star distance calculation error (solved)

phoenix · Post #21by **phoenix** » 14.09.2006, 21:45

rra wrote:
The problem is with Sol's position, and NOT the coordinate system. Actually, Sol isn't placed at (0, 0, 0). It should be there ! That's all.

I know that, but what has that got to do with the fact that the distance between the stars is not calculated correctly. I shouldn't matter at all.

Ren?©

exactly.
also changing Sol's position to 0,0,0 didn't change anything.

Cham · Post #22by **Cham** » 15.09.2006, 01:07

Oh, you mean distance between star A (not Sol) and star B (not Sol either) ? Of course then, changing Sol's position wont make any difference.

phoenix · Post #23by **phoenix** » 15.09.2006, 13:36

I found the bug which is no bug at all

I've just taken bad examples to compare distances.

HIP 32349 (Sirus A) is a binary starsystem and instead of calculation the distance from Sirius A I should have taken the Sirius barycenter.

the other examples i checked were all binary or even tri-starsystems... the funny thing is I chose those examples just by clicking around in the univerese and never intend to check distances of binary-systems

the calculation from the barycenter gives just perfect results :!:

Post #24by **t00fri** » 15.09.2006, 14:02

phoenix wrote:I found the bug which is no bug at all
...
:

phoenix wrote:I'm getting the same margin of error between 0.02 and 0.05 ly for any random star-combination.

Well, I ask myself why you chose multiple star systems in the first place, to check distance calculations at the precision level?? Why not using well-known single stars. There are plenty.

Also picking multiple star systems is ALL but a random star-combination. You can't have checked many such "random star combinations" to make up for your above general statement? ;-)

Bye Fridger

phoenix · Post #25by **phoenix** » 15.09.2006, 14:34

Also picking multiple star systems is ALL but a random star-combination. You can't have checked many such "random star combinations" to make up for your above general statement?

Bye Fridger

the problem was i picked 5 random stars which now all checked out as multiple star systems (seriously picked by accident) and then testet over 5000 star distances using those 5 stars as origin...

I compared the values with the top-left "distance to" field and the star-browser.
and because my origin star was moving and naturally not at the position given by the raw stars.dat all calculated distances were off.

:oops:

Post #26by **selden** » 15.09.2006, 14:53

About 20% of all "stars" are actually multiple star systems, so it's not all that unlikely that you'd manage to pick doubles by accident.
[edit]
There are only about 200 binaries in the STC catalogs included with Celestia, however. So that makes it a little less likely

[/edit]

http://spider.ipac.caltech.edu/staff/ja ... iples.html
discusses the distribution of multiple star systems relative to the galactic plane.

Post #27by **t00fri** » 15.09.2006, 15:21

selden wrote:About 20% of all "stars" are actually multiple star systems, so it's not all that unlikely that you'd manage to pick doubles by accident.
[edit]
There are only about 200 binaries in the STC catalogs included with Celestia, however. So that makes it a little less likely

Since it was me who added the 200 binary star orbits, I took that tiny binary fraction 200/112519 = 1.78 per mille (!) as an indicator that phoenix's statement

phoenix wrote:I'm getting the same margin of error between 0.02 and 0.05 ly for any random star-combination.

must have been based on a vanishingly small statistical sample which does leave us with a fair amount of questions... ;-)

[edit]
Statement weakened after PM by Selden, instructing me about the possible occurrence of improbable events... ;-)

[/edit]

phoenix wrote: ...the problem was i picked 5 random stars which now all checked out as multiple star systems (seriously picked by accident)

Put differently: given the probability of only 1.8 per mille to randomly pick one binary (with known orbit) in Celestia, phoenix' claim to have picked accidentally 5 such guys at the first go clearly points towards a breakdown of the laws of statistics ;-)

Never mind ...

Bye Fridger

PS: I did know about the true fraction of binaries that is indeed quite high.

Chuft-Captain · Post #28by **Chuft-Captain** » 15.09.2006, 15:53

Put differently: given the probability of only 1.8 per mille to randomly pick a binary (with known orbit) in Celestia, phoenix' claim to have picked accidentally 5 such guys at the first go clearly points towards a breakdown of the laws of statistics

A sequence of highly improbable events does not nescessarily point to a breakdown of the laws of statistics, as it is also included as a possibility in such laws. :wink:

Post #29by **t00fri** » 15.09.2006, 15:58

Chuft-Captain wrote:
Put differently: given the probability of only 1.8 per mille to randomly pick a binary (with known orbit) in Celestia, phoenix' claim to have picked accidentally 5 such guys at the first go clearly points towards a breakdown of the laws of statistics
A sequence of highly improbable events does not nescessarily point to a breakdown of the laws of statistics, as it is also included as a possibility in such laws.

Of course I know how to deal with statistics, but after a PM lecture by Selden about statistics and politeness, that was the only polite way out... ;-)

Bye Fridger

Post #30by **t00fri** » 15.09.2006, 16:30

phoenix wrote:...the problem was I picked 5 random stars which now all checked out as multiple star systems (seriously picked by accident)

...with a probability of 1.8 per mille for one such pick.

This reminds me of a standard joke long ago when we were young math students, discussing about how statistics had to be interpreted:

( it was also long before terrorism on airplanes was an issue)

One guy told another one:

Oh I am always horribly afraid in airplanes that one of the passengers is a terrorist and blows up a bomb on board.
Replied the other one: I have no problems with this. I usually take a bomb along myself, since the probability that TWO bombs explode in the same plane is vanishingly small...

Bye Fridger

phoenix · Post #31by **phoenix** » 15.09.2006, 16:31

t00fri wrote:must have been based on a vanishingly small statistical sample which does leave us with a fair amount of questions...

yes you are right my statement was a bit too exaggerated only testing 5 stars as origin :wink:

*edit* but...

Since it was me who added the 200 binary star orbits, I took that tiny binary fraction 200/112519 = 1.78 per mille (!) as an indicator that phoenix's statement

is also not exatly true.
i'm usually navigation with an auto-magnitude of 8.5 which shows me far less than 112519 stars to pick from

phoenix · Post #32by **phoenix** » 15.09.2006, 16:33

Oh I am always horribly afraid in airplanes that one of the passengers is a terrorist and blows up a bomb a board.
Replied the other one: I have no problems with this. I usually take a bomb along myself, since the probability that TWO bombs explode in the same plane is vanishingly small...

Bye Fridger

you made my day :lol:

Chuft-Captain · Post #33by **Chuft-Captain** » 15.09.2006, 16:39

t00fri wrote:Oh I am always horribly afraid in airplanes that one of the passengers is a terrorist and blows up a bomb a board.
Replied the other one: I have no problems with this. I usually take a bomb along myself, since the probability that TWO bombs explode in the same plane is vanishingly small...

Now you're scaring me Fridger! ... friends of yours?? :lol:

Post #34by **t00fri** » 15.09.2006, 17:07

phoenix wrote:...
*edit* but...
Since it was me who added the 200 binary star orbits, I took that tiny binary fraction 200/112519 = 1.78 per mille (!) as an indicator that phoenix's statement

is also not exatly true.
i'm usually navigation with an auto-magnitude of 8.5 which shows me far less than 112519 stars to pick from

Since it was me who incorporated the automag-scheme, let me point out that your setting refers to a magnitude threshold of 8.5 at 45 degrees field of view! When you decrease the field of view with this setting you catch stars exceeding apparent magnitude 10 (try it!) . Hence more or less the full HIPPARCOS sample was at your disposal for picking!

Too bad...

Bye Fridger