Broken Bank Notes Message Board

I just exchanged some e-mails with the editor of Paper Money magazine. He indicated that there are no hard and fast submission guideline apart from those printed on the first page of the magazine. No size limits either. As a result, I'm considering a series of three articles that present the information in the three main chunks we are tracking today (SENC, Proof, and Unlisted).

Bernie, I think you and I ought to collaborate on the content and co-author this article. I'll send you a draft once I have the data transcribed... unless you tell me not to.

- Greg

I finally finished the filtering of the data. Here's the final tally:

SENC Uniques
1301 potentials
10 of which a proof has been seen on Heritage that may be the one sold by Christie's
8 other notes seen on Heritage that were either released version or declared not from the ABNC sale
Final total of 1283 likely uniques listed as SENC

Proof Uniques
668 potentials
10 of which a proof has been seen on Heritage that may be the one sold by Christie's
6 other notes seen on Heritage that were either released version or declared not from the ABNC sale
Final total of 653 notes listed by Haxby as Proof

Unlisted uniques
127 potentials
1 of which was seen on Heritage that may be the one sold by Christie's
Final total of 127 Unlisted notes

Grand total for all groups is 2061 unseen potential uniques, which is almost exactly 10% of the 20566 notes we recorded from the Christie's sale.

By the way, SC and TN presented an interesting situation. A number of the potential uniques from those states were what I would call design proofs. Instead of being directly tied to a Haxby number, they were proofs of the basic designs used by the branches of a bank. Thus, though the designs were called out by Haxby, they weren't really assigned a Haxby number. So I treated them as Unlisted. They are almost certainly unique, with a higher degree of certainty than the notes that had NO Haxby reference information. I intend to call them out more clearly in the article.

- Greg

Could you post or send me the newest Unique spreadsheet?

I am still somewhat confused by the UNL part of the spreadsheet. I thought that you had removed the UNL notes from this spreadsheet? I certainly disregarded all of the UNL proofs in the NY section. See your posting of 5/29/08.

_________________
Bernie

Here is the latest Uniques file.

I don't totally disregard the Unlisted uniques. I'm simply not tracking ALL of the unlisted uniques... just those that interest me for some reason. It's hard to understand, I know, but here's the basic drill:

If there was a grouping of notes on a sheet and one or more of the notes on that sheet was unlisted and the others were SENC, I assume the unlisted note was probably also SENC... so I track it.

If there was a grouping of notes that had another reference number but no Haxby number, I'd track that until I could look up the alternate reference source... which is how I got started on the cross reference work.

If something about the description caught my attention as being a distinguishing characteristic, I'd flag it... just in case I found some way to check it out.

If all I had was an imprint and a denomination, I generally did not flag it. Too many of the alternate references are sloppy about tracking imprints, and they may have that denomination listed but I'd have no way to verify the note in question was indeed the one described by the alternate reference source. In short, if I had no way to verify a note, I let it slide.

MN presented an interesting twist. A couple of the notes in that group were "pre-proof". If you look through the Hewitt book you'll see "notes" that were really just rough guidelines for what the design would eventually be... vignette placement instructions, bank name, denomination... those are almost certainly unique (albeit not impressions of an engraved plate). So I thought I'd track those as unlisted uniques, just for the heck of it.

"Unlisted Uniques" is a sort of dumping ground, to be sure, but I expect that dumping ground has some real uniques in it.

- Greg

Take a look at:
http://cgi.ebay.com/Cleveland-OH-10-For ... dZViewItem

Christie's OH-165-16a comes in the following varieties:
red OP; stamped - 5 copies
brownish red; stamped - 2 copies
brownish-orange; mounted on card - 1 copy

This one looks like it might be the last one which is also a unique impression. It has a starting bid of $1495

_________________
Bernie

Greg, take a look at 5 unique impressions posted by 08822 at the CU forum.
My response to him is a few posts down.
http://forums.collectors.com/messageview.cfm?catid=23&threadid=674821&STARTPAGE=2

_________________
Bernie

Tonight I did some analysis using the All_Lots file. That file currently has over 7000 auction lots described. Of those 1472 have been determined or declared to describe notes that were traded before the 1990 Christie's auction. I checked these against the Uniques5 file and only crossed off two of the unique notes for sure, and have another 3 that bear deeper investigation.

Interestingly, during my analysis I found one or two notes that claimed to be from previous sales and were also said to be in the Christie's sale. I'd be interested to know whether Christie's accepted consignments for that sale, or whether all the notes offered there came directly from the archive. The wording in the catalogue suggests that they all came from the archive, but the wording is not ironclad in that regard. In either case the number of notes that were asserted to come across that bridge was vanishingly small.

I'm also seeing that only one note declared as a Haxby Plate Note was being asserted as also showing in the Christie's sale.

In any case, the effect on the Uniques numbers is (so far) tiny. This gives me confidence to go forward with the Unique Impressions article with minimal corrections.

- Greg

I've done some additional processing (and thinking) on the All_Lots data, and I expect that there are quite a few more notes in there that were not in the ABN archive sale that I missed on the preliminary pass. I'm nearly done tagging the records with obvious indications about their availability in the ABN sale, and will soon do a pass to compare the records that are not definitively tagged with the ABN sales catalog data.

The intent in all this is to insure that we have some type of placeholder in the census that indicates a note was or was not seen in a public sale. Where possible, I will also note when additional examples of the same note are seen.

One thing is becoming obvious during this processing phase. I need some way to indicate which example of a note is seen in any given auction to avopid counting it in the census multiple times.

Greg,
Take a look at lot 7328 in the March 24, 2009 Stacks Auction.

This is described as: "... OH-470 Unlisted series. Wolka 2495-11. High Rarity 7, likely Unique. This is also the only example seen by us (or anyone we believe) from the lone sheet in the 1990 Sale. This was an interesting sheet with all different notes. The bottom $5 impression was payable at Philadelphia and appeared in an EAHA Sale last year...."

This note was not captured in your "Unique impressions" spreadsheet as "Unique in C but UNL in H" because Christie's sheet had two $5 proofs. However the two notes were different!

_________________
Bernie

Thanks. I'm almost to that point in the processing the 03/23 auction. I'll take a look today.

One thing... the reason I didn't publish the Unique Impressions article is because I suspected the data set had errors like this, which is why I am actively processing all the auction information I can get my hands on. At present my All_Lots data set is approaching 9000 entries, and will probably crest that milestone before I'm done with the catalogues I have now.

To be sure, I have not processed the data enough to know how many other errors or omissions in the Unique Impression set there might be. And I'm glad to know that you uncovered two additional items for the Uniques set with this discovery. I'll update the Uniques file as soon as I understand the error better (i.e. which two impressions were combined in the one Christies lot).

- Greg

Now that I've hit a lull in the database expansion phase of the project, I think it's time to revisit the Unique Impressions project and determine how many of the previously determined "Unique Impressions" are no longer considered unique.

- Greg

I presume you will be using the new data base of all lots sold at auction for some of this?
How complete do you feel that this data base is?
I am not sure how the data base (without images) is going to eliminate many proofs from the unique spreadsheet?
How are you going to proceed?

I have kept track of "notes that prove non-uniqueness." I posted some of this information previously. Remember that these are not just proofs but also issued (or remainder) SENC and UNL Haxby notes.

If I can help in some way, tell me.

_________________
Bernie

Interesting. I wasn't aware that you were doing a more complete "non-unique" study. Maybe you could take a look at the list I develop and eliminate the notes from that list that are determined to be non-unique based on non-proof items.

One small wrinkle... if a note changed between its proof state and its printed state, I'd still declare the proof to be unique.

- Greg

So what does "changed" mean?
I think the criteria should be that it would be a different Haxby number.

_________________
Bernie

I would agree, if the Haxby variants had a different number. Even in the latest Stacks auction there were several notes listed with "Variant" at the end of the Haxby number. To my way of thinking it is or isn't a Haxby IDed note. In other wards, a note attributed to a Haxby type with a variation (OP, no OP, Date, etc) would not be the same as the issued note of that Haxby type.

Make sense?