The below would apply to any 100% mechanical watch, and as such could be placed several places other than the Grand Seiko forum. That said, it's as applicable to Grand Seiko as any other brand and this is where I usually post, so I'm placing it here.
Positional correction (also called auto-regulation and probably a bunch of other things) seems fairly well known in Rolex circles (possibly because they officially recommended it at one point), but maybe less so in other circles.
The idea is that a mechanical watch will tend to gain or lose at different rates depending upon the position it's in. Assuming you have an accurate time source for comparison you can use this information to correct the accuracy of your watch when you take it off at the end of the day.
For example, when I took my SARW027 off and glanced at the seconds hand and compared to the GPS clock I noticed the watch was +2 seconds. If I know what the watch gains or loses in various positions, and I know the watch is going to be off-wrist for about 16 hours, I can choose to lay the watch in a position where it is about back to +/- zero when I put it back on.
If you only have one watch, then it might be easy to just remember what positions do what. But if you have a whole box of watches on your dresser, unless you have an eidetic memory you probably won't remember all of the numbers for each.
So I made up little cards for the watches I wear most, and then used the Timegrapher to test the watches in 6 positions. I tested the watches while wound to about the percentage they'd probably be wound when taken off, and tested the watches after they had been sitting at room temperature quite a while so they would be "cold" since that's the temperature they'll be when they aren't on my wrist. My SARW027 for example runs a good bit different when warm than it does when "cold" (as most mechanical watches will). I stopped the Timegrapher during position changes, and in each new position I let the watch settle until I got 4 identical readings in a row.
The resulting card (which I will stick in the bay where this watch lives in my watch box) looks like this:
So back to the scenario of taking my watch off, glancing at the GPS clock, and seeing that the watch is 2 seconds fast, if I know the watch will be "resting" for 16 hours (about 2/3 of a day) then I'd want to rest this particular watch in the 12-up position so I'd lose 2/3 of its SPD of -3 sec/day (about 2 seconds) by the time I put it on again.
Of course the problem with all this is that all sorts of things make the watch vary in terms of what it gains or loses, even just getting older the watch is going to run different. So this may all be an act of futility, but I thought it would be interesting to play with.
I've only done it for a couple of my watches that I tend to wear consecutive days in a row and have generally found that being deliberate about what position the watch rests in does indeed allow me to keep it close to +/- zero. Of course it's possible to let the watch get so far off in one direction that a single overnight period won't be enough to correct it. For example when cold my SARW027 gains in 4 of 6 positions and only loses in two positions and not very much in those two. If it got too far ahead then positional correction wouldn't be a good option and a quick hacking would be necessary to drop it back however many seconds was needed.
For those who are curious, the standard Rolex positional correction advice was:
- To gain a few seconds: Lay the watch flat with the dial uppermost.
- To lose a few seconds: Lay the watch vertically with the winding-button downwards.
- To lose rather more seconds: Lay the watch vertically with the winding-button uppermost.
I know Rolex owners vary in their opinions about the correctness of the above advice, but many also say "works for me". Personally I don't wear mine enough for it to matter as it is not my "daily driver".
Anyway, this was too long, probably stupid, or useless, or whatever, but I thought I'd share it for whatever value it might have. Even if all it does is prompt discussion of how stupid it is it will be useful.