Written by Ctein
[Important: Please read the whole post before clicking the payment tab. A payment by itself does not constitute an order! —Ed.]
As Mike wrote last week, the best suggestion I got for what to do with my dye transfer matrices was, "Sell them!" At a reasonably low price, enough people seemed to be interested in having some printmaker Ctein work product, not to mention a piece of now-nearly-gone photographic history.
Ask and you shall receive! I have 280 matrix-and-work-print sets for sale, most of them one-of-a-kinds, although I have multiple sets for a few photographs. For the next five days (until midnight, Monday, Central Standard Time) or until I run out, you can buy these sets for a nominal price that won't be repeated.
Each set includes a work print (usually the first print pulled from the matrices), the three matrices (seen above) that were used to make the print in their protective paper folders, and a two-page letter from me that explains what the annotations on the work print mean. The work prints are exactly the same size as finished prints—90% of the sets are for prints in the 16x20" size range; most of the rest are 10x12" prints.
The price? $150 per set, plus a flat charge of $25 for shipping and handling, regardless of the number of sets you order (so long as they're all going to the same destination). International customers pay no more than domestic. I have to collect sales tax from California customers, so orders shipped to a California destination will cost $162 a set, tax included, plus the $25 for S&H (20% of all sales goes to Mike, so you're also supporting TOP when you buy a set).
The twist
There is a twist to this sale.
Originally, my idea was simple—people would send me money, and I'd pull
a work set from my files and send it to them. It would be entirely
random. Then I decided to let you tell me which print you'd prefer to
get and see how many of those requests I could satisfy. The odds are
against being able to make everyone happy, but it will be a fun game for
me to see how many requests I can match (I have an odd idea of "fun").
Here's the way it works. For each set you order, you can give me up to three picture titles of sets that you would prefer to get (you don't have to give me any choices if you're entirely happy with the luck of the draw). What I'll try to do is optimize all those preferences across all the orders to maximize the number of people who get one of their three choices.
Please don't try to game the system; it won't work. Making fewer than three choices won't improve your chances of getting the one you want. Meta-choices like "anything from Hawaii is fine" I will have to ignore, and will constitute no choice at all. You can't specify size.
This isn't first come first served. I'm going to assign prints to people based on what will maximize the number of happy people. Your odds of getting the picture(s) you want are just as good on the final day of the sale as the first (assuming I'm not sold out by then).
Beyond that, there are no hidden rules and no catches. You make three picks; I'll make you happy if I can. If people's choices were entirely randomly distributed (they won't be), 95% would get one of their choices. But....
There are no guarantees in this game, and, of necessity, no returns. I'll try to fill your request...but you have to be willing to take whatever I send you. You're not guaranteed to get one of your choices. It's still a grab-bag. If that doesn't sit right with you, then please don't buy.
The work print quality varies. A small percentage of them are so pristine and on the mark that you could hang them on the wall; a small percentage of them are so far off that it will amaze you how I could get from there to a finished print. Some are flawless, some badly cracked. Most are somewhere in between the extremes. Which will you get? No guarantees. Luck of the draw.
Ready to play? Then here's what you have to do. Please follow these steps in order! Don't jump ahead; it will just confuse things for me.
1. Make your choices. All the available photographs are shown on this page.
You can make up to three choices per order (e.g., if you're ordering two sets of matrices, you can make six choices, which would improve your chances of getting at least one of them).
It should contain the following information, in this order:
• How many sets you are ordering.
• The titles of your choices, if any.
• Whether you are paying by personal check or PayPal and the amount you are paying.
• Your postal mailing address, exactly as it should appear on the mailing label. I copy this into a database, so plain text, please. No HTML, no fancy formatting, written on three lines (foreign customers will have a fourth line for country, of course), thusly:
Ctein
42 Skyline Drive
Daly City CA 94015
This e-mail constitutes your order. This is what I use to record and track your purchase. If I do not get an e-mail with all this information, you don't have an order! If you choose to pay by PayPal, do NOT put this information in your PayPal order and expect me to see it. It needs to be put in an e-mail to me.
3. Pay me. You can pay by personal check drawn on a U.S. bank in U.S. funds, which will save me the PayPal fees, made out to "Ctein" and mailed to the above address. Or you can pay via PayPal with the buttons below. There's one for California buyers and one for everyone else.
[Ordering information removed—Sale ended midnight 2/17]
The PayPal price is the price for one set including S&H, because I think that is what the vast majority of people will be buying. If you want to order more than one set, you need to click through the order button once for each set. You'll get charged each time for the shipping and handling, but I'll catch that when I review the order and refund the difference to you, so you'll only pay the flat $25.
Orders will start shipping about two weeks after the sale ends. All orders will ship within 60 days. I will send you back a confirmation e-mail when I get your order, and another one when your order ships. If you do not get your confirmation e-mail within a couple of days of sending me your order e-mail, contact me. If you don't get a shipping e-mail within 60 days, contact me then.
Any questions? Please post'em and I'll reply!
Ctein
Ctein's column appears on Wednesday mornings on TOP.
©2014 by Ctein, all rights reserved
(To see all the comments, click on the "Comments" link below.)
Featured Comments from:
ross: "When in the process should we submit payment? I don't want to mail off a check or send PayPal to later find you have sold out."
Ctein replies: I will be tracking the number of sets sold, and the order links will get killed if and when the sets sell out. You should submit your payment right after sending me your email. Don't wait for my response.
Ken Sky: "Are the prints signed?"
Mike replies: Ctein is out for the afternoon but
I'll step in—the prints aren't signed, and they shouldn't be. When a
photographer signs a print it in effect says "this is finished work and
meets my approval." A signature proclaims that the work fully expresses
the artist's intentions. A first work-, test-, or guide-print would by
definition not meet this standard.
Kevin Purcell: "I sense a Zipf-like (or similar Pareto-like) distribution in demand going for 'space-related,' 'lit up stuff at night' and 'anything with strong blue sky.' Then 'geology.' Then the rest. I'd be interested to see Ctein summarize the demand after the sale."
Ctein replies: Well, with approximately half the sets sold, you're right about a power distribution, although the exponent is somewhat less than one. The range of choices, though, is considerably more eclectic than I would've guessed. Some are no-brainers to me, but there are a few real surprises, including one photograph that gets no traction in my portfolio but has four requests already for the matrix set. What this says is that TOP readers have broad and eclectic tastes.
(Before anyone asks, no I'm not going to say what particular photographs are being chosen heavily and which aren't.)
That will come as no surprise.
In terms of my workload, a power distribution is good and what I was hoping for. So far something like 35% of the photographs people have chosen were picked by only one person. That makes those easy—it's a no-brainer that if someone requested a photograph that no one else did, then of course they get it! I don't know if a global optimization on this kind of problem is an NP-complete problem (one of the mathematicians here, like DD-B, could say), but the local optimization is definitely in the N-cubed range. So, if this 35% holds through the rest of the sale (no way of knowing) it'll reduce the difficulty of my optimization by more than two-thirds.
I had no idea what to really expect. The pathological extremes made it very unpredictable. For example, if everyone had chosen only prints that were offered in TOP sales, it would be really easy: A double handful of people would get their choices and then everyone else would get random pics. Conversely, if it was a perfect distribution, with each and every photograph getting three "votes," that's the worst of all possible worlds in terms of the work required of me.
The next time someone proclaims that it's only the pictures that matter, we can use the success (i am assuming) of this sale as a great counterpoint. Process matters!
Posted by: saagar | Thursday, 13 February 2014 at 08:58 AM
There certainly are matching-related optimization problems that are NP-complete. Usually it involves adding some kind of weighting or cost to the matches and then requiring that you minimize this cost.
The matching of medical residents to residencies is one example of this if I recall correctly.
Posted by: psu | Thursday, 13 February 2014 at 06:14 PM
The images on ctein's website aren't appearing. Is there anywhere else I could see the images to choose from?
Posted by: Nick | Friday, 14 February 2014 at 02:03 AM
The pathological extremes
Which TOP readers did you have in mind ?
:)
Posted by: Nigel | Friday, 14 February 2014 at 02:46 AM
I find the optimization aspect of this whole thing fascinating. (Of course, I would.) It reminds me very much of a simpler version the inheritance optimization procedure from Neal Stephenson's "Cryptonomicon." Given Ctein's known genre affections, I have some small suspicion that the resemblance isn't entirely coincidental.
Posted by: Nicholas Condon | Friday, 14 February 2014 at 09:40 AM
I'm guessing Jim Marshall's John Coltrane images aren't part of this deal...
Posted by: k4kafka | Friday, 14 February 2014 at 10:05 AM
The problem you're trying to solve is not NP-complete; it's called the "Stable Marriage Problem": http://en.wikipedia.org/wiki/Stable_marriage_problem
Posted by: Bob Blakley | Friday, 14 February 2014 at 12:00 PM
Dear Nick,
Do the usual: clear your browser cache and reload this TOP page. The link should direct you to http://ctein.com/topmatrix.htm and you should see the images of the photographs just fine.
If your difficulties persist, this is beyond my expertise to resolve; all I can tell you then is that it's a problem with your system, not with the webpage.
~~~~
Dear Nigel,
Since all readers are above average at Lake Waukesha, the pathological individuals are probably the normal ones.
~~~~
Dear Nicholas,
I have not read Cryptonomicon, so any resemblance is entirely coincidental.
~~~~
Dear k...,
I don't own the rights to any photographs that I've printed for other people. I have lots of matrix sets for photographs I've printed for other photographers, but those are not mine to sell, because I don't have authorization to do so. Those photographers get to establish their own markets for their work.
I would have no interest in even making such an inquiry. By the time I got done splitting the money between them, me, and Mike, it wouldn't be worth my time or effort.
pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com
-- Digital Restorations http://photo-repair.com
======================================
Posted by: ctein | Friday, 14 February 2014 at 01:02 PM
I don't think this is the stable marriage (as the preferences only go in one direction -- the prints don't care about the buyer but the buyer cares about the prints) or the hospital/residents problem (you don't care about the residents specialty i.e. the buyers preference. IRL those end up in the "Scramble") but rather a simpler assignment problem as N people have to get N items.
http://en.wikipedia.org/wiki/Assignment_problem
As the weights are all binary on three things you might apply the Hugarian Algorithm which is polynomial (P not NP :-) ).
http://en.wikipedia.org/wiki/Hungarian_algorithm
Posted by: Kevin Purcell | Friday, 14 February 2014 at 06:58 PM
Ctein,
What are your plans for the unsold matrices?
Posted by: John Yuda | Monday, 17 February 2014 at 05:00 PM
"Some are no-brainers to me, but there are a few real surprises, including one photograph that gets no traction in my portfolio but has four requests already for the matrix set. What this says is that TOP readers have broad and eclectic tastes."
After sorting through your portfolio, the photos I selected are probably not my overall favorite photographs in your portfolio. Instead, I selected photographs which I thought might be most interesting when presented with the matrices, each as a piece of art itself (though I admit, my limited knowledge of the details of dye transfer may negate these efforts). Had I been shopping for prints, only 1 of my selections would have been in my top choices.
Posted by: Scotto | Monday, 17 February 2014 at 08:48 PM
Dear John,
About half the sets are unsold, and those I'll offer on my website. The selection page is up already:
https://ctein.com/matrixsets.htm
but I haven't coded in the explanatory or ordering information.
Customers will be able to pick and choose among those as they like, but they'll be paying a considerably higher price (probably $275 plus shipping), in keeping with the philosophy of sales on Mike's site. If educational institutions or museums contact me about acquiring sets, I'll cut them a much better deal. Or individuals buying in bulk––prior to last week's sale, one conservator bought 30 sets to use in his classes.
~~~~
Dear Scotto,
People will take more “risks” when they're looking at a price in three figures instead of four, so I expected a considerably broader range of selections than what people usually buy. But still, “Smiley Seal,” which isn't remotely typical of my work and never gets any interest in my portfolio, got five requests. That's way, way above average; in fact, only 15 sets got more requests than that.
pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com
-- Digital Restorations http://photo-repair.com
======================================
Posted by: ctein | Tuesday, 18 February 2014 at 07:04 PM