anyone know what triggers double xp?
Have to agree with Aineko on this. Addition of a related FIR will assist. Also as we say to folk who work in the RF environment. DO. NOT. TOUCH. AUTOMATIC. GAIN. CONTROL !!! Let the user have a choice of auto or manual gain control on the y-axis as this auto adjustment when zooming and moving the window is not helping. Also please let us set the central axis point of y along with the min/max marks.
Merci
So I Love project discovery but I have been finding a lot of slides with transits not able to be marked. This specifically applies to slides with multiple object transits. If you mark the smaller object also and not just the largest object transit. I.E. 2 different spikes unrelated to each other it will tell you analysis failed even though the second object was blatantly obvious.
I dunno I woulda marked 4 transits. May not be a planet but there is for sure a pattern and points/times of interest. As far as the accuracy rating youāll probably get it wrong because currently it wont accept a second transit and still give u credit.
I am not very good at this, but I managed to find a sneaky transit by trying a new (for me) tactic based on folding a spectra as tight as possible and then side scrolling and looking for dips. I havenāt read anything similar, so decided to share it here. But it still fails on multiple occasions, when the dips are hidden in the noise.
The tactic I used was
- Locate a dip that is most likely a transit. If you are wrong the dip will be off-centre when you side scroll later, making it harder but not impossible to spot it.
- Fold to the smallest possible (0.5 days) and start scrolling sideways (by at least +0.5 days, to 1.0 days orbital period, to allow everything a chance to line up) while looking for overlapping dips. Dips with a longer or shorter orbital period will have noise mixed in, so pay extra attention to the lower half of the spectra as you do this. If you are unsure about a dip, try looking at it with the double orbital period to see if it is still there (write down the epoch and orbital period if you think you might loose track of it).
- Sometimes a tightly folded spectra can look like a zig-zagging wave, I believe this is a recurring pattern that is caused by the sun itself and this should not be mistaken for a transit.
Some work in progress pics I took on the 200075050 spectra
(I noticed that this is the exact same spectra as Bienator II posted above, post 153)
a) Messy spectra, I canāt see any dips
b) Zoomed in and looked around, found a possible dip that I decided to work on. If I am wrong, the dip will be off-centre, but it will still show up.
c) Folded at 0.5 days length (the smallest possible) and started scrolling to the right, looking pretty flat at first
d) What do we have here?
e) Last pic before submit
f) This time it was a success. But I still have many failures
3 Likes
Overall, Iām very happy w/ Exoplanets. Itās a lot of fun.
Itās good to have the feedback when you get a failed sample so you can learn and improve.
However, Iām not sure what to make of some (seems a lot) of the standard samples where it seems all but impossible to determine the correct results even when they are pointed out. Many have posted examples aboveā¦hereās mine;
Edit; ānew users canāt post picturesā. Hereās a link, http://i.imgur.com/zGZavSi.png
Now that the community has some experience w/ Exoplanets, it would be great to get some further guidance or tutorials on how to analyze the more difficult samples.
Some are evil, but not wrong. 
Why are we getting a blueprint copy when we hit level 50 instead of the infinite runs blueprint that shows as the reward for level 50? Seems like CCP is misrepresenting what you get for hitting that level.
Iām having issues with some transits not being discernible from the background. Even when I get a fail and it shows where the suspect transits are I donāt see them standing out from the clutter.
My feedback on project discovery:
- So-called grade 5 (and possibly grade 4) difficulty tasks are 0 fun to me.
- The forums have āguidesā on how to āwinā at these, and they are all pure ā ā ā ā ā ā ā ā . Iāve used them and not found a single pattern like what is described. I see ā ā ā ā ā ā ā noise, and PD claims I missed some super-frequent transit.
- 80% or more of the analysis slides I get, where I can have my accuracy adjusted, are this type of āfind the invisible needle in the noise-stackā. I bounce between 55% and 58% accuracy.
Give us a UI option to blacklist grade 4 and 5 challenges if we arenāt good enough to solve them, or limit those challenges to people with 75% or higher accuracy. I am seriously considering uninstalling EVE over this. I installed the game specifically because of project discovery, and my SSD space is precious to me. There are far less irritating games that could be occupying that space.
THINK! There is no reason to frustrate people WHO WANT TO HELP YOU! If we donāt find playing invisible-needle-in-the-noisestack, then donāt make us ā ā ā ā ā ā ā play it! Let us work on the easier transit data. Or give us the tools to actually find these things consistently. Until I see a livestream where these scientists manage to find 100% of these super-hard transits using the in-game UI I will maintain this position because there is no ā ā ā ā ā ā ā way they actually found these things using the tools we have.
Fix this broken ass idea and make it great. Like it should be. I donāt care whose fuckup it is, the scientists, CCPās, or someone in between. Fix it!
That really sucks. How many runs did it have?
I was grinding this with the expectation of a BPO with infinite runs that was shown in the game. This greatly reduces the incentive and it upsets me that this issue has not been corrected. CCP should honor the reward promised in game to those who have earned it and update the reward so others donāt expend their energies with false hopes of a BPO.
some are hard, some are maybe too hard. but at sub 60%, youre just overseeing a lot.
Also, being annoyed and loud doesnt make you right. Look up the livestream, if you want more information.
Iāve started doing that with the noisier data sets, and I almost want to ask for an couple buttons to set the folded period on āfast forwardā or ārewindā so I donāt have to mess around with the scaling scroll wheel.
Iām done with it, no point leveling up past 25.
I think that part of the problem is that some people come to Project Discovery expecting it to be fun, and to be a game. Analyzing these data samples is neither, after that shiny āIām experiencing something newā period has passed. If it were fun, theyād be able to put this ā ā ā ā up on some mobile phone app and let the community consensus roll in. Iām pretty sure this was touched on in the recent twitch stream from CCP. Citizen science projects donāt generally do so well because most people will play with it for an hour or two, say āThatās neat!ā and then never touch it again.
Project Discovery exists explicitly because MMO gamers areā¦ I think the popular term on Reddit is āautistsā. We are willing to sit for hours at a time doing explicitly un-fun grindy things for hours at a time if it gets us a shiny new ship or skin or labcoat for our avatar.
You donāt find invisible-needle-in-the-noisestack to be entertaining? Guess what, neither does anyone else. But thatās exactly the job that PD is paying you to do, at 50,000 ISK per sample. Grade 5 and 4 analysis tests are zero fun to you? You think that the people behind PD are going out of their way to frustrate you? Itās a frustrating job!
I guarantee you that the actual astronomers who work this data for a living can do it with the tools weāve been given. The difference is that they might actually take ten minutes per sample, poring over every bit of the graph, stretching and refolding it to make sure theyāve found everything. You, and everyone else who complains that they canāt find the needles in the noisestack, arenāt taking the time to do it right, because you expect it to be a minigame, not a job.
TLDR: This is EVE. HTFU.
Itās a one run copy. CCP pulled a fast one on us again.
Itās frustrating to fail analysis with low-SNR patterns like many screenshots in this thread.
Two tools that could help me notice those patterns:
- Horizontal line at 100%
- Pencil tool:
Every click on a potential transit sets a new pencil marker
Recalculate the matrix of time deltas between every pair of pencil markers:
f(t0, t1, t2 marker) dt {
return [t1-t0, t2-t0, t2-t1]
}
Out of all the deltas, look for a set of deltas that are almost equal.
Flag these markers for closer inspection
1 Like
I have to agree; Showing the people they will get a certain reward, and then claiming its a bug. It is a really bad move. It doesnt matter the bug has been there for a long time. Its just a proper fu to the players.
A horizontal line would be very helpful, Iām waving around a ruler infront of the screen when I need a reference line to compare dips against.
Not so sure about a seperate pencil tool though. Maybe if it was built into the current line tool, so that min/max numbers within a certain distance of the marker (in the detrended mode) were shown. Otheriwse I donāt see how it would help to find reoccuring trends. Perhaps I didnāt understand what you meant.
One issue I have with spotting overlapping dips in the folded view is the lack of depth differentiation. When I canāt tell the layers apart the dots can merge into blobs if there is even a slight y-axis separation between each fold.
Maybe it would be easier to tell different folds apart if a moving average line could be turned on, one that also becomes folded when the spectra is.
Scrolling a folded spectra long distances with the mouse is bad ergonomics. Arrow keys would help.
@Buzz_Driver I want to mark anything that looks like a dip, and only then look for patterns in the time deltas between my marks. I called it pencil because Iād discard the marks when i no longer needed them.