe it is that VeriMarx can incorporate user-submitted material, along with certain open and available source material, as part of a broader analytical process. Those inputs are not blindly accepted as authoritative. They are assessed, filtered, weighted, and evaluated for consistency and reliability before being given analytical value. So no, that is not the same as indiscriminately scouring the internet and relying on whatever happens to be out there.
At the same time, yes, VeriMarx does retain submitted autograph images so the system can continue to learn and improve over time. That is not materially different from what third-party authenticators do when they retain prior submissions, exemplars, notes, and comparison files as part of developing internal reference material and informing future evaluations.
The important distinction is not simply whether prior images are kept. The important issue is how source material is evaluated and used. Retaining submitted images for learning purposes is very different from blindly harvesting random online images and treating them as trusted exemplars.
I also want to be careful not to overstate or over-disclose proprietary details about exactly how VeriMarx ranks, filters, and weights source material, because that gets into internal methodology. But your concern is a legitimate one: source quality matters enormously in authentication, and separating stronger material from weaker material is a core part of the process.
-
Permalink Reply by Steve Cyrkin, Admin on
-
Dom, Jackie Gleason reminds me. Secretarials: You often don't see general signs of forgery in secretarials signed by people who often proxy-sign for a person. You'll rarely find hesitation, shakiness, and other things VeriMarx probably looks for, because they're experienced and confident.
What you often find, though, are inscriptions commonly signed by a proxy signer that you often don't find from the actual celeb, autographs on photos or other items the celeb rarely signs in public—and in most cases, pre-eBay celebs signed slips of paper, cards, or whatever was around.
-
Permalink Reply by Eric Keith Longo on
-
Good point Steve. I have yet to see Gleason use a dash like Spear did.
-
-
That is a very fair point, and I agree. Secretarials can be especially challenging because they often do not show the obvious signs people associate with forgery. When a proxy signer is experienced, the writing may look smooth, natural, and confident, so the usual red flags are not always there.
That is why the surrounding context matters so much, including the type of item signed, the inscription style, and whether the piece matches what the celebrity was actually known to sign in public. With many pre-eBay celebrities, authentic examples were often on simple paper, cards, or whatever was available, so those historical habits are important.
It is also one of the reasons we were asking the community to help. We want VeriMarx to become a system that is genuinely valuable and accurate, and that only happens by learning from knowledgeable collectors who understand these nuances. Input like this helps us improve the system in the areas where nuance matters most.
-
-
The Jackie Gleason "Challenge" - Eric, we thought it might be interesting to perform our own "Jackie Gleason" submission test to see how VeriMarx would perform. The first text utilized a readily available sample of his signature from his passport. We used this as a basis. As expected, VeriMarx identified this highly consistent with his signature, and did not identify significant forgery indicators.SEE REPORT ending in 8NNW. The second report (Ending in 3A3RY) utilized a submission that presented what we felt were reasonable characteristics, but also seemed questionable. This sample provided very interesting results. It did not present as forged. However, it also failed to meet the typical standard when compared against other known authentic examples. So, what does this mean? In practical terms, this may either be a good forgery, or perhaps more likely...secretarial.
-
-
That is an obvious forgery based on a Sydell Spear secretarial. That took me a second. The passport is from my own study 5.24.54 IIRC. These images are a bit blurry as far as reading the text. The matching "a's" in the last example are a quick tell, the first name baseline is unusual. To blurry. Not authentic certainly. Might be a Spear but I can see it well enough. Baseline is unusual in the first name for her (the "ie" is a bit high).
-
-
Eric,
Correct. Of course the point is to verify VeriMarx ability to identify such things. "Obvious" to you...but to many collectors unfamiliar with these idiosyncrasies...not so much. The images are very clear in the actual submissions. Please see attached.
- Attachments: No photo uploads here
-
-
These appear just like the others - fuzzy, Very hard to read and when enlarged. Thank you anyway..
-
-
Here are the actual
images submitted....
-
-
-
Here is 70's Gleason. 1974 and 1977:
-
-
Correction - I meant to type the non-articulated "k: is a bit high
-
-
The 5th, 6th, 8th, 9th and 10th observations in the second bullet point "Key Findings" list are incorrect BTW. To me, its own findings are internally contradictory and factually misaligned with the historical record as the exemplars I posted show. It would be good if the thing could assign an approximate date. For one thing, taking #5, Gleason never signed like this at any point, late or otherwise. Only Pat Saddleman. There are other diagnostics to mention as well.
Sign up for our Newsletter!
Photos
© 2026 Created by Steve Cyrkin, Admin.
Powered by
Badges | Report an Issue | Privacy Policy | Terms of Service
