Devblog: Exoplanets: The Next Phase Of Project Discovery


(Arielle en Distel) #21

@CCP_Phantom You guys didn’t think about accessibility when developing this did you? The tutorial text is extremely small and I can’t figure out how to make anything in the Project Discovery window bigger…

(Lain Deudigren) #22

As someone with many hours, well over level 1000 and a 99.9% accuracy in stage 1 after playing this for half an hour and failing every time all I can say and advise is that CCP rethinks what it has done.

After being told that I did everything right except 2 samples in the tutorial I am now at 26.9% accuracy. I cannot spot ANY of the patterns I was asked to find in the tutorial except I get shown patterns I could not identify even when folding.

All my hard work in stage 1 seems to have been spat in my face and I hope that if I am not the only one CCP immediately rethinks the entire system.

I like to help but it seems ‘stage 2’ is currently not for me.

(Mirsa Attor) #23

I too was at 90+% accuracy on sisi and cant seem to find a single pattern on TQ anymore. I’m now at <20% accuracy and stopped playing.
Whatever changes were made on the transition from sisi to TQ -> they werent good :frowning_face:

Edit: My most interesting sample so far:

(Kiauze) #24

One of Discovery’s “patterns”

I’d like to ask how are we supposed to detect these sort of patterns, which for all intents and purposes are virtually indistinguishable from the background noise.

I love the concept of this project, but I’ve been failing almost every attempt due to either these sort of “patterns”, or the detection being off by just a few pixels to the right/left.

(DiDDleR) #25

Tutorial samples were easy - now dropping in accuracy with nearly every sample incorrect. I’m being told that I’m missing transits that are hidden within the noise of the sample which are impossible to spot.

Also the window cannot be stretched or maximised to be as large as the screen. My window currently won’t go any bigger than approximately 1/2 the screen size. Making it bigger would help us to spot the transits easier without having to zoom in and scroll the sample repeatedly.

(DiDDleR) #26

Nearly all the samples I’m getting are coming back Analysis Failed when there are no obvious transits in the data.

July release - Issues
(Joia Crenca) #27

I’m at 2/10 “Editing the orbital period” and can’t seem to find the action needed to satisfy this section of the tutorial.

I’m also wondering if there will be replies to this thread or if a different one was intended to be used for Exoplanets?

CCP, intentions are great, but working projects are important. Please allow staff time to get new features and projects in good shape before deploying. Some ‘Corporate Bean Counter’ causing projects to be rushed shouldn’t be allowed to bring CCP down via a string of broken features.

(Winter Archipelago) #28

When narrowing the time window to less than max, the displayed percentage for the luminosity changes to only show the range needed for that particular segment of time. This causes the graph to both bounce and stretch quite badly while scrolling through it, and makes it significantly harder to see what’s going on because the graph data is basically changing every time.

I’d make a strong recommendation that the graph be kept static, displaying the full range of the luminosity, even if it doesn’t need all of it for that individual segment.

(Ideki) #29

I agree with @DiDDleR

Also the window cannot be stretched or maximised to be as large as the screen. My window currently won’t go any bigger than approximately 1/2 the screen size. Making it bigger would help us to spot the transits easier without having to zoom in and scroll the sample repeatedly.

the window needs to go has big as the screen can.
Because I have a very high resolution and wide screen too, and it’s annoying to have to zoom/scroll continuously.

(Ideki) #30

I have been toying with the project Discovery a bit and I got to say that while it is fun, it gives me sometime weird (unjustified ?) results.

Take this one for example:

It says that I failed, when my markers are right on the spot.

Or in this one:

It mark the ‘correct’ answers where there are no sharp changes, nor patterns.

So I am curious a little bit about the validity of the results that we are compared against.
Some explanations about those results would be welcomed to avoid further frustration and ensure the success of project Discovery.

(Ideki) #31

@Mirsa_Attor, I got that exact same sample too. :grin:

(Winter Archipelago) #32

I’m not sure what’s wrong with this one:

There are clearly passes of something there, but if I had tried to set a second object, it would have had an identical orbit.

(Duo Roman) #33

You failed what may be an obvious transit of a planet because it was not a transient, it probably is a binary star.

(luftwraith) #34

It seems like the “grading” algorithm is making transits where no one would see one. The tutorial says finding the transits would be rare, but the machine says there are some on every sample, I miss them and am graded as a failed analysis, my accuracy is now so low that I get nothing for my effort.

So . . . back to ratting.

(Julie Morano) #35

Got to 75% accuracy in an hour or so on SISI.

On TQ I can’t get anything right.

Accuracy down to 39%.

I’m outta here…

(Edlorna Tinebe) #36

I, too, have been surprised when the computer tells me I missed a transit in all the data (Usually ones with periods less than 24 hours - I’ll have to keep that in mind and improve on it). But I’m not crying foul. It took me a bit to get used to the cell analysis, too.

One thing that I think would be a HUGE help for myself and everyone else would be the ability to continue operating the zoom and fold and period controls after we’ve been graded. Let me zoom in, fold the graph, and SEE what it looks like so I can catch my mistakes next time. Seeing a couple of boxes on the full 30 day graph isn’t helpful.

(Kiauze) #37

This is actually different from the cell analysis.
Apart from the oddball sample there, most errors I did were done by a miss-classification of present structures.
Here, we are being compared against something the human eye can only perceive as random noise (check picture I posted earlier), especially since the tutorial made it VERY clear we should be searching for U or V shaped dips, and most of the errors I got had neither.

(Ideki) #38

Another useful thing would be to be able to zoom with the scroll wheel of the mouse when the mouse.
So that we could faster check specific portions of the graphs instead of having to drag both corners of the full graph.

(Kiauze) #39

You can actually do that on the zoom bar. :stuck_out_tongue:

(Ideki) #40

Yeah[quote=“Kiauze, post:39, topic:8026”]
You can actually do that on the zoom bar. :stuck_out_tongue:

Yeah, just found out.
It’s just that the scroll is very slow for some reason.
Like I have to scroll A LOT to get it to shrink/grow a little bit