What's With The Bank Notes?

" Easily see watermarks (including on postal cards) without fluid."
The problem is some watermarks are simply weak. Technology can not fix that.

" Easily determine paper types."
Trouble is there isn't just hard & soft, but also there is intermediate.

"Easily measure and count the grills (even the faint ones)."
With a good enough image, software should be able to do that, although that still leaves the question of whether it is real or a forgery.

Indeed, grill recognition is ideally suited to machine learning systems, the difficulty being preparing a large enough labeled training dataset. The problem is little different than human facial recognition.

One more banknote that gives me heartburn at times, the 15 cent Webster, Scott numbers 152 and 163. The so-called secret mark that is likely not a secret mark depending upon who you read or listen to. Plate wear or secret mark? In either case it is real judgement call most times.

The arc at the top of the design can help, 152 is strong, and 163 is usually weaker. that can help in unsure cases. Also the 152 paper is very thin, but 163 is usually a bit thicker. Obviously it helps to have multiple examples at hand when looking.

I agree with rog about the Websters. I also have a difficult time with ink shades ... and hence with some of the higher denominations that don't (probably) have secret marks. Hamilton & Perry specifically.

The 90 cent Perry has always been difficult for me. I once mentioned that to the dealer Richard Champagne. He told me it is simple to distinguish. Just spread out 100 or more Perry's on a table and the two shades will be easy to sort.
If only.

That Richard..... They are tough, just when you think you really know what carmine rose looks like, you discover you really don't. At least not without some comparison examples.

I posted above link a year ago but got no reply about 30c Hamilton and using position dot left side of vignette with line going horizontal from it to determine Continental vs National.

I got this from an old educational card to help philatelist determine difference as just so difficult by color.

Anyone ver heard of this

rgstamp - That was an interesting read (your 1/2021 post) and it spurred me on to do a bit of looking, but I must say I'm a bit more confused now. Checking the PF listings for 143/154/190 I see some without dots, some with dots and no lines, and others with the dot and line certed for each issuance. It is made a bit more difficult too as the dot seems to move left and right and can be hard to see on some when moved right into the shading lines of the vignette. Just from that I would say either the PF is wrong on many of their certs or it is not a reliable indicator to follow.

I have a used Scott 165 that I acquired in November of 2020 ... with 3 certs (2 older PSE, 1 relatively new June 2020 PF). It does have the "position line". Of course, a cert is merely an educated opinion rendered by an expert. Interesting, nonetheless.
Added: Decided I should not trust my aging eyes, so I scanned the stamp & enlarged it on my computer screen. The line disappeared, apparently a figment of my imagination. Will eventually do the same with my 143 as it also appears to have the line.
Added: The line on the 143 does not disappear. Sometimes the eye really does see what is there and not just what the brain tells it is there.
Thanks for an interesting philatelic tidbit.
Added: Final note. A closer look at the 143 line revealed that, when compared with your example, was slightly above where it should be & is likely just a remnant of the very light cancel. I.e., there is no "position line" on my 143. C'est la vie.

RE: 'old education card'
I am an engineer with a quality background and as such I tend to rely upon certain statical criterial for proving/disproving theories/opinions/facts.

Some times in our hobby and in this community we see folks make declarative statements about things based upon their personal observation(s). While interesting posts, I often think to myself, 'an observation is not the singular of data'. In other words, when trying to prove something we should not confuse some observations as being the same thing as actual data that would be used to prove something. As was touched on above, a large random sample size would be needed to for a theory to be verified as factual.

Samples need to be randomly collected since examining a specific collection or group may or may not represent the actual population of all stamps. For example, you score a group of 250 30 cent Hamilton bank note stamps. A big group for sure but you have no idea how they were gathered by the person(s) before you. So you cannot assume that that the 250 represented the total population of stamps printed.

The second thing is the sample size that needs to be assembled to be considered valid (at least by scientistic or QA standards). MilSpec sampling inspection criteria is often used and is one that I have some experience with. Given that there were 750,000+ of both the Continental and National printings, you would want a minimum of 1250 of each printing in a sample size. For the ABN printing (4 million+) you would want something closer to 2000 stamps as a sample size.

I think we should always be aware that our hobby has a LOT of information which is based upon casual observation (anecdotal) but is presented as factual. The take away is that we should always dig deeper and try to understand the actual data.
Don

If there were quick, easy answers to these banknote questions they would have been no trouble all these decades.

Agreed.
I was trying to make the point (and probably failed) that knowing how data is either implicitly or explicitly collected is fundamental to understanding its value AND its limitations. I think that our hobby tends to ignore this; we start with a few stamp(s) in hand and then try to work backwards to a factual conclusion about millions of bank note stamps that were printed.
Don

For difficult stamps like the Hamilton I never did arrive at a comfortable place even after many tries at different methods. I did have many hundreds in hand at once and after much frustration ended up buying some cheaper certified copies of the different printings to use as guides. No dice. I came to the conclusion that what one person easily sees another person just cannot. I believe that is a function of being human and how we perceive things visually and how our brains process information. I showed another person the pile of stamps and they went through lickety-split separating them by issue. Were the they correct? I would like to think so although I had no way of telling that because of what I stated above. It just would not register for me.

This is where technology is IMO invaluable.