These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

Test Server Feedback

 
  • Topic is locked indefinitely.
 

Project Discovery - March Update feedback thread

First post First post
Author
Previous Topic Next Topic
Yadaryon Vondawn
Vicanthya
#21 - 2016-02-26 16:39:57 UTC
CCP Paradox wrote:
Yadaryon Vondawn wrote:
CCP Paradox wrote:
Fantastic thank you, I'm already on them :)


By the way, are there more items planned which can be acquired solely through Analysis Kredits? In your Drifter Booster post you stated that they can also be acquired through other means. This does mean that effectively only the clothing is unique to Project Discovery. I do not know in which other way the Drifter Boosters can be obtained but if it is to simple I fear it may lower the incentive to actually do it.

Are there any thoughts going on about this?


There are more items to come. There is also no maximum rank or cap set in place. We can keep adding new items as demand drives it, and we will be watching this closely after launch. Feedback on rewards is more than welcome here, we want to hear from you! Big smile



Thanks for that! This does begs another question tho:
The ranking still appears (or I missed something) to be linear. This means that one can level very quickly to very high levels. The AK reward keeps going up. This means that a high player can do the same amount of classifications and get way, way more AK. Think 80 AK vs 6000 AK for the same amount of classifications (rank 3 vs rank 102). How is this being tackled?

As for the rewards, unique items are always nice. Any other item will probably hurt the market for that particular item making it hard to give those out. Personally I like usable items, like the boosters, but I can imagine a special SoE skin that can be bought with AK would be nice aswell.
CCP Paradox
#22 - 2016-02-26 16:58:31 UTC
The ranking is indeed linear for the test server, and will be changed for launch.

CCP Paradox | EVE QA | Team Phenomenon

Space Magician
Nana Skalski
Taisaanat Kotei
EDENCOM DEFENSIVE INITIATIVE
#23 - 2016-02-26 17:24:12 UTC
I had Project Discovery icon moved from business part of Neocom menu to the Neocom bar before the update. Now when I removed it, I cant find it anywhere. Its nowhere to be found. What?

ᴇᴠᴇʀʏ ᴘᴀʀᴛ ᴏғ ᴀ ɢᴀᴍᴇ ʜᴇʟᴘs ᴛᴏ ᴛᴇʟʟ ᴀ sᴛᴏʀʏ 📕

ᴡʜᴇʀᴇ ɪs ᴀɴɢʀʏ ᴄᴏɴᴄᴏʀᴅ ɢᴜʏ ᴡʜᴇɴ ʏᴏᴜ ɴᴇᴇᴅ ʜɪᴍ

ᴏsᴘʀᴇʏ 🚀

≡⋁≡ ♥
Yadaryon Vondawn
Vicanthya
#24 - 2016-02-26 17:31:11 UTC
Nana Skalski wrote:
I had Project Discovery icon moved from business part of Neocom menu to the Neocom bar before the update. Now when I removed it, I cant find it anywhere. Its nowhere to be found. What?


I had this problem aswell but realized you could map Project Discovery to a key in the shortcut menu ;)
Nana Skalski
Taisaanat Kotei
EDENCOM DEFENSIVE INITIATIVE
#25 - 2016-02-26 18:04:02 UTC
Yadaryon Vondawn wrote:
Nana Skalski wrote:
I had Project Discovery icon moved from business part of Neocom menu to the Neocom bar before the update. Now when I removed it, I cant find it anywhere. Its nowhere to be found. What?


I had this problem aswell but realized you could map Project Discovery to a key in the shortcut menu ;)

Thanks, but if it will behave like that on tranquility, people will not be happy.

ᴇᴠᴇʀʏ ᴘᴀʀᴛ ᴏғ ᴀ ɢᴀᴍᴇ ʜᴇʟᴘs ᴛᴏ ᴛᴇʟʟ ᴀ sᴛᴏʀʏ 📕

ᴡʜᴇʀᴇ ɪs ᴀɴɢʀʏ ᴄᴏɴᴄᴏʀᴅ ɢᴜʏ ᴡʜᴇɴ ʏᴏᴜ ɴᴇᴇᴅ ʜɪᴍ

ᴏsᴘʀᴇʏ 🚀

≡⋁≡ ♥
Oktura Ostus
THE CORP WITH NO NAME
#26 - 2016-02-27 00:49:58 UTC  |  Edited by: Oktura Ostus
Please, add 'blue only' filter. It would be helpfull to find holes in blue "necleus", so it will be easier to find out where is "Nucleoli (fibliar center)". 160018302 - good example where it would be helpful to find out are there any holes under 'green' dots or not.


Also other feature requests:
- ability to zoom into examples. It's would be good to see details and there
- ability to filter colors in examples
- also it would be good to have examples of "misleading" pictures. For example, when something looks like "nucleus" but it's not "nucleus" and why not.
- ability to learn from errors - open images from my history, and look what I answered, and what is correct answer. Especially for pending submissions that pass verification.
Nana Skalski
Taisaanat Kotei
EDENCOM DEFENSIVE INITIATIVE
#27 - 2016-02-27 09:18:00 UTC  |  Edited by: Nana Skalski
This small image inside image is not really needed for me, in fact it is completely obstructing part of image and doesnt really help in orienting where I am. I see it more distractng than helping.

I know these images have fixed resolution, so zoom is fixed.

Keep it simple. leave zoom lock, get rid of the small image.

And the category samples, they are staying in place, showing the image if I get cursor on them. I would want to change category rather than move my cursor around category sample image. Now i have to be carefull not to move cursor on the category sample image. It is slightly annoying.

ᴇᴠᴇʀʏ ᴘᴀʀᴛ ᴏғ ᴀ ɢᴀᴍᴇ ʜᴇʟᴘs ᴛᴏ ᴛᴇʟʟ ᴀ sᴛᴏʀʏ 📕

ᴡʜᴇʀᴇ ɪs ᴀɴɢʀʏ ᴄᴏɴᴄᴏʀᴅ ɢᴜʏ ᴡʜᴇɴ ʏᴏᴜ ɴᴇᴇᴅ ʜɪᴍ

ᴏsᴘʀᴇʏ 🚀

≡⋁≡ ♥
Akkerman Shalyga
Soviet W-Space Warriors
#28 - 2016-02-29 13:59:01 UTC
There are a lot of confusing samples without any hint. For example, here (ex. 160000370):

https://dl.dropboxusercontent.com/u/4262011/wrong.png

I can't see green thin strands. I can't see any yellow point. Maybe we need some more hints about how to detect such features?
Yadaryon Vondawn
Vicanthya
#29 - 2016-02-29 14:28:20 UTC
Akkerman Shalyga wrote:
There are a lot of confusing samples without any hint. For example, here (ex. 160000370):

https://dl.dropboxusercontent.com/u/4262011/wrong.png

I can't see green thin strands. I can't see any yellow point. Maybe we need some more hints about how to detect such features?


The sample is not confusing, either the players or the system behind it is giving a false positive. There are a few of these examples where the 'good' option is just plain wrong.
Akkerman Shalyga
Soviet W-Space Warriors
#30 - 2016-02-29 14:43:38 UTC
Yadaryon Vondawn wrote:
Akkerman Shalyga wrote:
There are a lot of confusing samples without any hint. For example, here (ex. 160000370):

https://dl.dropboxusercontent.com/u/4262011/wrong.png

I can't see green thin strands. I can't see any yellow point. Maybe we need some more hints about how to detect such features?


The sample is not confusing, either the players or the system behind it is giving a false positive. There are a few of these examples where the 'good' option is just plain wrong.

Not a few really. Seems like every 5-10 samples I can see a "plain wrong". And it can be easily achieved if system accepts most popular answer as a right answer. Because a lot of people just click like "ok, it seems good, no time to think, I need my money and AK" =(
CCP Wonderboy
C C P
C C P Alliance
#31 - 2016-03-01 11:57:57 UTC
Akkerman Shalyga wrote:
Yadaryon Vondawn wrote:
Akkerman Shalyga wrote:
There are a lot of confusing samples without any hint. For example, here (ex. 160000370):

https://dl.dropboxusercontent.com/u/4262011/wrong.png

I can't see green thin strands. I can't see any yellow point. Maybe we need some more hints about how to detect such features?


The sample is not confusing, either the players or the system behind it is giving a false positive. There are a few of these examples where the 'good' option is just plain wrong.

Not a few really. Seems like every 5-10 samples I can see a "plain wrong". And it can be easily achieved if system accepts most popular answer as a right answer. Because a lot of people just click like "ok, it seems good, no time to think, I need my money and AK" =(



We are aware of these false positive samples. They are a result of a minor error in the long production line from the scientists microscope to Project Discovery.

Human error is mostly at fault here, we are already working on it and we believe this will be fixed for the TQ release.
Terrah Chain
Space Isolation
#32 - 2016-03-01 17:10:54 UTC
CCP Paradox wrote:
Known Issues
Soon to come

  • Multiple example images per classification.
    • Many asked for this also, as sometimes one example image is sometimes not clear enough. We are working on this now.

  • Flight Academy video.
    • To better explain Project Discovery, your role and help aid you through the tutorial we will be creating a Flight Academy video that will be available to watch in client and on YouTube.


    Thank you for trying out Project Discovery,
    -Team Astro Sparkle

    One area I think could use some clarification is when to choose more than one classification. I did some analysis yesterday and on one slide is seems the consensus was tied for 2 classifications, nucleus and cytoplasm. But on another similar slide only nucleus was the consensus.

    When there are two shades of green are we focusing on the bright green only or all green areas? When dealing with classification outside the nucleus, is it possible that there is green below the nulceus? If so how can you let players know when to ignore it?

    On one slide with the bright dots outside the nucleus the text said the dots would never be in the nucleus, since dots appeared to be in the nucleus also, I chose a different classification but got it wrong.

    So, my point is that if you want accurate results the training may need to be a little bit better.

    Otherwise I like it and had fun with it. Not sure the rewards would make it an item of interest for me, but definitely worth doing if I am passing the time waiting for something else.
    CCP Wonderboy
    C C P
    C C P Alliance
    #33 - 2016-03-01 17:46:08 UTC
    Terrah Chain wrote:
    CCP Paradox wrote:
    Known Issues
    Soon to come

  • Multiple example images per classification.
    • Many asked for this also, as sometimes one example image is sometimes not clear enough. We are working on this now.

  • Flight Academy video.
    • To better explain Project Discovery, your role and help aid you through the tutorial we will be creating a Flight Academy video that will be available to watch in client and on YouTube.


    Thank you for trying out Project Discovery,
    -Team Astro Sparkle

    One area I think could use some clarification is when to choose more than one classification. I did some analysis yesterday and on one slide is seems the consensus was tied for 2 classifications, nucleus and cytoplasm. But on another similar slide only nucleus was the consensus.

    When there are two shades of green are we focusing on the bright green only or all green areas? When dealing with classification outside the nucleus, is it possible that there is green below the nulceus? If so how can you let players know when to ignore it?

    On one slide with the bright dots outside the nucleus the text said the dots would never be in the nucleus, since dots appeared to be in the nucleus also, I chose a different classification but got it wrong.

    So, my point is that if you want accurate results the training may need to be a little bit better.

    Otherwise I like it and had fun with it. Not sure the rewards would make it an item of interest for me, but definitely worth doing if I am passing the time waiting for something else.



    This is a pretty hard thing to convey in the tutorial, but i'll try my best.

    Remember that what you are looking at is actually a 3D structure represented in 2D, the nucleus is the doughey center of the cell, and the cytoplasm surrounds it.

    That means that there can be protein in the cytoplasm beneath the nucleus, giving the appearance that it is in the nucleus, when it actually is not.

    You can sort of tell when that is happening if there are green spots in the nucleus that are blurry and out of focus, that means they are either beneath or above the nucleus, which mean they are located in the cytoplasm.

    Another thing that is good to know is that you are not supposed to classify every little faint smudge of green. If you see green staining that is way weaker in brightness and contrast compared to other green staining in the image, you usually ignore it, but only if it is really weak.

    We are interested in the strongest proteins present in the cells, so classify *everything* you see, but leave out the super weak green staining.

    I hope that clears things up a little bit, we are of course always aiming to improve Project Discovery and we will try our best in training people to do this accurately.
    Terrah Chain
    Space Isolation
    #34 - 2016-03-01 18:20:46 UTC
    CCP Wonderboy wrote:
    Terrah Chain wrote:
    CCP Paradox wrote:
    Known Issues
    Soon to come

  • Multiple example images per classification.
    • Many asked for this also, as sometimes one example image is sometimes not clear enough. We are working on this now.

  • Flight Academy video.
    • To better explain Project Discovery, your role and help aid you through the tutorial we will be creating a Flight Academy video that will be available to watch in client and on YouTube.


    Thank you for trying out Project Discovery,
    -Team Astro Sparkle

    One area I think could use some clarification is when to choose more than one classification. I did some analysis yesterday and on one slide is seems the consensus was tied for 2 classifications, nucleus and cytoplasm. But on another similar slide only nucleus was the consensus.

    When there are two shades of green are we focusing on the bright green only or all green areas? When dealing with classification outside the nucleus, is it possible that there is green below the nulceus? If so how can you let players know when to ignore it?

    On one slide with the bright dots outside the nucleus the text said the dots would never be in the nucleus, since dots appeared to be in the nucleus also, I chose a different classification but got it wrong.

    So, my point is that if you want accurate results the training may need to be a little bit better.

    Otherwise I like it and had fun with it. Not sure the rewards would make it an item of interest for me, but definitely worth doing if I am passing the time waiting for something else.



    This is a pretty hard thing to convey in the tutorial, but i'll try my best.

    Remember that what you are looking at is actually a 3D structure represented in 2D, the nucleus is the doughey center of the cell, and the cytoplasm surrounds it.

    That means that there can be protein in the cytoplasm beneath the nucleus, giving the appearance that it is in the nucleus, when it actually is not.

    You can sort of tell when that is happening if there are green spots in the nucleus that are blurry and out of focus, that means they are either beneath or above the nucleus, which mean they are located in the cytoplasm.

    Another thing that is good to know is that you are not supposed to classify every little faint smudge of green. If you see green staining that is way weaker in brightness and contrast compared to other green staining in the image, you usually ignore it, but only if it is really weak.

    We are interested in the strongest proteins present in the cells, so classify *everything* you see, but leave out the super weak green staining.

    I hope that clears things up a little bit, we are of course always aiming to improve Project Discovery and we will try our best in training people to do this accurately.


    Yes, thank you. I knew some of what you said, but trying to figure it out and also seeing what others are choosing as classifications makes me think some addtional tips and pointers can be helpful. Trying to accurately classify what the project is looking for is obviously key and I wanted to convey my experience so that you could clarify and help others understand with my first time user point of view.

    Also the better equipped the user to analyze with the proper point of view and understanding will make the project more successful. So thank you for the time and I hope you can find ways to convey the areas I mentioned to new analyst.
    Yadaryon Vondawn
    Vicanthya
    #35 - 2016-03-02 10:06:49 UTC
    I noticed this morning that Project Discovery has a checkable box now that says 'Abnormal Sample'. What does that do? Oh and thanks for the way we can now select colours! It is great.

    And at the moment samples do not load :|
    CCP Wonderboy
    C C P
    C C P Alliance
    #36 - 2016-03-02 10:40:37 UTC
    Yadaryon Vondawn wrote:
    I noticed this morning that Project Discovery has a checkable box now that says 'Abnormal Sample'. What does that do? Oh and thanks for the way we can now select colours! It is great.

    And at the moment samples do not load :|



    You check the 'Abnormal Sample' box when you see something in the image that can not be classified in any of the available categories.

    So when you see something strange or out of place, check the box and the scientists at the HPA will review it.

    I too like the new color filter ^-^

    About the samples not loading, we are updating the MMOS API today and at the latest you should be getting samples tomorrow.

    Thank you :)
    Yadaryon Vondawn
    Vicanthya
    #37 - 2016-03-02 10:52:58 UTC
    CCP Wonderboy wrote:
    Yadaryon Vondawn wrote:
    I noticed this morning that Project Discovery has a checkable box now that says 'Abnormal Sample'. What does that do? Oh and thanks for the way we can now select colours! It is great.

    And at the moment samples do not load :|



    You check the 'Abnormal Sample' box when you see something in the image that can not be classified in any of the available categories.

    So when you see something strange or out of place, check the box and the scientists at the HPA will review it.

    I too like the new color filter ^-^

    About the samples not loading, we are updating the MMOS API today and at the latest you should be getting samples tomorrow.

    Thank you :)


    Ah I see, am I correct in guessing that using the Abnormal Sample will eventually lead to additional classification categories? During one of the talks Lundberg I believe said that the HPA has 20 classifications currently, one of the aims of the project is to extend that number right?

    From a gameplay point of view: Is it going to be part of the tutorial?

    And looking forward to the renewed Project Discovery tomorrow :)
    CCP Wonderboy
    C C P
    C C P Alliance
    #38 - 2016-03-02 11:03:06 UTC
    Yadaryon Vondawn wrote:
    CCP Wonderboy wrote:
    Yadaryon Vondawn wrote:
    I noticed this morning that Project Discovery has a checkable box now that says 'Abnormal Sample'. What does that do? Oh and thanks for the way we can now select colours! It is great.

    And at the moment samples do not load :|



    You check the 'Abnormal Sample' box when you see something in the image that can not be classified in any of the available categories.

    So when you see something strange or out of place, check the box and the scientists at the HPA will review it.

    I too like the new color filter ^-^

    About the samples not loading, we are updating the MMOS API today and at the latest you should be getting samples tomorrow.

    Thank you :)


    Ah I see, am I correct in guessing that using the Abnormal Sample will eventually lead to additional classification categories? During one of the talks Lundberg I believe said that the HPA has 20 classifications currently, one of the aims of the project is to extend that number right?

    From a gameplay point of view: Is it going to be part of the tutorial?

    And looking forward to the renewed Project Discovery tomorrow :)


    It could lead to additional categories, yes. And that would mean EVE players had made a real scientific discovery!

    However unlikely that is, it could happen :D
    Yadaryon Vondawn
    Vicanthya
    #39 - 2016-03-03 15:57:02 UTC  |  Edited by: Yadaryon Vondawn
    A new patch, a new feedback post :D


    • Wallet still blinks, I beg of you, please remove it.
    • Project Discovery is now taxed!? Why is this a thing? You get 50k~ per sample and you lose (in a NPC corp) 10% of that! I do not think this should be taxed
    • CCP Wonderboy wrote:
      • We are aware of these false positive samples. They are a result of a minor error in the long production line from the scientists microscope to Project Discovery.

      Human error is mostly at fault here, we are already working on it and we believe this will be fixed for the TQ release.

    I still see this happening, people identifying fibrillar centers as regular nucleoli's or nucleus as nucleoplasm. Based on the sample images and the sample images from the HPA these classifications are just plain wrong, yet it penalizes my accuracy.
  • Level progression was changed to non linear. So because the level stuff did not reset but only the accuracy I happen to be able to see how much I need to get to rank 105, *insert drumroll* 135.494 Experience points. Yep, 135K(!) that means that with an 75% accuracy I need to classify over 1800 to level up. In the meantime I will not get any AK since this is only given out when ranking up. That is insane. 1800 samples take a good 6~ hours plus, if you rush it. Since I wont be able to level up now I cannot see if the reward has been changed accordingly but can we get some clarification on this? This seems extremely offputting to me. Before this change I got 5K+ AK at level 100+. If after this change this will remain the same I would quit Project Discovery right away because from a gameplay point of view it is not viable at all. I feel this is the most important thing, I cannot stress this enough, without incentive the entire real life science tie in goes down the drain because players simply wont do it.
  • Abnormal sample is not part of the tutorial (not explained and not visible), maybe also include a sample image that is none of the options and has to be marked with Abnormal sample?
    • Oktura Ostus
    THE CORP WITH NO NAME
    #40 - 2016-03-03 16:59:21 UTC  |  Edited by: Oktura Ostus
    CCP Wonderboy wrote:

    I too like the new color filter ^-^


    Thanks for new filter! It now possible to view "holes" in blue cells. Small feature request to the filter: LMB on red color makes red visible, blue and green hidden. Is it possible that RMB on any color do reverse: hide it, and show all others?


    I've played a bit and find out this sample: http://i.imgur.com/MTQWznE.png 1600000854 As for me it's clear that there is "cytoskeleton" (marked in red circle), it's exactly the same as in examples, but it's marked as negative result. May be it's calculated from most often results, but such things disappoints a bit.

    May be it's possible to give ability to ask for recheck wrong results for players with high accuracy? With comment why to recheck.