I found this little watch—a Hamilton “Bagley”—in a local shop. The shop owner’s big passion is coins, but he’s also a NAWCC member and a pocket watch collector, and some good pieces occasionally find their way into his display case.

The Bagley is one of Hamilton’s iconic wristwatches from the pre-WWII era. I’d encourage you to visit Hamilton Dan’s nice overview of the model. The invaluable Pocket Watch Database places this watch at 1941. Dan’s review says the Bagley was made from 1939 to 1942, so it seems correct.

Hamilton’s 987A movement—a 17-jewel, 6/0 size—is in my view simply superior. It is modern, attractive, and capable of unusually good performance. From what I can find, Hamilton adjusted these to the 3 common wristwatch positions at the factory: dial up (DU), dial down (DD), and pendant down (PD).

The watch was in good condition, which bodes well for adjustment. I cleaned the watch, inserted a new white alloy mainspring, oiled and demagnetized it, and let it run overnight. Let’s see what we can do with this little guy. Adjusting a watch like this to 6 positions is perhaps excessive, but it can handle it.

Our 3 Steps

I set the regulator to the center of the fast/slow index and fully wound the watch. Amplitude was excellent: over 330 in the hanging and over 310 in the lying positions. And the rates were surprisingly good.

As we explained earlier, we have three tasks in positional timing. First, we must align the dial up (DU) and DD (DD) rates. The fates have smiled upon us: DU and DD are nearly identical.

Our second goal is to align the vertical rates. These rates aren’t so bad, but the 6th position (PR) is way out there. Given the watch’s great condition, I suspect it spent some time in a capable adjuster’s hands at some point.

And our third goal is to align the horizontal and vertical rates. There’s a notable gap between DU/DD and the 3 main vertical positions, but it is better than I usually see on such an old wristwatch.

As explained in an earlier post, we’ll evaluate our progress using the average absolute deviation from zero (how far from 0, regardless of fast or slow) and the standard deviation of the 6 times (how far they are from each other). A mean near zero means the watch, on average, is close to 0 in all positions; a small standard deviation means that the rates in the 6 positions have converged and are less disparate.

Round 1

As explained in our dynamic poising series, we can identify the heavy spot on the balance by measuring the rates at a low amplitude, around 140 to 160 degrees. The watch will be fastest when the heavy spot is below the balance staff.

Here are the rates for the 8 vertical positions.

The fastest position, shown in red, is PL. The slowest position, shown in green, is opposite it (PR). (We’re viewing the watch from the movement side, so left and right are flipped.)

The heavy spot is thus below the balance when the watch is pendant left (PL), defined from the dial side. The red dot shows the heavy spot.

What should we do here? We can add weight or remove weight, but we want to do whatever makes the vertical rates closer to the horizontal ones.

Recall that Bagley is running fast DU/DD. If we remove weight at the heavy spot, it will run even faster DU/DD. The vertical positions will align with each other, but they won’t align with the horizontal ones.

But if we add weight, we’ll slow down the DU/DD positions, align the vertical ones (by increasing poise), and make the horizontal and vertical positions more similar.

Remember: if slow, remove weight; if fast, add weight.

So it’s time to break out the trusty stash of timing washers and balance screw holders. I removed the screw that was conveniently exactly opposite the heavy spot, put my lightest washer on it, and screwed it back in.

Round 2

Did our weight adjustment do some good? Let’s wind the watch fully and see.

The mean distance from 0 is basically unchanged (15 vs 14.8 seconds) but the standard deviation is much lower (5.2 vs 11.9), so the 6 times are converging closer together.

Time for another round of adjusting. Because I expected making only a small change in poise, I nudged the regulator so that the DU/DD positions were -2 seconds. By filing off a tiny amount of weight, the watch would gain in speed and the vertical positions would become yet closer to the DU/DD ones.

As before, I unwound the mainspring so the amplitude was around 140-160 degrees and measured the rates in the 8 vertical positions.

The heavy spot was somewhere else this time—below the staff in the position shown in the picture.

Time for the balance screw file, my go-to tool for fine adjustment. I lightly sanded the face of the screw directly below the staff when the watch is in this position.

And what was the result? I fully wound the watch and nudged the regulator to set it slightly fast. Here are the rates.

The mean absolute deviation across all 6 positions is now 4.5 seconds, and the standard deviation is only 2.3 seconds. Clearly, the times are converging toward zero.  The largest difference between any 2 positions is 11 seconds. The watch keeps “straight limits” of 7 seconds (i.e., no rate deviates from 0 more than +/- 7 seconds), a high accuracy standard for a 75-year-old wrist watch.

My obsessive side wants to take another crack at it and try to squeeze out a few more seconds; my rational side says “it’s late—leave well enough alone.” So we’ll close the books on this one and go to bed.

UPDATE: Trying to adjust old watches to tight limits is indeed irrational and quixotic. When our obsessive and rational sides disagree about watch-related matters, we all know how it usually ends.

So, after a few days, I decided to revisit this Bagley. After another round of adjusting, here are the timing results.

The mean is now an average of 2.5 seconds (in absolute distance from zero) per day, with a standard deviation of 1.9—our best values yet. The rates all fall within straight limits of +/- 5 seconds. The amplitudes are higher than 330 DU/DD and 310 in the vertical positions.

I hereby declare this watch the world’s most accurate little Bagley. Time to turn to a 12-size Illinois watch that needs some attention.

Conclusion

This vintage Hamilton wrist watch is 75 years old and was initially adjusted to only 3 positions. With some luck (finding one in good shape), a thorough cleaning, much patience, and a bit more luck, the watch could be adjusted to keep far better time across positions than it did when it was made back in 1941.

It’s a testament to the high quality of Hamilton’s watches and the advantages afforded by modern lubricants and timing machines.