Please Read First!

i added two screen caps from google earth of the top 10 candidates and a zoomed in screen cap of candidates 2-9 which are closely grouped. i might add hotlinks to the seti pages, but i don't have time right now.

http://www.lxicon.com/seti/all.candidates.2009-08-20.jpg
http://www.lxicon.com/seti/candidates.2-9.2009-08-20.jpg

Forum Database question. It says there is one thread in this forum, when I count eleven. Is there an issue in counting them or is it intentional because NTPCKR is creating the threads?

As mentioned before (by Jon Golding), on the sky-plot of an NTPCkr candidate it would be nice if the region of interest could be shown instead of only drawing the target center with a marker on the Google map. This enhancement to the NTPCkr should now be straightforward to implement, given my latest blog post that provides the needed equations and an example JavaScript implementation that can draw these areas on Google sky maps. See http://blog.tiaan.com/link/2009/09/13/pixel-boundaries-ntpckr-seti-at-home for details.

With the coded function I could also visually see how four recent candidates are neighbouring regions (20504984, 20504975, 20504996, 20504973), as shown in the thumbnail image of the post. It was also interesting to see that Google’s sky map shows stars in these candidate regions, even though the NTPCkr table lists zero stars for each candidate – I guess these stars are too far away to really count or be catalogued. Nonetheless, being able to see the approximate regions of the candidates made it much more interesting for me. I do hope this enhancement would make it into the online NTPCkr soon.

With the coded function I could also visually see how four recent candidates are neighbouring regions (20504984, 20504975, 20504996, 20504973), as shown in the thumbnail image of the post.

Ultimately in 3D, obviously, yes. But knowing they are neighbouring pixels on the celestial sphere at the finest resolution that SETI is looking, does make it more interesting – even if just to suggest these candidates might have a glitch in common leading to their detection or that they are perhaps the same signal that is just spread over several neighbouring pixels due to some accuracy errors.

@Tiaan

This is a really nice way to show the candidates - well done, and I hope it can become implemented into the NTPCKr soon.

Actually, although I suggested showing the region of interest (and not crosshairs), it was Joe Segur who suggested showing the outlines of the Healpix on the sky map, which I think is a much more valid way to present the data.

Cheers,
Jon
____________

Let's not worry now who gets credit for finding the signal. Let's find it first. :) We can always worry about giving the proper person credit later. :)

Hey, you did the maths wrong!

According to FAQ, the score should be the probability that this happens randomly. If I remember my maths right, probability is a number between 0 and 1 characterising the occurrence of an event, where 0 means that it happens never and 1 that it happens always (with every try). So a negative probability which all the top candidates have is mathematically meaningless. (If I calculated a negative one that means my calculation is wrong). Well, maybe the ranking is right and I don't want to get too nerdish about it :). Just wanted to point out, that these numbers cannot represent real probabilities.

Those numbers are no probabilities, they are (meta)scores.

Gruß,
Gundolf

Metascore: The cumulative score of everything that made up this candidate (the individual signals and multiplets). If we did the math right, this is the probability we'd see this candidate in random noise (so the lower the score, the better).

Any chance of NTPCkr being updated soon? It's been two months, unless I'm misunderstanding the purpose and interval of the tool.

Well, it looks like the web page access to the science data went down a month ago: since then I cannot look at the candidate details. The number of signals at the current spot in the sky on the science status page has vanished, too. Not that these numbers would actually have meant anything because the telescope position and status hasn't been updated for half a year as well.

So... there are some issues to fix before any update would make sense even though the NTPCkr is planned to update in near real-time in the future. But if I read through the technical news, they seem currently occupied by fixing other issues and getting that software radar-blanked data to us, so the web stuff is apparently one of the lower things on their priority list. They simply don't have the manpower to get more done in shorter time.

ap_08se06aa_B2_P0_00097_20091028_17830.wu_4 WILL ARIVE LATE!

it will NOT be done by 07:09:59 pm 11/28/08.

If this task is withdrawn, I request credit for the 14 and a half CPU-days my system has put into this monster.

I don't know how I was assigned this job in the first place

I also fear even cosmic rays on a job this big.

Any chance of NTPCkr being updated soon? It's been two months, unless I'm misunderstanding the purpose and interval of the tool.

Yeah! Looks like the NTPCKr is running again.
8 new candidates, although we don't yet have access to the full WU details.
Well done everyone at Berkeley.
____________