Accuracy - the evidence.

[ratty]

Originally Posted by Rock

Being retired I have plenty of time on my hands so I do a daily 'Time check/record' on the Sub and DJ. I now have several years data.
The data is very consistent year to year.

I can state that the Subs accuracy is best in Winter (May - October) when accuracy is +1 - 1.5 secs/day and worst in Summer (November - April) when accuracy is +2 - 2.5 secs/day.

The DJ is the opposite being worst in Winter being +2 - 3.5 secs/day and best in Summer at +1 - 2 secs/day.

How about that. Anyone else done an empirical study?

Thanks for this, very interesting.

I would be surprised if the temperature difference makes much difference to accuracy bearing in mind the watch will be either being worn on will be inside a house!

At the risk of sounding really sad, have you kept any info on average temperatures during the different seasons? Do you regularly get extremes of weather?

Thanks again

[Andad]

Originally Posted by Abdullah71601

Accuracy is irrelevant to your argument. Your argument requires a watch that is synchronized to the train conductors watch. You can have the most accurate watch in the world, but if it doesn't match the train system you will still be left standing on the platform.

You first need to define the time reference before you can measure whether your timepiece is accurate in context of your argument. The variation between your reference and your watch will define how accurate your watch is. Any other measurement is useless for your needs.

The train conductors in Australia all have Seiko Astron watches.

So I sync my Rolex watches to the trains.

Or my one.....

[kromdom]

If I wake up the next day...WIN. if I remember where I last left my watch....WIN. Accuracy? meh

[Ian Macdermott]

Both my Rolexes are accurate however if I need split second accuracy then i wear one if these two.






Sent by the power of thought

[Abdullah71601]

Originally Posted by directioneng

The train conductors in Australia all have Seiko Astron watches.

So I sync my Rolex watches to the trains.

Or my one.....

This is the correct answer to the train example. How does it account for transmission delays?

I usually arrive early for trains, planes, and buses. They aren't supposed to depart before their scheduled time, but they occasionally deviate from their schedule. Rolex precision is well within the limits I need to be on time.

I really don't need to time events with great accuracy down to the second, but if I did, my time piece would be sync'd to NIST.

[mailman]

Bought a BNIB 116610LN on June 1st. I set it June 2nd. Just checked it and for 29 days it's -6 seconds. That's pretty much wearing it 10-12 hours a day

[Rock]

Interesting comments Guys.
The "standard" I use is www.timeanddate.com/time/zones/aest
The thread is not really about "accuracy" so much as "temp variation and its effect on accuracy".
Graham, the data I have collected says clearly that a temp variation does have an effect on accuracy. Where I live, the summer and winter temperatures are very different and pretty consistent year to year. Equally, the 'inside' temperature varies by as much as 10 degrees Summer to Winter. With only a few days exception, we don't use heating in Winter or AirCon in Summer.
I didn't start keeping the data with any hypothesis in mind. Just looked at it and noticed the clear consistent variation from Summer to Winter each year.
Thought it might be a talking point.

[dysondiver]

Originally Posted by padi56

Last time I checked was July 1983 just checked my watch now and its gained 1.9999987 of a second since then.But on a serious note all this checking stuff I gave up decades ago, I now check my watches around once a month or so. Then they are seldom more than a minute or two out that's accurate enough for me.Now the escapement of a mechanical watch in 24 hours pushes the gears 432,000 times and a day has 86,400 seconds.So for any mechanical watch to run within a few seconds a day is a mechanical marvel, given the fact that the movement is constantly affected by the earth's gravity, metal expansion and contraction, temperature variations, subtle changes in lubrication and friction, shocks, and so on.The fact is that no mechanical watch made will keep perfect time, very close yes but perfect no. But please guys to not to anal about a few seconds either way trust me far more important things in daily life.

all of that ,,, and listen to Baz Lurman , and wear sun screen.

[Racerdj]

While I do not have the amount of data you have achieved, sometimes the season of the year affects the accuracy slightly.

[Andad]

Originally Posted by mailman

Bought a BNIB 116610LN on June 1st. I set it June 2nd. Just checked it and for 29 days it's -6 seconds. That's pretty much wearing it 10-12 hours a day

Excellent accuracy Mr P but wouldn't you rather have it +......

'Cos the train has left the station....

[jshepard]

My study was a bit simpler. When I am called out for being late for something, I check my watch and go, "yep, I'm late! My bad..." If I am not called out for being late, I check my watch and go "alright, I am NOT late!" So far, none of my watches have failed me in these experiments expect for an instance when I forgot to wind the speedmaster as I was leaving the house because I was running late, but during that incident I already knew I was late so the speedmaster was not needed to confirm my data I gathered from the look of my boss' face.


Sent from my iPhone using Tapatalk

[The Libertine]

Originally Posted by directioneng

Excellent accuracy Mr P but wouldn't you rather have it +......

'Cos the train has left the station....

[mailman]

Originally Posted by directioneng

Excellent accuracy Mr P but wouldn't you rather have it +......

'Cos the train has left the station....

I take the bus It's always late

[ref1655]

Originally Posted by padi56

The fact is that no mechanical watch made will keep perfect time, very close yes but perfect no. But please guys to not to anal about a few seconds either way trust me far more important things in daily life.

it's a newbie mentality padi56...a natural preoccupation for those with 5-6 digit references. old-timers & owners of 4 digit models usually can't be bothered. the only time I might 'check' is when re-setting the calendar on months with less than 31 days.

[Jefferyasher]

While my post is not specific to accuracy +/- COSC time, there are a couple of ways I have/am using to measure my 14060m's accuracy. The first is the Microsoft Excel worksheet where you can add starting times and variations to determine accuracy per day/month. The newest, which I recently came across is a website synced to USNO Master Clock. It seems pretty interesting.....will update as time goes on. The website is www.toolwatch.io.