Thursday, March 11, 2010

Off on a tangent ... a testing time

For a change from my usual bike blogging, I'm going to bang on a bit about one of my other interests / passions, watches, and more specifically, accuracy of watches.

I have a number, which I'm not disclosing, of watches ranging from a Timex Ironman (a plastic cased, quartz watch I wear when cycling - lots & lots of useful functions), through an assortment of Seiko chronographs from the 70's, a few Omegas, a Longines, a Rolex, a Graham, a Maurice Lacroix, and a few others. I'm not going to declare the total number, and i'm not going to list the entire collection here. Hey, someone from might be reading this.

Inspired by a recent post by The Sydney Tarts, I asked myself "which is the most accurate?", along with "how accurate is accurate?" and "what level of inaccuracy am I prepared to put up with?"

Let's start with "how accurate is accurate?". To be a COSC (Contrôle Officiel Suisse des Chronomètre Certified Chronometer (watch latin for "an expensive watch that should keep time") a mechanical watch must be accurate, over a period of days, in a variety of positions, to plus or minus 3 seconds.

24 hours in a day, 60 minutes in an hour, 60 seconds in a minute equals 86,400 seconds in a day, so we're aiming for 0.003% (or better).

What am I prepared to put up with? Well, that's a good question. I tend to change my watch on at least a daily basis (yes, sometimes more than once a day), and any given watch is typically in 'rotation' for about a week at a time. By which I mean, I've got three or four that I'm wearing "this week" and each will be worn three or four times in a fortnight then put away. Some (favourites or new watches) last longer in rotation, some head back to the cupboard a little quicker (I'm fickle).

While a watch is in rotation, if it's not being worn, and an automatic (most are) I'll manually give it a few winds or shakes - to keep it going - at least daily. I had a winder but it broke, can't be fixed, and I've not got around to replacing it.

I set my watches to the speaking clock, and check them against the time pips on ABC radio.

After that LONG WINDED preamble, I'd say that plus or minus 10 seconds against the time pips is accurate enough. Any less, I'm happy; any more, I'll mutter under my breath about inaccurate watches and then reset it.

Which is the most accurate? I've devised a round robin elimination process to try to work out which is the most accurate of my time keepers:

  1. Set to the speaking clock on 5 minute intervals, i.e.: it's easier to set on a 'full' 5 minute mark than it is to set to say 8:57 pm.
  2. No more than three in rotation at any one time (quartz excluded).
  3. Not on the wrist for less than 4 hours, or more than a day at a time.
  4. When not being worn, must be wound (manual watches) or wound/shaken (gently please) once a day
  5. At the end of seven days (168 hours, 10,080 minutes or 604,800 seconds - give or take), whoever is closest to the pin, wins.
  6. Repeat for another week with three more, and for a third week with the last three.
  7. Don't bother with pieces like that Seiko I know keeps crap time.
  8. In the event of a tie, I'll come up with some sort of arbitrary way of assessing, like "which looks best with a pair of jeans?"
Having set the rules (and I'd welcome any input), heading to the starting gates for the first week are ...... The big guns of my collection: Early 90's Rolex GMT Master II (movement? anyone?) 2009 Graham Chronofighter GMT Big Date Oversize (inhouse - I think - movement: G1733) and my new Longines Master triple date moon chrono (inhouse movement L678).

And of course I'll be twittering observations and updates, using the hashtag #watchtest

Quick update - thanks to the encyclopaedic knowledge stored in @twinck's head, I can advise the Rolex has a 3185 movement, the Graham is running a valjoux 7750, my longines is keeping time with the mightly valjoux 7753.


  1. Ok, I think you've got the basis for a fun testing regimen there.

    One other thing to consider would be purchase of a watch timing machine which can give you some pretty good time keeping information about each watch in a short space of time.

    A couple of good overviews are these:


    Of course these machines are expensive. An alternative to purchase that I have used is to take some of my watches to a preferred watchmaker who has a timing machine and see if he or she can spare the time to do it for you. It's most fun to have it done in front of you, and he or she can give a commentary and explanation.

    Like you, I own a number of watches. We've already established that my standards of date keeping are lax compared to yours i.e. I usually don't set the date on the watch I wear on any particular day to save stem and tube wear.

    My current personal attitude to accuracy is that I'm not too fussed. I'm never disappointed in the accuracy of my watches I never wear the same one 2 days in a row unless travelling and I deliberately don't check how the watch has been going 24-hours after setting the time. That way I don't get upset if it's lost 30 seconds, or even a couple of minutes.

    New morning, a different watch.

    Mind you, I have to make sure I'm always 5 minutes early for any appointment so if the watch is running slow there's no embarrassment.

    But that's just me.

  2. I have worn the same watch for more than a year, more or less (because a) it's easier to read with my failing eyesight than the dozen precious -to me- mechanical watches I don't wear and b) it was $70 -I could care less if I break it). While it keeps excellent time (bloody quartz) I can't say that timing accuracy has ever been all that important to me.

    Mind you I drive an automatic, use a Macintosh and have never bought a Chronograph because they have too many buttons.

    I've always looked at it like this. A hand made watch is by it's very nature inherently flawed: I know it needs to be serviced regularly because stuff wears out or comes loose. So it follows that it will be out, to some degree, at any given time (except two moments each day).

    I have a Rolex I haven't worn more than twice in 5 years, not because it slips a few seconds over a week, but because it has a lousy clasp that I'm too mean to replace (yes I've tried to have it repaired - it's a "known issue" with these sir). But the watch still gives me pleasure when I see it, because it is a real piece of watchmaking craft, external aesthetics aside.

    And since your test is about mechanical movements, I will be keeping watch. (wrong time to pun?)