[Sjors]

It took some time to complete and juggle a bit with the data, but I think I finished the test. Since I cannot make up a post like an html page, I posted it on my server.

You'll find all about it here.

I didn't use computer programs, but a crappy calculator, a pen and paper (I should take a pic of my note book to complete this page), so I had to cut intermediate results into whle numbers to keep it easy calculatable. The σ shouldn't differ very much from a computer calculated σ.

Cheers, I'm off to bed.

Sjors

 

[solar g-shocker]

Sjors,

Is it fair to say that the master's of G, like AW-57X models, GW-200 Froggie's, and Raysman were best overall, even with a small "n"? The new Muddy's and the 3 DW-004's were good also!

I was just wondering if this had finished, so thanks for sharing this result! |>

Neil

PS
I might be looking for a reason to spread some Raysman love though...

 

[DW-5600E]

Ok, I am in a particularly silly mood...

1. At the bottom of your list of results, you state:

*A average month is 265/12 = 30.4 days.

2. The number of seconds in a month would therefore be:

60*60*24*30.4 = 2626560, not 2635200 which you've printed on the last line of the page.

You've somehow ended up with an extra 2.4 hours for the month!

Sleep well, sjors! :-!

 

[FreeDiver]

thx Sjors for all that work, my additional concluding note:

all those watches are worthwhile to own because they look cool and because even +-30/month is acceptable in my book.

cheers

 

[az12vman]

Now that was totally awesome, dude! WOW!

Thanks for the fun data!!!!!!!! I was wondering what that whole "Horizontal Bop"* thing was that you have been mentioning in your posts...






*Sorry- Just saw Bob Seeger and the Silver Bullet Band (Hometown Legend here in Michigan) night before last. |>

 

[dingosrun]

dam sjors you got way to much time on your hands lol

 

[kpatt]

That's brilliant Sjors ! It's interesting how much variation there is between models of the same module although things don't look too good for the DW-5600E or the DW-6900.

Next test could be to run the stopwatches and timers against each other.;-)

cheers,

 

[Gordon]

Cheers for that Sjors!

Much appreciated! :-!

 

[SeikoAutomatic_01]

Interesting. Some digital watches are not that accurate. Now I know that a Wave Ceptor is good for more than the time change twice a year.

 

[logician]

Ok, I am in a particularly silly mood...

1. At the bottom of your list of results, you state:

*A average month is 265/12 = 30.4 days.

2. The number of seconds in a month would therefore be:

60*60*24*30.4 = 2626560, not 2635200 which you've printed on the last line of the page.

You've somehow ended up with an extra 2.4 hours for the month!

Sleep well, sjors! :-!

Actually, Sjors was right, a generic "month" is considered 30 days for most conventional uses, though it would come up to 2592000 seconds, not either of the figures above.

For this use, Sjors started on Nov 23, and ended on Dec 22, which would be a grand total of 29 days, assuming he started and ended his test at the exact same time of day for each, which would make the figure he should use 86400 seconds short of the 2592000, or 2505600 seconds. In reality though, a single day's worth margin of error is only 3 1/3%, and I would imagine +/- 5% would be adequate for comparison - it wouldn't affect the test at all, though, since all watches would be subject to the same innacuracy, lending it to be a self-correcting "error"...

 

[UBU]

Hi Sjors,

well done, a very good test :-!

I use my atomic GW 1200 as a reference for all my watches and it's amazing

that some watches keep time for several weeks in a range

of + - a few seconds!

(sorry not a G - it's the old 7A48 Seiko Quartz movement that wins all the competitions)

Best regards and Mery X-Mas to all santasmile

Uwe

 

[Sjors]

Hi Logician and DW-5600,

Thanks for your comments.The 2635200 was a 30.5 day month. I corrected it on my page. The answers stayed the same. I actually started around 10:00 am on the 23rd and read the watches in the evening of december the 22nd. This compensates for the 0.4, while of course this will probably not result much in the outcome. Actually, if I had ead the watches next day, I guess that most watches woud have been off so little, that it falls into the error increment of 1 second. Actualy this would theoreticly only be noticable with watches that are off more than 30 seconds.

For the daycounting, November 23rd also counts as a day, which makes 30 days (see next post).

December 22nd the 30th day. Actualy I have puzzelled this out the 22nd by just using my fingers, because I first though I made a mistake too by counting November 31 days.

Uwe,

Nice to see you around here, The best Chrismas and New Year wishes for you too!

Everyone,

Thanks for your support and comments. Have a great time!

Cheers,

Sjsantasmilers

 

[Sjors]

[column]23
3
13[/column][column]24
4
14[/column][column]25
5
15[/column][column]26
6
16[/column][column]27
7
17[/column][column]28
8
18[/column][column]29
9
19[/column][column]30
10
20[/column][column]1
11
21[/column][column]2
12
22[/column]

 

[Resinbandit]

Hi Sjors,
A mammoth feat;-) . An interesting read for me. I not one for this kind of accuracy (always running late) in general life. It's just always interesting to read of the accuracy you can get for cheap today.

Kind of like the fact that man landed on the moon (or...did they...:-x ) with less computational power than an old "286".

Good work Sjors, and thanks for sharing.

 

[Agustín]

Thanks for this evaluation Sjors. You have a great collection of G. So you maybe buy baterys by bulk :-d. This make me question. The age of a batery in a watch can make a diference in the performance even with the same module? From watch to watch? I hope this was clear :-s
Greetings santasmile:-!

 

[dingosrun]

great test sjors love your work
now we need a bash em good test lol

sjors after doing his test

 

[Schirra]

:thanks Sjors for your job, I like the note book with equations, very scientific mood!!!!

|> :gold