The Watch Site banner

If you have a Seiko from 1967 (or 1966)

1995 Views 22 Replies 12 Participants Last post by  NorCalSeiko
As some of you know, I am studying out the Seiko World time series, and currently puzzling out anomalies in the serial number data in 1966 and 1967. At this point I am looking for non-world time seiko case backs from these years, in hopes that these can give me some better context and clues to help resolve my issues below.

If you have a seiko from one of these two years and wouldn't mind sharing a photo of the case back (with at least the first 4 or 5 digits of the serial) - I would be very grateful. In this thread or via PM.

Here are the anomalies I am trying to understand:

In all other years of the world timers, the serial numbers reset each month, so you might have 430001 through 4319999 then you will again see 44##### start with a lower number and climb up. NOT in 1966 or 1967.

1. In 1966, all World Timers have seven digit serials, and so far they ALL start with 688#### or 681####. There is just enough data to make it seem odd to have these two groupings of numbers, as though perhaps the 3rd digit has another meaning, or the counter was used cross model?

The data: Data Explorer: Vintage Seiko World Time Watches

2. In 1967, the serial numbers I have recorded keep sequentially growing from Feb through June. The data set is thin, but it seems like perhaps these 7-digit serial numbers might NOT reset each month. Oddly this is true for the 6217-7010, but NOT for the re-edition of the 6217-7000. This MIGHT just be a case of too little data.

The 7010 data: Data Explorer: Vintage Seiko World Time Watches .
The 7000 data: Data Explorer: Vintage Seiko World Time Watches

3. Again in 1967 the serial numbers all change from 7 digits to 6 digits, starting in November. Is this true of other Seiko models?

The data: Data Explorer: Vintage Seiko World Time Watches

Thanks in advance to those of you who have a data point to share.
See less See more
1 - 20 of 23 Posts
  • Like
Reactions: 2
That link to the 62MAS study is awesome. I need to read through it with a clear head (and some coffee). It inspired me to make a few similar charts...

For initial reference, here's 1964 - where number appears to reset each month:
Azure Rectangle Slope Font Parallel



With the World Timers so far, 1967 looks like this (for the 6217-7010):
Rectangle Slope Plot Line Font



Like this for the 6217-7000 in 1967 (re-edition of the original)
Rectangle Slope Plot Font Parallel



And the 1966 Asian Games is all in one month, like this:

Rectangle Font Slope Parallel Plot

Attachments

See less See more
5
  • Like
Reactions: 1
By 1968 and 1969 the pattern solidly shows a numbering reset each month
Slope Rectangle Plot Parallel Font


EDIT, here's 1968 and 1969 broken out for better legibility
1968
Rectangle Slope Font Plot Parallel


1969, 6117-601X
Rectangle Slope Font Plot Parallel

Footnote - June to July do not reset in the sample data here either...
See less See more
3
  • Like
Reactions: 1
While I doing this, might as well drop each year out - then I will need to go back through it and study it out a bit.

Here's 1969 for the 6117-640X series:
Rectangle Slope Font Plot Parallel


1970
Rectangle Slope Plot Font Parallel


1971
Rectangle Slope Plot Font Parallel


1972
Rectangle Plot Slope Font Parallel


1973
Rectangle Slope Font Parallel Plot


1974
Plot Slope Rectangle Parallel Font


1975
Rectangle Slope Plot Line Font


1976
Slope Rectangle Plot Font Parallel



Data gets thinner as the years advance. Here's the overall production picture:

Slope Plot Font Rectangle Parallel
See less See more
9
  • Like
Reactions: 1
Really nice stuff! I am such a serial number nerd. How do the gaps between the numbers look? Are there discernable pattern that suggest that the gaps are not random, but a consequence of something else - ie not existing?
It's a lot to look at. If you start with 1964, it seems to me that there is a pretty obvious pattern where the production numbers reset each month, and fill out fairly evenly. There are some gaps, but I think those could just be lack of data points:

Azure Rectangle Slope Font Parallel


In 1965 there were no world timers produced.

Then in 1966, we get the Asian Games version, all from August 1966, or at least all starting with "68", in two or maybe three major number groupings. First we see the [68]14193 through at least [68]15662 group, which looks pretty well distributed in that range. Then a huge jump to [68]80268 through [68]80987. Then another sizable jump to [68]86261 from there steadily marching to [68]88140

Rectangle Font Slope Plot Parallel

These are uncommon watches, so data points are thin, but with the 22 examples here, they seem to be pretty clearly grouped. It took Seiko 10 months to manufacture around 100,000 world timers in 1964. Could they have really made 88,000 in a single month in 1966? It doesn't seem likely. Perhaps the third digit has another meaning, or maybe Seiko was sharing numbering across models this year/month, or maybe they just skipped large sequences... ?

Then 1967 happened. With the world timer series, we see a re-edition of the 6217-7000 in low production numbers, again seemly fitting the 1964 model, where the production counts reset each month they were produced (with perhaps August to September being an exception?):
Rectangle Slope Plot Font Parallel


Then we see the 6217-7010, looking more like the 1966 Asian Games, but across multiple months - these sort of production number "plateaus":
Rectangle Slope Plot Font Parallel


Data is thin on all of the watches from 1967, but of the few data points we have, it certainly seems like the production numbers are NOT reseting.

Then carry on through the rest of the data from 1968 to 1976 and the trend is again pretty clear - production numbers reseting each month.

So the odd ducks are 1966 and 1967.
See less See more
4
Perhaps of relevance, the 6217-7000's are marked JAPAN A ...

1964:
Wheel Automotive tire Drinkware Automotive lighting Motor vehicle


1967:
Automotive tire Automotive lighting Rim Audio equipment Material property


Where as the 6217-7010's appear to be marked JAPAN J:
Brown Fluid Wood Cosmetics Tints and shades
See less See more
3
I'm not going to get my act together enough to take caseback shots, but I'm hoping this will do - it is a copy/paste from my collection spreadsheet of all my '66 and '67 Seikos:

SeikoSeikomatic Weekdater6206-804061052521966
SeikoSkyliner6220-997062011551966
SeikoSkyliner6222-001063037801966
SeikoSkyliner6220-997063048491966
SeikoSeikomatic6206-804063055061966
SeikoSeikomatic-R8305-101064117621966
SeikoSeikomatic6206-810064417951966
SeikoSeikomatic6216-900067008261966
Seiko"Braille Watch"6618-800069022481966
SeikoBusiness-A8346-800069182461966
Seiko"Chronostop"5717-59706D007321966
SeikoSportsmatic6619-70506D011111966
SeikoNA21-71106N001991966
SeikoLord Marvel5740-80006N015761966
SeikoBusiness-A8346-700070015941967
SeikoBusiness6206-810071445121967
SeikoBell-matic4006-700072066421967
Seiko62MAS, Daini6217-800176006051967
SeikoBusiness-A8346-901076028801967
SeikoBusiness Bell4006-701077017751967
SeikoNA6119-703078010901967
SeikoSeikomatic-P5106-700079002081967
SeikoNA5139-701079009311967
SeikoKing Seiko4402-80007D028261967
SeikoLord Matic5606-70107D31441967
See less See more
  • Like
Reactions: 2
Thank you for the additional data, and keep it coming! Nothing is standing out to me, but each bit is helpful in trying to get a complete picture
1966 Sportsman

Watch Analog watch Photograph White Light
Motor vehicle Automotive lighting Bicycle part Gas Audio equipment
See less See more
2
  • Like
Reactions: 1
1967 Sportsmatic

Watch Analog watch Photograph White Light
Watch Analog watch Black Camera accessory Finger
See less See more
2
  • Like
Reactions: 1
Then in 1966, we get the Asian Games version, all from August 1966, or at least all starting with "68", in two or maybe three major number groupings. First we see the [68]14193 through at least [68]15662 group, which looks pretty well distributed in that range. Then a huge jump to [68]80268 through [68]80987. Then another sizable jump to [68]86261 from there steadily marching to [68]88140

View attachment 499298
These are uncommon watches, so data points are thin, but with the 22 examples here, they seem to be pretty clearly grouped. It took Seiko 10 months to manufacture around 100,000 world timers in 1964. Could they have really made 88,000 in a single month in 1966? It doesn't seem likely. Perhaps the third digit has another meaning, or maybe Seiko was sharing numbering across models this year/month, or maybe they just skipped large sequences... ?
I’m fairly certain that here the third digit is not part of the sequential numbering, but rather indicates something else - possibly a production line.

For reference, all examples of the first Grand Seiko have a “1” as the third digit in the serial number, so the serial number is of the format YM1xxxx (or MY for November and December back then), where xxxx is the counter.

If you think about it logically, the purpose of the serial number is to uniquely identify a watch of a specific reference. This has a couple of implications. Firstly, you can have identical serial numbers on different references. More interestingly, it means that you have to apply different numbering strategies depending on the planned production volume, and those production volumes can vary by quite a bit over the lifetime of a product.

If you are planning to manufacture more than 10,000 examples of a reference in a given month, you have to have a five digit counter.

But if you know you won’t be getting close to that kind of volume, then you can get away with a four digit counter. Your dataset for 1968 and 1969 is very intriguing because monthly production volumes are really close to 10k, and yet (assuming you have not dropped a digit), they only used a 4 digit counter.

Regardless of whether you have a four or five digit counter, in either instance you then have a choice as to what to do with that counter at the end of the month. You can either reset it, or you can just let it continue rolling on, since the change in the YM part of the serial number will still result in a unique number for the watches.

In the case of the first Grand Seiko, over the course of its production life (April 1960 to August 1963), we see that both strategies (letting the counter continue to roll over at the end of the month, or resetting it) are followed.

But what is even more interesting is that it is very clear by comparing the case serial data with that for the movement serial, it is not correct to assume you can deduce production numbers from the case serials. Just as one example, I have seen GS Firsts with serials 1710566 and 1715236. And there is no way that Seiko produced 5,000 examples of the
first Grand Seiko in a single month.

It’s fascinating to see that pretty much all the “anomalies” in the dataset I have for the first Grand Seiko are repeated in your analysis of the World Timers - logical resetting of the counter at the start of some months; leaving the counter running from month to month; big gaps in the serial numbers for a particular month that cannot possibly be explained away by inferred production volume; a mysterious coding of the third digit…
See less See more
  • Like
Reactions: 2
Thanks for that context. You are correct that the numbers in 1968 and 1969 are 4-digits here - and they do get very close to the max.

In this data it seems the numbering strategy changes by model perhaps instead of month/year. There are some large jumps, that may represent "gaps" of no production - but as I have collected more and more, I have seen these fill in nicely in many cases - so I am not sure what to make of them. In models where the numbering seems to reset each month and follow a pretty steady progression, I would suspect the gaps are just missing examples in my dataset - but definitely a riddle.
I was curious to see if all of these production sequences would fold together nicely without any dups by model - the thought being perhaps the casebacks were all produced, then shipped to various production facilities and randomly used up across the year/mo. It does NOT seem like that is the case.

Running a quick dup check by Production No (omitting year/month) and by model, I get these duplicates, which I guess at least proves that the same stack of numbers was not just used once per model or year, but instead repeated.

  • 6217-7000 - 1523
    • 4401523
    • 4901523
    • Gap: 5 months
  • 6217-7000 - 12163
    • 4812163
    • 4912163
    • Gap: 1 months
  • 6217-7000 - 1523
    • 4401523
    • 4901523
    • Gap: 5 months
  • 6217-7000 - 12163
    • 4812163
    • 4912163
    • Gap: 1 months
  • 6117-640X - 1368
    • 9D1368
    • 221368
    • Gap: 26 months
  • 6117-640X - 1315
    • 021315
    • 181315
    • Gap: 18 months
  • 6117-640X - 2327
    • 022327
    • 452327
    • Gap: 51 months
  • 6117-640X - 738
    • 070738
    • 4D0738
    • Gap: 53 months
  • 6117-640X - 425
    • 080425
    • 520425
    • Gap: 54 months
  • 6117-640X - 643
    • 0D0643
    • 4D0643
    • Gap: 48 months
  • 6117-640X - 765
    • 180765
    • 270765
    • Gap: 11 months
  • 6117-640X - 1315
    • 021315
    • 181315
    • Gap: 18 months
  • 6117-640X - 1217
    • 191217
    • 341217
    • Gap: 19 months
  • 6117-640X - 900
    • 1D0900
    • 230900
    • Gap: 3 months
  • 6117-640X - 322
    • 210322
    • 6N0322
    • Gap: 58 months
  • 6117-640X - 1368
    • 9D1368
    • 221368
    • Gap: 26 months
  • 6117-640X - 900
    • 1D0900
    • 230900
    • Gap: 3 months
  • 6117-640X - 862
    • 260862
    • 4D0862
    • Gap: 30 months
  • 6117-640X - 765
    • 180765
    • 270765
    • Gap: 11 months
  • 6117-640X - 1217
    • 191217
    • 341217
    • Gap: 19 months
  • 6117-640X - 2327
    • 022327
    • 452327
    • Gap: 51 months
  • 6117-640X - 643
    • 0D0643
    • 4D0643
    • Gap: 48 months
  • 6117-640X - 738
    • 070738
    • 4D0738
    • Gap: 53 months
  • 6117-640X - 862
    • 260862
    • 4D0862
    • Gap: 30 months
  • 6117-640X - 425
    • 080425
    • 520425
    • Gap: 54 months
  • 6117-640X - 322
    • 210322
    • 6N0322
    • Gap: 58 months
See less See more
  • Like
Reactions: 1
Now I've meandered a great distance from the original purpose of this thread - to riddle out the 1966 and 1967 numbers... but since we are deep in the weeds anyways...

In this dataset there were two models that had sub-model designations - 6117-6010 / 6019 in 1968, and 6117-6400 / 6409 in 1970-1972. Here's what those serial numbers look like. As the footnote says, red plots are models that end in "9", blue end in "0".







See less See more
4
  • Like
Reactions: 1
1 - 20 of 23 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top