Why I Vote "No" On Adding Projects to the Gridcoin Whitelist If they Do Not have Sufficient Work-Units

in #gridcoin7 years ago (edited)

gridcoin.png
A Reddit user was kind enough to do a bit of research on some of the projects that are currently proposed to be added to the Gridcoin whitelist. I'll link them here:

The issue with these projects --as the Reddit user notes-- is that they all currently have some ambiguity surrounding whether or not they will have workunits available for an influx of Gridcoin users. (Please note that SourceFinder anticipates having many WUs available at some point in the near future due to receiving work for ASKAP as @peppernrino notes in this article).

Why does a project having Workunits available matter?

Gridcoin could be described as one big supercomputer, right? On those BOINC projects where volunteer computing is not sufficient to crunch the amount of data the project has, we can add the project to the Gridcoin whitelist and point thousands of individual computers' worth of crunching ability in their direction. When a project does not have enough workunits available, we are pointing our resources at a project that, quite frankly, doesn't need it.

Big deal, right? We'll just let Gridcoiners camp out on the project and get paid for whenever they are doing work. Unfortunately, that's not currently how the Gridcoin Proof of Research (PoR) reward mechanism works. No matter how many workunits a project has available and how hard their contributors have to work, it receives an equal distribution of daily PoR rewards. That means a project like NFS@Home, for example, where every cruncher is crunching 100%, 24-7 receives the exact same amount of GRC distributed among its Gridcoin team as Sztaki Desktop Grid, which has zero WUs available most days.

Another way of saying that: If you have a 200 Magnitude on Sztaki Desktop Grid and a 200 Magnitude on NFS@Home, you earn the exact same amount of daily Gridcoin, except your computers are busy crunching 24-7 on NFS@Home whereas you might go entire days with no workunits on Sztaki Desktop Grid.

If we allowed several projects with insufficient workunits into the Gridcoin whitelist, a savvy Gridcoiner could restrict their efforts to only those projects. In this way, they could earn a high magnitude and GRC while only intermittently utilizing their hardware and minimizing their costs.

Other than the issue of gaming the Gridcoin system in this way, it also represents a very inefficient use of such a great resource. If a project is able to sufficiently get by with volunteer contributions, Gridcoin is better off "pointing our supercomputer" toward projects that need our assistance. If we don't do that, we are not leveraging this outstanding scientific resource to the best of its abilities.

Thoughts? Let's hear them below.

Sort:  

This is a situation that I feel I need more information on to take a stance, so, unless I find a surfeit of time soon-ish, I will probably abstain also.

I'm unsure whether gaming the system is the best term, as it brings out negative connotations, but, it will do as long as we agree that there might be "first blush" reactions.

Personally, I have and continue to try and optimize my ROI, but even if I fall behind on that, I know that I am still helping, that is the beauty of the BOINC / Gridcoin partnership I think.

Excellent read with some great info.

TL;DR I will continue to crunch the projects which I support regardless of their immediate position on the Gridcoin white-list as the GRC dev team continues to improve the relationship, both physical and coding, between Gridcoin and BOINC.

I agree that availability of WU should be considered when deciding to whitelist/delist a poll.

Regarding your position:

There is fine line between a project which intentionally limits and distributes WU and one which simply does not have WU or the ability to maintain itself. For me, the project itself and its development and admin activity is a bigger factor than the WU structure of the project.

SZTAKI, for example, rewards users which have been working the project for an extended length of time. I think that is fine. It encourages new users to join the project for the project's stake instead of just for the cobblestone. Those who have been dedicated to the project do in fact get a lot of work for SZTAKI. I do not think their model is perfect, but I am not certain that it deserves removal from GRC whitelist.

I think the root issues regarding projects and white-listing and distribution lie with the DPoR protocol of gridcoin, so I will probably abstain from the SZTAKI vote. The dev team is working on the relationship between Gridcoin and BOINC (right now by working on superblocks), so I feel like in the next 6 months to a year this will be a non-issue.

In terms of gaming the system: I haven't thought this all the way through, so please, guide me, but if someone wanted to increase their mag on a project which limits WU, wouldn't they need to have been working that project before they limited WU?

Ultimately, dividing total GRC magnitude/daily coins evenly among each whitelisted project is the primary issue. If the PoR mechanism somehow was able to reward people purely based on how much work they were doing (IE: two identical computers maxing out their CPU/GPU at 100% for a full 24 hours earn the exact same amount of credits regardless of which projects they were working on respectively) we could simply whitelist projects based on how valuable they are to science and make this process so much simpler.

Regarding gaming of the GRC reward system, it would basically work like this: if we had 10 projects that only divided out workunits once per week for a short time, I could simply dump my current projects and sign myself up for all 10 of those projects. Since most of these projects simply put their workunits in a big bucket and crunchers take them "first come, first serve," I would effectively be in 10 different "lottery" systems. I might only be crunching 25% of the time, but there's a pretty good chance that I could achieve a magnitude equal to someone crunching 24-7 for SETI, PrimeGrid, or any other project that requires 24-7 crunching to compete.

In a situation like that, you wouldn't even necessarily need a good crunching computer. Load up a BOINC client on a dozen old laptops to maximize your chances of getting workunits. Set each client to accept as many WUs as possible, and you've basically built yourself a Gridcoin mining farm that could get 1,000+ Magnitude on a cumulative computing power less than a single GTX 1070 and maybe an average of 200 watts electricy.

Yes! Finally another person that thinks equally dividing the rewards among all projects is a bad idea! I think voting on which projects deserve the most computing power and having a magnitude multiplier based on that vote is the best idea I've heard so far. In that case, I'm guessing most of those unpopular projects would have very low multipliers, so gaming the system would be less profitable.

Unfortunately Investors have the most voting weight in Gridcoin, not the miners.

Congratulations @xaqfields! You have completed some achievement on Steemit and have been rewarded with new badge(s) :

Award for the number of posts published
Award for the total payout received

Click on any badge to view your own Board of Honor on SteemitBoard.
For more information about SteemitBoard, click here

If you no longer want to receive notifications, reply to this comment with the word STOP

By upvoting this notification, you can help all Steemit users. Learn how here!

You've been UpVoted via the UpVote Experiment 002 Bot. Depending on my VP & the price of STEEM you should get a $.01-$.03 for your trouble.
Read more about this experiment here.

Thank You - @blueorgy

Coin Marketplace

STEEM 0.18
TRX 0.13
JST 0.030
BTC 57962.42
ETH 3050.85
USDT 1.00
SBD 2.25