The Known Survivor Rate of certain vintage watches. Some observations

One of my hobbies is researching watches. I have datebases with a.o. Universal Geneve, Movado polyplan & chronographs, Longines 5699 & 5347 imperiables. A very interesting question (IMHO) is always how many watches of a certain reference were made. Sometimes this number is relatively ease to obtain - at least a first glance - because for instance Longines numbered the individual watches in a certain batch.

Other brands are more challenging but there are ways to make an educated guess. Universal Geneve also watches in batches and used - very likely - consecutive numbers within this batch. Using the https://en.wikipedia.org/wiki/German_tank_problem one can then calculate an estimate of the batch size. Combining the number of known survivors and the batch number leads the Known Survival Rate© (KSR) of a certain reference. Known in this respect means ‘to be found somewhere in the public domain’. This can be the internet but also books.

My - general - theory is that the survivor rate is influenced by a couple of factors:

• Age: older is lower
• Brand: well known is higher
• Complications (including chrono): more is higher
• Intrinsic value of the case:
• Chrono’s: gold is higher
• Non-chrono’s: gold is lower
• Size: bigger is higher

Perhaps this can be summarised as: ‘the higher the value, the higher the number of (known) survivors’.

My research show me that for most UG’s the KSR is between 1 and 10%. Of course there is always the unknown unknown: references which were produced but of which no survivors are known.

Most brand never published the number of watches made of a certain reference but a new book published in collaboration with Rolex includes these numbers. Mike Wood already posted some of them on insta.



I spent some time looking for the examples of the smaller batches so ref. 6200, 6204 & 6205. There will absolutely be more examples known than the number I found but on the other end these (very) high watches have a tendency to be traded in very public places (but not always serials numbers are published). The aforementioned references were unfortunately not produced with consequetive serial numbers so the tank formula would have been of little help.

Ref 6200: production 303. Examples found: 22. KSR: 7.3%

Ref 6204: production 2,881. Examples found: 26. KSR: 0.9%

Ref 6205: production 810. Examples found: 43. KSR: 5.2%

Very interesting to see that even for the very top end of the market the % is still in line with what I found earlier!

Where exactly on the internet are you looking? It seems difficult to scientifically scour the internet for survivors. Additionally, finding a watch pictured in a book or magazine does not mean it survives still.

But I would expect your summarization to be wholly accurate. The higher the value, possibly rarity, the more likely it is in a box/safe, and not on a wrist, therefore more survivors.

I am new to this but I would imagine it is generally as simple as that. Biggest factor price, second rarity. The main difficulty in determining the number survivors though, is weighting the collectability. More collectible watches will surface less often, and less likely on the internet.

I find this really interesting and agree with a lot of what you’ve posted.

Two points spring to mind:
1. The German tank problem methodology would only be useful for those models with a chronologically discrete serial range. This would exclude watches like Speedmasters where the only usable serial is on the movement and the range is shared with other non Speedmaster chronographs at the time.

2. Although known survival rate is interesting, I’d be much more interested in unknown survival rate (how many are still out there waiting to be found).
It would take a decent amount of time to wait for the info, especially for rarer models, but the mark-recapture method could potentially be used to estimate the total population of survivors.
There would likely have to be some kind of adjustment for the fact that movement and recapture is not random and is biased towards watches that are already known/found but I’m sure it could be done.

In case of UG and Movado: let's say everywhere 😁 On the Rolexes: not that deep.

Of course you can't be sure about watches in a book not surviving but in general I assume that 99.9% will do. Bit of a philosophical discussion IMHO
I think I stated that but perhaps I was not clear enough.
Simply put: the number of steel watches which will be melted will be very limited. However a large number of them will live in places where they will never be found i.e. landfills or on the bottom of the ocean. Hundreds of milsubs were dumped - in oil barrels filled with concrete - by the MOD in the Irish Sea. Golden watches is a different story. At least once a month a loose UG movement is listed on the bay which used to live in a tri compax yet you never see a loose case for sale.

I wouldn’t class watches in landfill or at the bottom of the ocean as survivors, they’re gone for all intents and purposes and even if retrieved, almost certainly in terrible condition.

By unknown survivors, I’m mean those watches which are in useable condition, largely original are unknown to the market as they don’t appear in the various spreadsheets let by collectors nor are there any sales records on the internet.

I’m sure I’ve read somewhere that movements where made in batches and thus a given serial range would only contain cal 321 or 552 etc.

That said, unless we know the boundaries of these ranges (which I’m sure we don’t) it’s of no use

Some more numbers were published:
Ref 5510: production 500. Examples found: 31. KSR: 6.2%

It is funny to see that the numbers are kind of what I expected based on other brands.

For instance for the Longines sommatore (5699/23086) the KSR is 11%. Movado Polyplan is around 4% based on both serial and movement numbers.

Someone added: radium. This obviously leads to a lower KSR

Very interesting. My rule of thumb for UG chrono's is a Known Survivor Rate of about 3% (based on a very large database) and for Movado chrono's about 4% (based on a - more or less 'guaranteed' - production of 20k pieces). I assume that Speedy's should have a similar number