Pages: 1 2 3 4 [5] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 7 post(s) |
Duke Killem
Center for Advanced Studies Gallente Federation
47
|
Posted - 2016.03.12 08:13:29 -
[121] - Quote
HPA Illuminator wrote:
Makes sense, and I'll fwd the input. Thanks!
You've been really great on the forum thank you.
To everyone else there's an in game chat channel called 'Project Discovery' with folks in that seem to know a thing or two. A few of them have over 90% accuracy all at varied levels. |
Horiz Rin
School of Applied Knowledge Caldari State
16
|
Posted - 2016.03.12 14:25:32 -
[122] - Quote
I just wanted to give my thoughts regarding the opinions that the results will be garbage and its just a click fest to get isk.
If your accuracy is below 50%, you get an image where the answer is already known. Get that wrong, you stay below 50%, and you get another known image. Obviously no scientists are going to bother looking at those. Button mash, and your accuracy is going to drop pretty fast, so not only will no one see your results, you'll also end up not getting any rewards.
Then there's the 'match the general consensus' players. I'm fairly sure that those images with a high consensus will be looked at by an experienced scientist. If he agrees, then yay, everyones accuracy goes up, if he disagrees, everyones score goes down. Those players just hitting the 'consensus' answers are, again, going to drop below 50% - it just takes longer.
Eventually - with the thousands of samples submitted every day - the images that 'rise to the top' are going to be those that have both a high consensus, and have been submitted by accurate participants. Win-win for the scientists :)
Also, if you're at 50% accuracy (an 'average' participant) you get 50,000isk per submitted image You have to submit 20 images getting at least half of those correct (to stay 'average') to earn 1m isk ......... no ones going to get rich on that. |
Owen Levanth
Sagittarius Unlimited Exploration
447
|
Posted - 2016.03.12 22:45:11 -
[123] - Quote
Horiz Rin wrote:I just wanted to give my thoughts regarding the opinions that the results will be garbage and its just a click fest to get isk.
If your accuracy is below 50%, you get an image where the answer is already known. Get that wrong, you stay below 50%, and you get another known image. Obviously no scientists are going to bother looking at those. Button mash, and your accuracy is going to drop pretty fast, so not only will no one see your results, you'll also end up not getting any rewards.
Then there's the 'match the general consensus' players. I'm fairly sure that those images with a high consensus will be looked at by an experienced scientist. If he agrees, then yay, everyones accuracy goes up, if he disagrees, everyones score goes down. Those players just hitting the 'consensus' answers are, again, going to drop below 50% - it just takes longer.
Eventually - with the thousands of samples submitted every day - the images that 'rise to the top' are going to be those that have both a high consensus, and have been submitted by accurate participants. Win-win for the scientists :)
Also, if you're at 50% accuracy (an 'average' participant) you get 50,000isk per submitted image You have to submit 20 images getting at least half of those correct (to stay 'average') to earn 1m isk ......... no ones going to get rich on that.
Ah, so that's how that works. Would be nice if that would be explained in-game. I was wondering why I never saw one of those "consensus"-images before. (Constantly below 50% since the beginning.)
Also, I'm slowly getting better, so right now I'm hovering at 38-39% instead of dropping further.
Sometimes I still get confused by pictures where I'm reasonably sure there's something wrong with the "correct" answer, so I've started to make some screenshots. I'll probably upload them tomorrow and ask some questions about this. |
Sheeth Athonille
Rabid Dogz Mining The Bastion
77
|
Posted - 2016.03.14 01:49:11 -
[124] - Quote
Has anyone else noticed that sometimes when you choose a selection and go to the screen that shows what the consensus has been, it'll change what you chose? I noticed this a few times and at first thought I'd screwed up and just miss-clicked, but I've since confirmed that it has happened. It seems to happen randomly so I don't have any before and after photos to prove it unfortunately, I may try to take some random screen pics to see if I can catch it. Basically the green check mark shows up in places other than what I'd selected, and not the selections I've made were checked.
Just wondering if anyone else has noticed this as well? |
Nevyn Auscent
Broke Sauce
3059
|
Posted - 2016.03.14 02:13:20 -
[125] - Quote
Owen wrote:Ah, so that's how that works. Would be nice if that would be explained in-game. I was wondering why I never saw one of those "consensus"-images before. (Constantly below 50% since the beginning.) Also, I'm slowly getting better, so right now I'm hovering at 38-39% instead of dropping further. Sometimes I still get confused by pictures where I'm reasonably sure there's something wrong with the "correct" answer, so I've started to make some screenshots. I'll probably upload them tomorrow and ask some questions about this. That's not how it works as far as most of us have worked out. Consensus is sometimes done as soon as you press your submission. It's about reaching a certain number of submissions with a certain consistency. They then review the image once it's reached a second threshold. Or from what they said when an entire set of data has reached that second threshold.
However join the "project discovery" chat room where we share difficult images including with the hpa staff playing the game themselves. |
Elmund Egivand
Federal Defense Union Gallente Federation
1055
|
Posted - 2016.03.14 03:06:10 -
[126] - Quote
My only complaint with Project Discovery is that the program window stops loading pictures every two or three submissions.
Other than that, I rather enjoy seeing exactly how much of my University education I still remember.
A Minmatar warship is like a rusting Beetle with 500 horsepower Cardillac engines in the rear, armour plating bolted to chassis and a M2 Browning stuck on top.
|
|
HPA Illuminator
State War Academy Caldari State
25
|
Posted - 2016.03.14 07:32:02 -
[127] - Quote
Nevyn Auscent wrote:Owen wrote:Ah, so that's how that works. Would be nice if that would be explained in-game. I was wondering why I never saw one of those "consensus"-images before. (Constantly below 50% since the beginning.) Also, I'm slowly getting better, so right now I'm hovering at 38-39% instead of dropping further. Sometimes I still get confused by pictures where I'm reasonably sure there's something wrong with the "correct" answer, so I've started to make some screenshots. I'll probably upload them tomorrow and ask some questions about this. That's not how it works as far as most of us have worked out. Consensus is sometimes done as soon as you press your submission. It's about reaching a certain number of submissions with a certain consistency. They then review the image once it's reached a second threshold. Or from what they said when an entire set of data has reached that second threshold. However join the "project discovery" chat room where we share difficult images including with the hpa staff playing the game themselves.
more like the hpa staff trolling the chat room (while taking notes of high ranking players ;))
But good explanation otherwise! ^^ |
|
HPA Illuminator
State War Academy Caldari State
25
|
Posted - 2016.03.14 07:33:25 -
[128] - Quote
Sheeth Athonille wrote:Has anyone else noticed that sometimes when you choose a selection and go to the screen that shows what the consensus has been, it'll change what you chose? I noticed this a few times and at first thought I'd screwed up and just miss-clicked, but I've since confirmed that it has happened. It seems to happen randomly so I don't have any before and after photos to prove it unfortunately, I may try to take some random screen pics to see if I can catch it. Basically the green check mark shows up in places other than what I'd selected, and not the selections I've made were checked.
Just wondering if anyone else has noticed this as well?
It does happen, and from what I've heard, it's due to too many ppl playing/more than servers can handle. I honestly don't know what is saved as your answer though. |
LigrayOu Kva
Deep Space Conquerors Soviet-Union
12
|
Posted - 2016.03.14 22:38:08 -
[129] - Quote
Stop it. Enough. button mashers won. http://i.imgur.com/rfxGahh.png |
beakerax
Pator Tech School
230
|
Posted - 2016.03.14 23:19:05 -
[130] - Quote
Up to 73.5% :)
Is it just me, or have the number of control slides increased? I had three out of something like 15 slides just now. That's not a big sample, but it was a little weird.
Sheeth Athonille wrote:Has anyone else noticed that sometimes when you choose a selection and go to the screen that shows what the consensus has been, it'll change what you chose? I noticed this a few times and at first thought I'd screwed up and just miss-clicked, but I've since confirmed that it has happened. It seems to show the selections you chose on the previous slide. My completely unfounded belief is that it's caused by the client timing out while it waits on the % results from PD. |
|
Krevnos
Back Door Burglars The Otherworld
84
|
Posted - 2016.03.15 10:19:00 -
[131] - Quote
beakerax wrote:Up to 73.5% :) Is it just me, or have the number of control slides increased? I had three out of something like 15 slides just now. That's not a big sample, but it was a little weird. Sheeth Athonille wrote:Has anyone else noticed that sometimes when you choose a selection and go to the screen that shows what the consensus has been, it'll change what you chose? I noticed this a few times and at first thought I'd screwed up and just miss-clicked, but I've since confirmed that it has happened. It seems to show the selections you chose on the previous slide. My completely unfounded belief is that it's caused by the client timing out while it waits on the % results from PD.
Yes, this is exactly what is happening. We can't be sure if it affects the submission though.
We do need to get the server upgraded or a caching system in place. I suppose the HPA guys didn't expect the level of participation they received. Not a bad thing I guess Gÿ¦n+Å
|
Nevyn Auscent
Broke Sauce
3065
|
Posted - 2016.03.15 10:58:20 -
[132] - Quote
CCP have said it submits whatever you had actually chosen, and that seems mostly consistent at least with practice. And yes they changed the slide ratio |
Ria Nieyli
41950
|
Posted - 2016.03.15 11:32:59 -
[133] - Quote
It seems that the number of people doing this has dropped off sharply, I suddenly piled up several hundred unsolved cases. They were chugging along rather fast two days ago.
Vote Cytoplasm for CSM 11!
|
HPA Illuminator
State War Academy Caldari State
26
|
Posted - 2016.03.15 11:40:08 -
[134] - Quote
Krevnos wrote:beakerax wrote:Up to 73.5% :) Is it just me, or have the number of control slides increased? I had three out of something like 15 slides just now. That's not a big sample, but it was a little weird. Sheeth Athonille wrote:Has anyone else noticed that sometimes when you choose a selection and go to the screen that shows what the consensus has been, it'll change what you chose? I noticed this a few times and at first thought I'd screwed up and just miss-clicked, but I've since confirmed that it has happened. It seems to show the selections you chose on the previous slide. My completely unfounded belief is that it's caused by the client timing out while it waits on the % results from PD. Yes, this is exactly what is happening. We can't be sure if it affects the submission though. We do need to get the server upgraded or a caching system in place. I suppose the HPA guys didn't expect the level of participation they received. Not a bad thing I guess Gÿ¦n+Å
We're doing happy dance in the lab about the AMAZING number of people participating (of which the majority are doing their best) :) |
Krevnos
Back Door Burglars The Otherworld
84
|
Posted - 2016.03.15 15:54:02 -
[135] - Quote
HPA Illuminator wrote:
We're doing happy dance in the lab about the AMAZING number of people participating (of which the majority are doing their best) :)
We love them! They provide a very unique challenge since there is no algorithm or set answer. Our accuracy will improve over the coming weeks as be become more familiar with them. I'm glad to see more control slides popping up - they really help to hone our eyes.
I hope that we do you proud! |
TigerXtrm
KarmaFleet Goonswarm Federation
1406
|
Posted - 2016.03.15 16:58:01 -
[136] - Quote
Demica Diaz wrote:I had feeling that something like this will happen.
Wait, you actually get rewarded for getting it wrong? Please tell me no, because that would be insanely ********.
My YouTube Channel - EVE Tutorials & other game related things!
My Website - Blogs, Livestreams & Forums
|
Sp3ktr3
Caldari Provisions Caldari State
4
|
Posted - 2016.03.15 17:23:32 -
[137] - Quote
HPA Illuminator wrote: We're doing happy dance in the lab about the AMAZING number of people participating (of which the majority are doing their best) :)
I'm glad you're seeing such a positive response. Is there any way to show off the badge that we earn for ranking up in the project? |
Owen Levanth
Sagittarius Unlimited Exploration
449
|
Posted - 2016.03.15 18:50:11 -
[138] - Quote
Nevyn Auscent wrote:Owen wrote:Ah, so that's how that works. Would be nice if that would be explained in-game. I was wondering why I never saw one of those "consensus"-images before. (Constantly below 50% since the beginning.) Also, I'm slowly getting better, so right now I'm hovering at 38-39% instead of dropping further. Sometimes I still get confused by pictures where I'm reasonably sure there's something wrong with the "correct" answer, so I've started to make some screenshots. I'll probably upload them tomorrow and ask some questions about this. That's not how it works as far as most of us have worked out. Consensus is sometimes done as soon as you press your submission. It's about reaching a certain number of submissions with a certain consistency. They then review the image once it's reached a second threshold. Or from what they said when an entire set of data has reached that second threshold. However join the "project discovery" chat room where we share difficult images including with the hpa staff playing the game themselves.
I never had one of those "Consensus"-pictures, so I guess constant failure isn't the kind of consistency the project wants.
But I'm OK with this now, my initial excitement has cooled down and I've moved on to other things. Right now I'm willing to do 1-2 cell probes every time I log into my industry-alt, but only because the mini-game is still fun, even without compensation. |
PAPULA
Black Aces I N F A M O U S
84
|
Posted - 2016.03.15 19:20:20 -
[139] - Quote
Today it's very bad, pictures don't load at all, sometimes i get a message it's not available.
|
HPA Illuminator
State War Academy Caldari State
27
|
Posted - 2016.03.15 19:54:25 -
[140] - Quote
Mark as abnormal and we'll take a look at it and remove it. It's not that it's not loading, the image is broken (someone stopped the microscope too soon). |
|
Skyweir Kinnison
Coldhammer Harvest LLC
208
|
Posted - 2016.03.15 20:50:28 -
[141] - Quote
Those people who have got an armour suit as a reward must have the patience of Job. I've given over about an hour a day since the launch and have only made it to 25k points! I'm going to be continuing, but much less time as I'm getting bored and frustrated. It's a good cause, but reminds me why I left a science career to start a business
The increased number of control slides was much appreciated as useful feedback. I was pleased to see I got almost all right, which encouraged me at times of frustration.
I've given up on any rewards though, I'll just do it to support the project.
Humanity has won its battle. Liberty now has a country.
|
Sp3ktr3
Caldari Provisions Caldari State
4
|
Posted - 2016.03.15 20:54:54 -
[142] - Quote
Skyweir Kinnison wrote:Those people who have got an armour suit as a reward must have the patience of Job. I've given over about an hour a day since the launch and have only made it to 25k points! I'm going to be continuing, but much less time as I'm getting bored and frustrated. It's a good cause, but reminds me why I left a science career to start a business The increased number of control slides was much appreciated as useful feedback. I was pleased to see I got almost all right, which encouraged me at times of frustration. I've given up on any rewards though, I'll just do it to support the project.
The awarded credits increase with each rank you gain. Also I'm sitting at around 88% accuracy, which also proportionally increases the number of experience points you gain per submission. |
beakerax
Pator Tech School
230
|
Posted - 2016.03.15 22:36:34 -
[143] - Quote
90.5%!
Skyweir Kinnison wrote:I've given up on any rewards though, I'll just do it to support the project. The AK reward is 100 * accuracy * rank, so hang in there :) |
Ismael Caleb Echart
Signal Cartel EvE-Scout Enclave
4
|
Posted - 2016.03.16 21:53:29 -
[144] - Quote
While this project is a worthy idea, personally I'm no longer going to take part in it.
Most of the submissions are confusing and lots of features appear similar in them. So much so that you may get close, but not close enough with your final submissions making it difficult to have a high accuracy any really worthwhile reward. I've been using an alt to take part.
Nor is it a matter of a few minutes per submission, more like between 5 and 10 minutes to try and get it right and as accurate as possible.
Sorry, but we are paying a plex for thirty days game time, and I can earn more in 15 minutes mining than I'm ever going to make in rewards with this analysis project. I'm wasting no more effort on trying to improve with it. It's a waste of my time against other earning in the game.
As I say great idea, but thumbs down long term from me. |
Sheeth Athonille
Rabid Dogz Mining The Bastion
78
|
Posted - 2016.03.17 13:33:58 -
[145] - Quote
The reward can't be all that great or else people would farm them to hell and back. If you do enough of them you begin to differentiate what's what easier and can do them much faster. Depending on the slide it's usually 30 seconds to a min for me. If the slides where easier to pick out what's what they'd be able to use a computer program and wouldn't need human eyes to do them :) |
HPA Dichroic
Polaris Corporation
22
|
Posted - 2016.05.11 09:38:05 -
[146] - Quote
This has now been officially updated in-game. We now have a system, so updates should happen faster. Thanks for the help and keep em coming! |
|
|
|
Pages: 1 2 3 4 [5] :: one page |
First page | Previous page | Next page | Last page |