terencew's Feed

terencew
RSS
11-08-2018 at 05:55 AM
8 Comments
Rate this Entry

A quick thought about CR penalties...

... just as it popped into my head. Even if it is flawed and I should've pondered a bit more about it, I'm just gonna throw it in here and see how it lands. Apologies if it's been suggested before, or if it's just rubbish.


It comes with my ignorance of the formula of when votes are finally tallied. It also assumes that CR is still a thing for at least the near future, and incorrect votes are penalized. I still have problems with this whole CR reward/penalty concept, but here goes...


What if we had a base penalty rate for an incorrect vote (as we currently do at 5%), but it was scaled back somewhat proportionally to the final yes/no ratio?


If the final vote is close, the penalty should be much less than if you're on the wrong side of a landslide.


This would encourage people to vote on conscience, rather than protecting their CR by abstaining, or watching the opinions of those with higher CR and basing votes accordingly.


Probably needs more thought...

Comments
  1. bensweeneyonbass's Avatar

    I like it!

    If it's a hung jury with 60% casting the "correct" vote, you'd think the other 40% weren't "as wrong" as when 98% cast the correct vote and only 2% were wrong.

    ThanksJJT_Defender thanked this post
    LikesJJT_Defender, xcess dead rew liked this post
  2. GibGirl's Avatar

    I do like the idea on first thought, as the more contentious a submission is, the more we should want people to vote - and vote honestly.


    There is, after all, a big difference between the bad vote someone makes from blind voting, and informed honest votes that just don't agree with the eventual outcome. We want to penalize the former, but the latter should be welcomed if we want the system to work well.

  3. bensweeneyonbass's Avatar

    ^I agree. As of now most people abstain on contentious submissions in fear of a CR hit. Then it sits forever.

  4. Snowflake's Avatar

    I agree -- and mind you, i nomally thing 5% isnt enough. so i'll elaborate why in this case i agree with the leniency

    it stands to reason getting a tough thing correct deserves a higher "reward" (again i dont agree its a reward/punishment per se so much as a relection of accurcy, but ill go with the terms) since well, it shows more expertise. Likewise it stands to reason getting a tough one wrong show less ignorace (or whatever the opposite of expertise is) and so should come with a smaller penalty.

    there is one danger to this. part of expertise, is knowing yourself. i'm not sure i want to encourage noobs to vote on complex submissions.

    I read an interesting article once about how to manage votes. All votes are bets -- but you have to let the gambler choose the value of the bet. This way the most sure people bet the most, and the least sure bet the least. At first the system just measures confidence, not expetise, but with time, only people who know their limitations make a profit by only voting on what they know. eventually these weighted gambling votes become very accurate. It was stated the stock markert (which has gambling votes on the price of a stocker when you think about it, every investor is essentilly saying how much they thikn the stock should be and bidding accordingly) can better predict the weather by looking at florida orange prices than metereoligsts can. the stock market works like tg, rather than a select few refs/expert scientists it pulls in a large crowd with weighted gambling votes and shockingly that works out more accurately.



    so although i agree in theory that making a mistake on a controversial sub doesnt really show as much lack of experience as mistakes on other subs and i would vote in favor of your proposition if tg put it to us decide, I just want to consider potential issues as well. the issue being i dont necessarily think its good to encourage people to vote on things they dont understand.

  5. Snowflake's Avatar

    oh another point. theres two different types of controversial subs. controversy over a factual debate "does this actually violate a rule" and controversy over true opinion "should this be the policy.....". For opinion things there should be no cred hit. i'd actually rather have a commision, or board of refs (hopefully elected by us), or a forum where dave or jace makes the call to decide policy, but if tg refuses to do that and insists on using case by case submissions to decide policy, then sure, votes that are actually about policy shouldnt come with a penalty.

  6. GibGirl's Avatar

    Aside: I'm not sure we should call it voting "right" or "wrong". It's more about voting "in agreement" with the community or not. Credibility still has relevance, as it indicates your understanding of the community's rules, but I think that someone's vote not matching doesn't necessarily mean they're "wrong".

  7. Snowflake's Avatar

    thats true gibgirl and its what i was trying to get at when talking about policy and opinion. in those cases i 100% agree its not at all about wrong/right but just popular. which is fine, many elections are about opinion, but in those case we should be honest what it is.

    as for the other cases, true it is techncially going with the flow, but presumably/hopefully the flow is correct.

    LikesGibGirl liked this post
  8. D.B. Cooper's Avatar

    At first I liked this idea. Then I read an article about how the state I live in is the last to allow jury verdicts that are less than 100% and it got me thinking. If a submission can't get to at least something like 80 or 90% consensus, does it really deserve to accepted? Obviously 100% isn't reasonable but if it's only getting a small majority, maybe it's ok that it doesn't get accepted. But I also think TG should have a policy that if a submission doesn't get accepted in a certain amount of time, it should be cancelled and no cred gained or lost by anyone. And I also think they should have a separate place for policy things to be voted on.

    LikesSnowflake liked this post
Join us