These forums have been archived and are now read-only.

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

Science & Industry

 
  • Topic is locked indefinitely.
 

The state of Project Discovery

First post
Author
#1 - 2016-07-07 12:21:09 UTC
Hey folks,

Today I want to discuss the state of Project Discovery. As you all know the project had a few goals:

  • Integrate into an existing game to ensure a steady amount of players
  • Help science go forwards by analysing way faster than a team of scientists could while still upholding acceptable accuracy


Today many players still actively play Project Discovery but as with many such projects, even this integrated one, we find that a small dedicated core of experts make up the bulk. Ofcourse when Project Discovery was just new everyone wanted to try it out. There were some launch bugs but in the end a fair amount of people tried Project Discovery.

Now that we are a few months further down the line I wanted to take the time and analyse the available data. This data shows a rather depressing trend.

Data ( from the monthly economics report):

This results in the following graphs:
ISK per month
Classifications per month

We can see a very clear trend here. And it is not a good trend for science. The question is, how can it be fixed?

Problems
The initial problem with Project Discovery, as many can attest, was that it was often too difficult for people to understand. With HPA_Illuminator's classes and various people helping each other in the ingame channels, here and on youtube I feel that that problem is largely tackled.

This brings us to incentives. Currently ISK payout is tied to accuracy and AK payout is tied to accuracy and rank. Because there is a near hardcap of samples per rank (50~), the AK payout increases a lot at higher levels. A person at rank 250@99% accuracy gets around 25K AK per 50~ samples. Whereas a person with rank 80@99% gets only 8K AK for only slightly fewer samples.

Besides this disparity there is the problem of the AK rewards. Farming AK (especially at lower ranks) is simply not really worth it for most players. One would need to get close to rank 80~ before one has enough AK to buy for example a suit. The suits go for less than 100M at the moment, and they are not even on the market. Couple this with the boosters sell price of 3.1-4M a piece and one may wonder why one would actually do Project Discovery, as a normal player seeking ISK.

There is also the problem of faulty samples, this is largely tackled by HPA's excellent response to reported samples. The report proces however requires a reddit account and a short list consisting of sample number, screendump and a comment. Because there are quite a few of the control samples faulty this sometimes discourages players from reporting it.

Possible fixes
The following possible fixes could, in my eyes, help Project Discovery to gain the amount of players it deserves. The more people help, the more we can help science after all!

Therefor I suggest the following changes:

  • More rewards that have a clear market (think temporary skins, better boosters and actual modules)
  • Even out the AK payout - samples per level at higher levels to something more reasonable
  • Allow players with a accuracy of 99% to mark a control image as faulty, this would turn up a new screen where one can set the correct classifications. If 5 players with 99% mark the same image with the same new classifications, alter the control image to that new standard and notify the HPA


With this piece of writing I have hoped to give a small overview of the current state of Project Discovery, the problems and the possible fixes. Do you have suggestions for CCP/HPA to make Project Discovery? Please discuss :)
C C P Alliance
#2 - 2016-07-08 08:12:36 UTC
Hey Yadaryon,

Thanks for continuing to be one of our top supporters! Big smile

I haven't seen that data about the ISK or classification counts. Where did you find it? It is indeed a bit sad to see the numbers declining, but they are still well above the expectations for a normal citizen science project, so keep it up!

With most of your suggestions, my hands are tied of course as I don't work at CCP, but let me take a minute to address the HPA related one.

As you say, there are some faulty control samples. I updated another 50-100 samples yesterday with corrections brought to us by players so that should help, but your point about the friction to getting these samples to us is well taken.

Your suggestion of players with high accuracy impacting the controls automatically is probably not something we would be comfortable with, but I think it's a good idea to look at the most disagreed with samples and modify/remove them from the pool now that we have sufficient data to do so. I will work on that a bit next week and see what it yields. One trick is to keep the really challenging ones but correct the ones that are incorrect, but I'll check it out and come up with something.

In the mean time, don't stop classifying!!! We continue to be impressed with everyone's support and you should be prepared to see EVE online in a scientific journal in the coming months. Blink

o7 Dichroic
#3 - 2016-07-08 12:31:25 UTC
HPA Dichroic wrote:
Hey Yadaryon,

Thanks for continuing to be one of our top supporters! Big smile

I haven't seen that data about the ISK or classification counts. Where did you find it? It is indeed a bit sad to see the numbers declining, but they are still well above the expectations for a normal citizen science project, so keep it up!

With most of your suggestions, my hands are tied of course as I don't work at CCP, but let me take a minute to address the HPA related one.

As you say, there are some faulty control samples. I updated another 50-100 samples yesterday with corrections brought to us by players so that should help, but your point about the friction to getting these samples to us is well taken.

Your suggestion of players with high accuracy impacting the controls automatically is probably not something we would be comfortable with, but I think it's a good idea to look at the most disagreed with samples and modify/remove them from the pool now that we have sufficient data to do so. I will work on that a bit next week and see what it yields. One trick is to keep the really challenging ones but correct the ones that are incorrect, but I'll check it out and come up with something.

In the mean time, don't stop classifying!!! We continue to be impressed with everyone's support and you should be prepared to see EVE online in a scientific journal in the coming months. Blink

o7 Dichroic




Hey Dichroic!

The data is pulled from the monthly economic report. I checked the ISK faucet data from that report and graphed it. Ofcourse since the payout varies between 30k ISK and 99k ISK based on ACC it is only a guess. But it does give an indication.

I hope indeed that CCP will continue to improve upon Project Discovery in terms of incentives, this post was meant to bring the decline to light. Maybe you guys can also give a heads up to CCP that it would be very much appreciated if there would be, at least visible, progress. The rank icons were updated but it wasn't even in the patchnotes :(

The suggestion about automatically altering control images was, as Marionette pointed out in hindsight not such a good idea. And I can imagine that we as players make mistakes that could cause problems. Maybe let that suggested system sent the 'new' classification to you guys for confirmation? That would avoid the hassle of using Reddit or direct messaging.

Naturally I won't stop Project Discovery! I am still doing it while waiting for a fleet or while travelling ;)

Can't wait for the journal tho!
#4 - 2016-07-11 09:40:20 UTC
Well, I re-started playing when the project was new and I am low-skilled so it was a good way to make ISK but since the price of the AK dropped it was mostly useless, even if i'm really fast and always at 99% accuracy, I even helped people to understand how it worked and get their % higher. The problem is that most of those items aren't really usefull and once you bought a skin it's done you have it and don't need an other copy... So I agree with your idea of new rewards, I definetly won't do something that pay less than my other activities.

Sorry for my bad english

Fly safe, Chakmal
Amarr Empire
#5 - 2016-07-16 16:06:44 UTC  |  Edited by: Ni Neith
Quote:
We can see a very clear trend here. And it is not a good trend for science. The question is, how can it be fixed?


It can't be fixed. Eve is entertainment. If you do want to do science, go do science (somewhere else where it belongs). I personally knew this is doomed and never klicked this bs.
#6 - 2016-07-30 04:23:38 UTC
Ni Neith wrote:
Quote:
We can see a very clear trend here. And it is not a good trend for science. The question is, how can it be fixed?


It can't be fixed. Eve is entertainment. If you do want to do science, go do science (somewhere else where it belongs). I personally knew this is doomed and never klicked this bs.



Well thanks for doing your part to contribute to the betterment of mankind, sunshine.
If you don't want to help out, then by all means carry on. But do you really need to be so unconstructive and disparaging to an activity that can potentially advance research on helping the sick?

I mean, it's a brand new feature and like everything introduced it will need it's fair share of tweaks, but try not to
discourage others from being more than self interest addicted gratification monkeys that are only here for the chicken wings,


mmmmmm........k?



Thank you O.P. for a concise & constructively critical post that can help the reform of a very very good idea.


Jack
Forum Jump