https://puu.sh/wCR19/70b3da99e3.png < I got the Discovery project stuck in an endless loop. EBR-125541 and again EBR-125543 and a few times after, opening and closing the window does not help. I get stuck after 4/5 tries. EBR-125546 < got it on the first opening now, got that one going after opening / closing window. EBR-125547 < same session, got a new sample after opening / closing the windows a few times. on opening and closing the window, haha < https://puu.sh/wCS9s/dd9b803d8b.png. It seems the Sample session loader is broken, the weird thing is, that the overlay UI is not effected.
I think you can push now for really that one thing before it will be closed, the previous project is being scraped anyway, filfilled its purpose. You dont have to worry about anything like accuracy rating going really bad.
Oh, levels reset ;)… lol, I get suddently a lot of sets with where samples with very small trends break over time in stead if the bigger ones, this is new.
https://puu.sh/wDHby/4de332d0ab.png and again, we really need color grouping per repeat, so we can see how the periods are stacking in the fold, this 2 outliers were in such a big noise stack, there is no way you could have forseen the failing.
https://puu.sh/wDHgw/bd15e58cde.png < a presubmit, how to see if this is a good position, I cannot zoom in, I cannot see different periods in the stack, how to tell if it is good positioned. If 2 out of 14 like the one before are a near match it should not be a problem for a successful analysis.
Why does a discarded fold reset the zoom level of the second pane?
When do you get consensus feedback on samples you are the first to classify?
You actually can, clicking the green or yellow markers will show you the fold that produced that solution.
You currently don’t get feedback on those samples you were the first to classify, in the original Project Discovery, we originally gave people retroactive consensus feedback because we used consensus data to set the accuracy of players. That method proved to be unreliable so we now only provide consensus feedback when a consensus already exists.
Regarding the trouble with the outliers, it’s a fairly hard problem to tackle. Because a small error in period accumulates over each repeat to eventually become big enough to overcome the margin of error that our samples use. The color method does help with some samples, but we found it that it actually increased visual noise for some samples, especially samples with a high number of transits.
https://www.youtube.com/watch?v=fAZEyDntQGo < made a gameplay video for new testers, hope CCP likes the way and setup, never publiced game material before, hope this is ok.
I’m currently up to about my 40th sample and i have a few comments suggestions.
When you complete an analysis you get what i think of as the “Results” page that highlights all points involved with the transit. When i get “analysis failed” the green and red dots tend to mask the data. would it be possible to add toggle button to turn them on and off so you can look at the raw data. Some of the transition events are very subtle and I’m having trouble determining what they look like.
So far i have only gotten “analysis successful” for zero transits or one transits. I’ve had at least two cases there i could see possibly a second transit. are there any known instances of multiple transits in the dataset?
Can you add a button somewhere that pops up a window to give pictorial examples of what transitions can look like. The tutorial gives you some, but by the time you have completed it, you have forgotten what some of them look like. Some of them are very subtle.
Is it possible to add the zoom function when in folded mode. sometimes when you have a period of 15d or so lining things up is a little difficult.
In one set I once had 3 transits. I should have made a printscreen, 2 is more common. https://puu.sh/wBfSF/2abbfecbd7.png < Screenshot of a 2 transit.
Can we get a setting that enables an extra vertical line halfway of a to set trend before folding, so, if a pattern between 2 points is obvious, it would be nice to see if there is something happening half way.
Also, can we get a “defold” button when you are in a folded state in stead of just a “discard”, thx.
The image you linked there looks very artificial, the one with the two transits. I’m having a hard time believing there is a binary system with a planet in a kind of a Larange point perpendicular to the mid point of the two stars… The planet must be huge to eclips both suns at once
I also thought about an “unfold” button, but then i realised the “complete” kindof does the same job.
Me thinks that suggestions now are a little late for the formal release tomorrow.
I do hope the isk rewareds get buffed a little. 57k isk after 2+ hours is a little stingy!
