These forums have been archived and are now read-only.

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

Upcoming Feature and Change Feedback Center

 
  • Topic is locked indefinitely.
 

Project Discovery - Changes to the scoring system

First post First post
Author
Memphis Baas
#61 - 2016-03-26 01:37:20 UTC
Accuracy now seems to be similar to gaining standings, what you get depends on how close to 100 you are (with diminishing returns). Which is ok.

We're still doing this, though, which I found hilarious.

Wimzy Chent-Shi
Garoun Investment Bank
Gallente Federation
#62 - 2016-03-26 09:23:56 UTC  |  Edited by: Wimzy Chent-Shi
Well I am just sad I missed out on the cytoplasm freefarm that has been practiced before.
I feel somewhat demotivated since 90% of my input is essentially randomly irrelevant to my gain.
And even when I put in the effort and try to identify the correct thing, I get to find out that scientist or whoever responsible for the control sample had a different idea. I started humbly at 50% working towards 85% in the end.
I also observed that over 50% of the non-control sample qualify for Unspecific, but nobody checks those, because you risk losing the accuracy rating, cause control samples might look like many things, but well, they have the holy truth in them for some reason.

If I were to have a personal statistic of where had I most frequently gone wrong between two options, let's say Microtubule organizing center vs Centrosome (yey silly me), I would probably pick options based on statics instead of my observations.

Come get some cancer @ my blog !

"This clash of opinions is like cutting onions. We are creating something here, that's productive, ...and then there is also salt." -Wimzy 2016

Memphis Baas
#63 - 2016-03-26 13:16:52 UTC
Ideas:

1. We can write notes on other pilots' Show Info pages, to keep track of behavior. If you're not doing so already, CCP, start keeping track of the choices each character makes for Project Discovery, so you can easily differentiate between those who pick cytoplasm 99% of the time, and those who seem to pick whatever choices seem appropriate. You can build all sorts of statistics from this, including how "unusual" each choice is when it's made.


2. Project Rediscovery: the option to go through the samples again, but with the consensus visible, to agree or disagree with the consensus. For high accuracy players, or if you can somehow detect those of us who are actually trying to be accurate rather than to game the system.


3. The samples can be categorized:

- 1 or more choices at 100% - maybe these can be reviewed and converted into practice / accuracy slides, since we ALL seem to agree on the choices

- 1 or more choices at 50% - 90% - there's some sort of consensus

- 0 choices at 50%, but more than 8 choices at less than 20% - there's complete disagreement - something is confusing about the sample; maybe some of these can be reviewed and published on the website as contrast examples, or explanations on how to get the answer via elimination.

- not yet rated - you can keep a tally so you can see our progress.
Typhoon Maru
The Wyrd Sisters
#64 - 2016-03-26 13:35:38 UTC
I know it's a bit nit picky and asking for the moon on a stick , but it would be nice if we could examine the Option pictures using the magnifier and the colour filters. That said i'm enjoying this project immensely keep up the good work Smile
Nam Dnilb
Universal Frog
#65 - 2016-03-26 14:18:30 UTC
Typhoon Maru wrote:
I know it's a bit nit picky and asking for the moon on a stick , but it would be nice if we could examine the Option pictures using the magnifier and the colour filters. That said i'm enjoying this project immensely keep up the good work Smile


The above and unlock the image controls on the result screen please. Right now it a bit of "You, Sir, were wrong and you'll never know why!" P
NorthCrossroad
EVE University
Ivy League
#66 - 2016-03-26 17:40:28 UTC
Just have to say it - female Sister of EVE armor is HOT. :) Best "clothing" ever. Though would be great to have incentives after 75k AK. Like Sister of EVE ships skins at 100k+ for example.

North
Mars Aspen
Emergency Banana Straigtening Service
#67 - 2016-03-26 18:01:05 UTC
Raven Dallacort wrote:
I'm no scientist, but how am I wrong on this?


I would guess that since the blue is quite strong the underlying green is weak, so the strongest green signal is as indicated. Just imo though.
Wimzy Chent-Shi
Garoun Investment Bank
Gallente Federation
#68 - 2016-03-26 18:43:17 UTC  |  Edited by: Wimzy Chent-Shi
NorthCrossroad wrote:
Just have to say it - female Sister of EVE armor is HOT. :) Best "clothing" ever. Though would be great to have incentives after 75k AK. Like Sister of EVE ships skins at 100k+ for example.

North

This.
The armor now stands for "not dealing in project discovery again" :-D although, if not purchased over a contract somewhere it also stands for a decent standard of accuracy and skill which should/could be further motivated by further rewards.

Come get some cancer @ my blog !

"This clash of opinions is like cutting onions. We are creating something here, that's productive, ...and then there is also salt." -Wimzy 2016

Nevyn Auscent
Broke Sauce
#69 - 2016-03-26 18:47:32 UTC
Mars Aspen wrote:
Raven Dallacort wrote:
I'm no scientist, but how am I wrong on this?


I would guess that since the blue is quite strong the underlying green is weak, so the strongest green signal is as indicated. Just imo though.

The splotchy uneven sized nature of the green is a good sign of nuclear speckles rather than Nucleoli. Those two can be one of the hardest to tell apart.
Pycu
Caldari Provisions
Caldari State
#70 - 2016-03-27 08:51:38 UTC
CCP Wonderboy wrote:
. Instead, we will increase the rate of training examples you get, for which you get an immediate rating.


Not sure if it's a bug or not but I have observed a not so rare case where my choice is the only correct one, marked with green etc. but my rating is not increasing when I submit it.

If the Submission history was working I was going to be able to provide an example.

Any ideas?
The Golden Serpent
A Drunken Squirrels' Conspiracy for Revenge
#71 - 2016-03-27 10:11:17 UTC
I have been really impressed with the whole Project Discovery thing and if anything will make me resub this would be it, as it feels like I'm actually doing something real.

-:¦:-•:'":•.-:¦:-•* K H A N I D •-:¦:-•:''''*:•-:¦:-

Alexhandr Shkarov
Deepsafe Orphanage Ltd.
#72 - 2016-03-27 10:51:47 UTC
Feedback:

Since our accuracy rating is based on the pre-selected samples, could you please for the love of everything that you believe in add a box which explains WHY you want something specific if we have it wrong? I've had some images which could be two things but they claim I saw it wrong and needed to choose the other. It would be amazing if there was some explanation on why they see it that way, so that we can learn more about the process and be more accurate on other samples.

All my posts are on my personal title and should not be confused as me speaking for anyone else.

Memphis Baas
#73 - 2016-03-27 13:45:07 UTC
Give them a couple weeks and then switch to "no more Project Discovery until they fix it" mode. Nothing more damaging than a rewards system that's unfair or that rewards wrong choices.
NorthCrossroad
EVE University
Ivy League
#74 - 2016-03-28 17:36:05 UTC
Another small thing. Would be great, if hovering over the Project Discovery button on the Neocom would show the amount of AKs in your "wallet". Just like when you hover over the Wallet button.

North
Arsine Mayhem
Doomheim
#75 - 2016-03-28 18:31:31 UTC
I just got around to trying this out, went part way through the tutorial. One thing.

How are we suppose to find the impairment in a cell when we have no good cell to reference.

Doesn't seem to realistic. Am I missing something?
Black Snapper
Pator Tech School
Minmatar Republic
#76 - 2016-03-28 23:28:27 UTC
Why oh why is my submission history always empty. Im at level 17 and have been doing this for 3 days...
Memphis Baas
#77 - 2016-03-29 18:17:04 UTC
The submission history is empty for everyone; they removed that functionality. We used to get rated by the Project Discovery people and that's how our accuracy increased; now the system automatically rates your accuracy by giving you the tutorial slides all over again and if you guess them correctly your accuracy goes up, otherwise it goes down. So the submission history no longer matters.
SurrenderMonkey
The Exchange Collective
Solyaris Chtonium
#78 - 2016-03-30 14:49:04 UTC
This is a good change, however, I feel like the training samples need to be harder. A lot harder.

I can actually predict, with pretty close to perfect accuracy (exceptions are, e.g., microtubules, where the unknown samples are usually pretty obvious, too), whether or not a sample will be a training sample just by its appearance. They're pretty much always much "cleaner" and less ambiguous than the unknown samples.

While this has suppressed the motivation to click-cytoplasm-because-everyone-else-probably-did-too, I don't think a high accuracy score on the existing training samples is representative of ability on the unknowns.

"Help, I'm bored with missions!"

http://swiftandbitter.com/eve/wtd/

Drake Carver
Minmatar Deep Space Expeditionary Corp
#79 - 2016-03-30 16:04:20 UTC
TL:DR >>This post is one big sloppy thank you for this wonderful game. Skip to the end for the actual question.



From the beginning I have supported project discovery for its purpose and a fascination with the microscopic world.

From my experience so far I find it disheartening that the system can and will be gamed however I am certain methods will arise and adapt to optimize for the best results for the hardworking researchers that will actually make use of the results to refine our knowledge to be less of a guessing game in whatever we MMO player assist in.

Now I may be sorely mistaken in the things I am about to suggest but there are some functions that I feel would at least allow serious project discovery players to get a little more satisfaction from a job well done.

I often find that sometimes I have made an obvious mistake only after checking all the marks.

If there could be a way to report that I am truly mistaken in my result so that I cannot knowingly skew the results I would be thankful. I would still gladly accept the consequences of reductions in accuracy or the nullification of any gains I would have stood to earn. Letting obviously bad results through does not feel right and if I could make reparations I would.



TL:DR >>Can there be a button for reporting that a mistake has been made but it was realized only after submitting.
Terminal Insanity
KarmaFleet
Goonswarm Federation
#80 - 2016-03-30 20:45:09 UTC
This is a really neat addition to the game.

I wonder if you could expand this to include other science projects as well? a massive real science lab in eve!

"War declarations are never officially considered griefing and are not a bannable offense, and it has been repeatedly stated by the developers that the possibility for non-consensual PvP is an intended feature." - CCP