Devblog: Exoplanets: The Next Phase Of Project Discovery


(Altrue) #61

Well that’s one feature I sure won’t try for now.

I love my screen too much to do something that will make me want to throw it out the window.

(Cliff Beauland) #62

Aight then I have to stop before I am punished any further from attempting to understand what feels to me an inaccessible minigame.

Things that may help.

  1. When placing the transit marker - reflective data points should be highlighted around (as it does in the analysis results) - this would aid in identifying the patterns a little clearer when folding. The high contrast could easily allow the the Us and Vs to see a bit better when layered on top of one another - rather then a mess of pale gallente/caldari blue.

  2. Softer resolution averaging - again reduce the noise, discover the curve better.

  3. A clear breakdown from CCP as to how these results are being computed in comparison. What is making the subjectivity of these answers so - when many of these results are looking objectively questionable at best.

  4. Consider resetting average accuracy scores after some fixes? I find it a slap in the face to punish those early adopters (some who had good results on SiSi) and make them work more now that their averages are down - than someone who comes along once some tweaks are made. Maybe that’s just me though.

For now ill keep being a Sig explorer until I can help the real explorers.

(Edlorna Tinebe) #63

If I recall correctly, the second example doesn’t have a period for you to orbit. That’s the one with only a single dip, right? Click on it twice to indicate there’s no repetition.

(Tipa Riot) #64

Have no problems, already level 1 after a couple of samples. As easy as on SiSi.

(Miiral) #65

Can confirm, something has changed within few hours, when I got nothing right. Now it seems you have to put an effort to miss :smirk:

(CCP Phantom) #66

Hello wonderful space explorers! I would like to point your attention to our news item regarding an issue with the evaluation set of exoplanet classifications.

To cut it short: Due to a mistake, the training set is too difficult. We are working on a fix (scheduled for July 12). After this has been fixed, we will reset the accuracy settings to 50% for those who had dropped below the 50% rate.

Thank you for your feedback and for pointing out this issue. Please keep the discussion about this specific issue in that thread.

(Kiauze) #67

@CCP_Phantom thank you for the update!

(Tipa Riot) #68

Just reached level 10 with 68.1 accuracy. There were couple of odd samples like on SiSi, but I don’t see any difference in difficulty.

BTW, you can’t trade the two caps on the market. Please fix.

BTW2, Novice Analyst level 2 is not the rank upgrade I expected … lol

(Edlorna Tinebe) #69

I have no idea what’s going on with this star:

There’s a very clear pattern there, with a sharp dip, followed by a longer curving dip. The period is only .07 days - 1.68 hours - which is well below the Project Discovery minimum period of 0.5 days.

I mean… I’m pretty sure it isn’t a planet with an orbital period of 1.7 hours, unless this is from some weird Dragons Egg style neutron star or something. What would produce that kind of pattern?

(Tvashnar Crendraven) #70

Get rid of the star shining bright light into my face, and add the result of an fft.

(Lu Kun) #71

I think there are some things you could do with the new Project Discovery to make it better. (and sorry for my english)

  1. Chat channel
    Add a chat channel under “Content” for “Project Discovery”.
    Add a “motto of the day” with links to:
  • current issues
  • eve-wiki (containing also a link to “the real stuff wiki”, screenshots of the tutorials (I expect this to be usefull when discussing things with other players))
  • feedback thread
  1. User Interface
    There are several points I think you can do here.

2.1 Alignment of tutorial windows
I set my UI scaling to 125% because my eyes aren’t getting better. Unfortunately, the tutorials started somewhere out of the window so it is impossible to read all of it. Moving the Project Discovery window helped abit but please… fix this.

2.2 Solar Activity Classification
I had to ask in the Help channel wether I had to identify the correct classification(s) or not. It is a N-of-M-selection (means if you click on any classification its border is highlighted and will stay that way till you click it again). If they are for information only, please make this visible somehow. Maybe even move the tiny little questionmark to restart the turorial to that place too.

2.3 Add a slider between the top, middle, bottom region of the window
I’d like to increase the hight of the topmost part because I think it would help focusing on details. The bottommost thing with the solar system is… nice and i would not throw it away but if I could hide it or make it smaller while having a look at the data would be nice.

2.4 Feedback when failing
I am failing to identify the transits very often. That’s ok. But the information why I was wrong, your reasoning, should be improved. Show me what was wrong. Maybe zoom into the block there I failed. I remember there was something in the tutorials where it said “look at the U shaped region” and i was “what the f***?” because you called some points that looked like a broken line a U . Zooming into the block with the problem highlighted would have made it visible - maybe.

2.5 Export of the shown data
Absolutely no idea if this is already possible or if I can download the data somewhere else but I think it might be interesting to be able to export the currently show data - similar to the market data stuff.

  1. Advanced Ideas
    You showed some nice things and I really like having a look at it and trying my best but I am not a physics guy. I can’t imagine some of the things I am looking at right now. What might help me would be if you’d add a simulation “the other way round”.
  • How would our solar system look from planet Vulcan (Star Trek… Spocks planet), if they would have looked to our sun?
  • Show our system in the same window, with all planets and the according measured data (which can’t be measured but had to be made up as if they were measured)
  • Allow to Remove a planet from the system so I can see how it affects the data
  • Allow changing the solar activity classification so I can see how the measured data might change

regards, Lu Kun

ps: anywhere a manual how to write in a forum? neither can i find a preview for my post nor how i remove numbering…

(Celise Katelo) #72

Started out getting everything correct, but tables have turned, i can’t get a single one right. The samples are currupted perhaps ?

(Blesniv) #73
patch failed (:

(Julie Morano) #74

Still have 26% accuracy.
Still have opaque samples.

(Tom Forager) #75

I have a sad. No idea of what is going on.

(CCP Phantom) #76

Thanks for the quick reply and for re-checking Project Discovery. Project Discovery and its light curve sets are not coupled to a patch in EVE. Those sets can be changed without a downtime or a patch. Please keep an eye on this thread for updates!

Also, I would like to point to our live stream today at 17:30 UTC about Project Discovery.

(Blesniv) #77

YAY! A real tutorial!

(Captain Campion) #78

Today I’m losing accuracy when I’m wrong but not gaining any when I’m right. So I’m just slowly bleeding accuracy. This is since the PD downtime earlier.

(Jian Mira) #79

I just don’t get it, far too complicated to understand and a waste of time

(Tipa Riot) #80

Was there anything patched today? I don’t see any quality improvement to the control samples. Today I was unlucky and got a couple of “impossible” samples in a row. Means those with noise marked as transits or obvious transits not market (but some noise).

I have to admit if I now spot a sample with some potential transit, I skip it and try the next one and only submit when absolutely sure it is not a control sample or the orbit is “110%” obvious (and a control sample).