No, the part on the right side of the peak was the transit, the peak was just included with the point scatter around the transit itself… What I was talking about earlier was graphs where there is literally nothing around but a peak or multiple of them and people mark that.
MMOS will be beginning their work to change out the data sets at 15:30UTC today (2017-07-12). We don’t have an estimated time for how long this will take, but we will update this thread to let you know once it has been completed. Please also be aware that there may be some interruptions to Project Discovery during this time.
Thank you for your patience
Damn that’s strict! I would have thought it’s a pretty good interpretation and kudos for spotting the pattern
Thanks for the update CCP Logibro. I am looking forward to give it a try.
Once the database update is complete, can we get a new thread to differentiate?
Still making my album.
Just so everyone knows, this is happening now. The service will be unavailable for some time.
Is it possible you guys set THE OLD LEVEL for the new PD? Because when i wondered why im failing all the time i retried it on an alt.
and EVERYTHING i got was like, solvable. While my main char, which was over level 400 in old PD, got the strangest samples ever.
I am trying to participate. Not sure what to say how well is it going. I am trying to follow up the patterns noted through tutorial and one video that is on youtube, covering this topic.
I tend to have an issue, so I am asking for a advice here.
I think that automatic grid is making it harder - is there any chance to make individual cue points/markers manually?
Or if anyone can actually let me know where/what I did wrong here?
No, this is not the case. There is an issue with the difficulty of the last set of evaluation samples that were being given out. This caused, in some cases, people to snowball-destroy their accuracy. This is being fixed now.
Okay, just trying to find out why my main got so many difficulties while an alt does it with ease.
Managed to scrape my way up from 13.7% accuracy to 49.5% using the new dataset.
It’s a symptom of you starting the project with your main when the problems were at their worst.
I got some cash payouts at least, and I had an accuracy run of 10/10 - its the “no transits” that confuse me now because I’m second guessing if there might be a hidden sequence in the noise.
Here’s the analysis that broke the cycle (when I was down at 13.7% accuracy) : http://imgur.com/a/G6Ti3
Well, ill just wait for the new set.
Cant say im… calm, though. =/
That one still you gain procentage of its just the adjsutment you got a little bit wrong had one simular and almost got it still i got procentage gain fromit
Have done some photometry and some differential photometry as an amateur astronomer, so I love the “game” here. A few minor complaints though
the color used for the data set is a bit too muted. Pushing the monitor doesn’t really help much as it also boosts the rest of the screen. I would love an option to be able to select colors, or at least transparency level, even at the cost of some aesthetics
since we are dealing with noise and patterns, it would be nice to either have some characteristics/info/SD lines (maybe we have those, a bit unsure about that)
some of the samples marked as actual transit are borderline - I am sure that given additional information or noise characteristics they have been scientifically confirmed, but I am not convinced that they are doable visually. A few of the very short period and weak transits I got wrong would have been transits I wouldn’t have dared to consider as successful observations in my own amateur attempts.
not sure the tutorial makes it very clear you can have transits by multiple planets. It is OK: when you fail one, your realize that there will be some (and incidentally, your rating still increases somewhat after such a partial failure)
Anyway, really cool - the only very frustrating part is the muted data visualization.
If it is post update then the number of difficult, if not impossible, graphs has increased.
Are you really deliberately trying to annoy people so much they have no intention of continuing.
I can forgive the fact that you threw l5 difficulty at players who were starting, After 100+ they did eventually get ONE sensible one that wasn’t just noise so they could actually progress.
but now, after getting to the high 80s, EVERY BLOODY “gold standard” sample is absolutely brain meltingly impossible, maybe .001% data amongst the noise. heading fast back to 1%
Of course if one had an entire day between individual samples, better tools, and a few years experience, one MIGHT get lucky one time in a hundred or so.
as it is, THIS IS A GAME. and we aren’t getting paid, and we are not going to be happy to write a paper about the ONE BLOODY transit we saw that a hundred other scientists would dispute.
Sort this out, if your dear scientist wants us to solve the samples they and their supercomputers cannot solve, he needs to modify his expectations.
our role is to be a first pass through the data, finding the more obvious transits for further investigation.
judging us by the “gold standard” of impossible transits, just won’t work.
I may or may not give this pile of rabid monkeys another go, or maybe not.
tl;dr sort your stuff out!
Edit, Right, that was the final straw, thats, 20% points lost, it seems if you fail a l5 every one after that “evaluates” uses a L5. so that’s it, send me a mail when this is not so absurdly broken.
Looks like the orbital period is off a tad. See where the lines look centered on the dip at the left, but drift leftward of the dip on the right ?