Watch Accuracy - What Satisfies You?

Not according to Omega. Omega specs for the 1861 are from -1 to +11 average daily rate, measured over only 3 positions. That last part, measured over only 3 positions, is important to understand fully. The rate specifications given by companies are not what you see wearing it on your wrist, but what the company sees in short duration timing tests on a timing machine, under very controlled conditions, under known states of wind, and in very limited positions.

Even COSC watches only test for 5 positions, when the watch can be in almost an infinite number of positions in a day on your wrist.

But you specifically mention variation. The specs for the 1861 allow as much as 15 seconds variation over the three positions tested (dial up, crown down, and crown left) at full wind, and as much as 20 seconds at full wind -24 hours. So -7 to +7 is 14 seconds measured over 24 hour periods, and falls well within Omegas specs for this movement. It's important to understand that the other positions the watch can be in are not tested, and therefore have no tolerances. As long as the variation in the 3 positions is within specs, and the average rate is within spec, the watch is good as far as Omega goes.

So saying that the variation stated by Jones in LA indicates the need for service is completely false IMO.

Can these watches do better? Yes certainly, but to assume they can all be made as accurate as the best examples is simply not realistic given the design of the movement. There is only so much that adjusting can accomplish on these, so if all the stars line up and all the wheels are perfectly true, all the tolerances fall just right on the bell curve, and someone who is skilled adjusts the watch, then yes you are good. Omega knows that this isn't the case with every example, and they set the timing tolerances accordingly.

Cheers, Al

What does "much better" mean? Setting aside that straw man for a minute, you asking that has nothing to do with the assertion that the result from Jones in LA indicates that watch is in need of service - it doesn't indicate any such thing...

Cheers, Al

Firstly, thanks Tom for your input and your concern. I didn't respond earlier to your comment about the need for a repair because I didn't want to get into a protracted debate. But I know that the motive for your original comment is 100% good. Before buying a watch with a movement from the 321 family I already knew what I was getting into as far as expectations for precision go. I guess my expectations are low since I'm not concerned with a 14 sec/day swing in timekeeping (but I do know exactly how and when that swing will occur with my particular watch). If someone else with a watch based on this family of movements sees a swing of only a few seconds per seconds per day, I would consider them lucky. But I don't consider myself unlucky, based on what I've been able to gather regarding the 321 family.

I'm surprised that Al didn't re-school us in the important distinction between precision and accuracy. The original title of this thread mentions accuracy, but the discussion to this point has really been about precision (if I had a poorly-built, imprecise watch I could still keep it quite accurate by simply re-synchronizing it with a time standard on a frequent basis). The fun thing about my particular 1861-powered watch is that I can keep its accuracy within a relatively narrow band without the need to re-synchronize it, simply by controlling the off-the-wrist resting position. In other words, I'm exploiting my watch's inherent imprecision to conveniently maintain its accuracy.

Depends on the watch and vintage, and how recently serviced. For my '60s and newer Rolex sports watches, I like within 5 sec/day. Although it wasn't necessary, my '30s and '40s bubblebacks are well under 10 sec as well (my vintage Rolex guy really knows what he is about). For my older Hamiltons, etc, 20-30 sec is fine (after service), many are better. 30T2s (non RG) can be less than 10 sec as well, as Constellations from the '50s-'60s (my cal. 551 is about +2 sec last time I checked).

Hi Tom,

The answer I give would depends entirely on what you mean by COSC standards? Do you mean the commonly quoted -4 to +6 per day average rates? If so, then yes it would be very unusual for me not to get an 1861 to fall within those standards (let's leave 321's out as wear and parts availability cloud the issues with those). Or do you mean what Omega requires for tolerances for positional variation and isochronism? Or do you mean full blown COSC testing with all tolerances met over all the temperatures they test to? The last 2 definitions are certainly more problematic to assert that every watch could meet these standards.

So in fact with Jones in LA's watch he is meeting those basic "-4 to +6 standards" with the average rate being +0.2, but still you say the variation is "too much for that movement" but what do you base that judgment on? It seems you are basing it on the one example of this movement that you own, and I guess my point here is that applying what you see on one watch, across all watches, is not really a good idea in terms of having reasonable expectations. And again it certainly does not indicate the need for service.

Let's look at the last watch I serviced yesterday as an example - it's a Marathon GSAR, so very much a tool watch, with a very basic ETA 2824-2 elabore grade movement. So it does not have the better mainspring, pallet fork jewels, balance wheel and balance springs that the Top or COSC versions would have. Photos of the watch and movement:







Again basic finishing and nothing special at all. ETA allows the Delta to be as much as 20 seconds (3 positions only) with another 15 added for isochronism, and the average rate can be anywhere from 0 to 14 seconds and still be within spec., so not great performance.

When it arrived it was not running well:



Low amplitude at full wind dial down (even though I didn't set the lift angle correctly), losing 30 seconds a day, and large beat error. There were a few worn out parts I replaced, new mainspring, clean, oil, and basic adjustments only. The watch was horribly adjusted when it came in, so the balance spring was not centered at all in the regulating pins for example, and I simply performed the normal adjustments I do on every watch. Here are the results:





Now by anyone's standards, these are good numbers. Delta is just 3.2 seconds over 6 positions at full wind, and average rate is +1. Note that over 1/2 of that Delta figure comes in just the last position (crown right) that is not included in COSC calculations, so if I just use the first 5 the Delta is 1.5 seconds over 5 positions. I can tell you they don't get much better than this, even in COSC guise.

So does this mean I should expect every 2824-2 of at least this grade to perform this well? Hell no! I didn't even have to do any dynamic poising to get this, so just basic adjusting and all the stars aligned and the sum of the parts fell into place to give very good results. This watch would be at the far right of the bell curve. Assuming that because one example is this good that all should be is just not reasonable, because there will be variation across all movement types where some will be average, some better than average, and some worse than average.

Cheers, Al

To pick up on Al's point, I've long been amazed at the differences in accuracy between models with the same movements. Even beyond that, some vintage watches are very accurate even before servicing, while others aren't so hot after.

I recently sold a Girard-Perreguax HF chronometer that, after service by an excellent watchmaker, gained eight seconds over 10 days on the bench. I considered that to be satisfactory. 🙄

I just need to find time to look into this accuracy thing 😀 I'm too busy managing rotation schedule and reading posts here. I'm always on time for every meeting so I suspect my watches are somewhat accurate. 😜

Just to add a few more data points to the discussion regarding my now notorious 1861. When I deliberately let my watch rest in a particular position to speed it up or slow it down, I usually choose the most extreme positions: 12-Up is my watch's fastest position, and Dial-Up is its slowest position. Other positions either speed or slow the watch compared to a 0 sec/day loss/gain, but at a lower rate. If I keep my watch on my wrist for a full 24-hours, its average position is probably closer to Dial-Up than 12-Up which explains why it loses 2-5 sec/day in this situation. One thing that I found to be rather counter-intuitive (refer back to what I said in the previous sentence) is that if I wear my watch continuously for 24 hours and keep the chronograph running the whole time, the watch gains 1-2 sec/day. I also wind the watch each day, generally around the same time of the day so that the affect on beat rate from decreasing mainspring tension is fairly consistent day-to-day.