Front Page Minimum Votes Raised from 20 Votes to 25 Votes + Rep Bonus 4 Deals/Wk for Popular Deals

• 

  17833 on 05/05/2016 - 07:07

  I've never understood how the front page deals work. I access ozbargain mainly through this link https://www.ozbargain.com.au/deals Does this mean I always see the newest deals, rather than the most popular? Thanks

  • 

    Moderatormoocher on 05/05/2016 - 07:10

    Yes, the /deals link is for all deals posted in order of time of posting.
    The 'front page' refers to https://www.ozbargain.com.au

    Thanks, I'll clarify above.

• 

  tassieeagle on 05/05/2016 - 07:50

  Sounds good

• 

  tightarse on 05/05/2016 - 08:35

  +2

  My prediction came true :) …

  • 

    battler on 05/05/2016 - 09:12

    +2

    what are tonight's powerball numbers?

    • 

      scotty on 05/05/2016 - 11:35

      +1

      TA predicted 3 months ago. You should have asked "what are August's powerball numbers?"

  • 

    GameChanger on 05/05/2016 - 14:27

    -4

    I hope your comment had 0 influence on the decision.

• 

  [Deactivated] on 05/05/2016 - 10:04

  -4

  Would 5 votes make any difference though?

  Because on the balance of probabilities, a deal which attracts 20 votes would likely also reach 25 votes, hence negating any effect this new policy is expected to have.

  It's like printing more money. If everybody suddenly had an extra million dollars printed for them, people would be spending that money to the point that prices are hiked and everything would cost a lot more, so nobody would be better off with that extra cash.

  Considering the ever-increasing population of the OzBargain community, shouldn't the front page threshold be increased to a statistically significant number, such as 30 or 40?

  • 

    GameChanger on 05/05/2016 - 14:25

    -1

    KaptnKaos thank you for bringing logic to this thread.

    I do agree the vote should be increased to a much higher number and perhaps applicable to the established users.

    Whats with your comment getting so many down votes; the down voters please give some of your reasons.

    • 

      gimli on 09/05/2016 - 21:18

      Bandwagon

  • 

    Moderatorhamza23 on 05/05/2016 - 19:56

    On the face of it that would seem the case, if a deal attracts 20 votes then it would be likely to reach 25 votes, however it is more complex than that. Firstly earlier votes are more common and easier to obtain than later votes, ie the first 5 votes are much easier and quicker to receive than the 20-25th votes.

    Just in the past 24 hours there are 4 deals that have received 20 votes, but not 25 votes and most probably won't.

    Example 1 20 votes received in 19 hours, first 5 votes received in 20 minutes, last 5 votes received after 2 hours 40 minutes. Deal usually would have been on the front page 7 hours and 30 minutes ago, now it's not on front page at all. No votes in the last 7 hours and 30 minutes, had it been on the front page for the past 7 hours and 30 minutes as it usually would it probably would have received some 10-15 votes.

    Example 2 20 votes received in 20.5 hours, first 5 votes received in 27 minutes, last 5 votes received after 2 hours 5 minutes. Deal usually would have been on the front page 7 hours and 51 minutes ago, now it's not on the front page at all. No votes in the last 7 hours and 51 minutes, had it been on the front page for the past 7 hours and 51 minutes as it usually would it probably would have received some 10-15 votes.

    Example 3 22 votes received in 21.75 hours, first 5 votes received in 53 minutes, last 5 votes received after 6 hours and 2 minutes. Deal usually would have been on the front page 11 hours and 15 minutes ago, now it's not on front page at all. No votes in the last 6 hours and 30 minutes, had it been on the front page for the past 11 hours and 15 minutes as it usually would, it would have received some 15-20 votes.

    Example 4 20 votes received in 22.5 hours, first 5 votes received after 1 hour and 12 minutes, last 5 votes received after 10 hours and 8 minutes. Deal usually would have been on the front page 1 hour and 50 minutes ago, now it's not on the front page at all. No votes in the last 1 hour and 50 minutes, had it been on the front page for the last 1 hour and 50 minutes as it usually would, it would have received some 2-3 votes.

    Secondly, the front page formula that determines deals hitting the front page earlier than (now) 25 votes is also tweaked, meaning it is more difficult for a deal to hit the front page earlier. A front page deal is much more likely to receive votes than non front page deals.

    Example This deal hit 13 votes in 1 hour and 9 minutes, at that stage the deal would've hit the front page due to the early algorithm (it would have stayed there for a further 21 minutes unless a 14th vote was received, which it would have been and from there would have most likely stayed on the front page up until the 20 vote mark). However due to the 25 vote increase, the algorithm wasn't reached as it's now 15 votes within 1 hour - 1 hour 13 instead of 13 and it never got there. So instead of hitting front page after 1 hour and 9 minutes, it's now not on the front page after 4 hours and 45 minutes and is still 7 votes off hitting the front page if it does at all.

    These examples are only within the last 24 hours, therefore you can see that it does in fact make some difference. There wasn't a problem as such with 20 votes, just that we were receiving slightly more deals on the front page than was ideal (due to increase in popularity), so we have tweaked it up to 25 votes and will study the % of deals to front page deal now and in the future. Should popularity increase and the stats tell us we need to, we can tweak it up to 30 votes or similar in the future. Increasing it by too much regardless is not ideal, increasing it by not enough can easily be increased further if the numbers tell us it needs it. Right now we are slightly increasing it just to push the bottom end of the front page deals off (eg those 5 examples) & to counter increased popularity/voting on deals overall.

    However as explained this change will in fact make some difference. The overall factor? In the past 24 hours we have 16 deals on the front page instead of at least 21. Without taking into account other possible algorithm effects as per the second point.

    Apologises for any typos or errors but I believe everything is correct and would have been that case bar a very unusual occurrence. This took some 20 minutes to write also so times/data might be slightly off.

    • 

      GameChanger on 07/05/2016 - 00:42

      Thanks for the detailed write up Hamza.

      Your post is based on a very small sample size, I will monitor the deals and see the impact of the change.

      Personally I think the system might punish the newbies and the more known users have a monopoly on the front page. Sure one can argue that if the deal is really good, it will rise to the top but sometimes the post isn't marketed to get the clicks.

      I haven't yet thought of a system that will address the voting problem on here, but I feel algorithm should be made that can help all posts to be treated equally.

      As to the rep system I'm quite happy with it, as I feel us Ozbargainers give the reps a very hard time.

      • 

        spackbace on 07/05/2016 - 10:42

        Personally I think the system might punish the newbies and the more known users have a monopoly on the front page.

        It all balances out.

        Highest voted deal for March was the 2nd post for that member, and subsequently scored themselves a monthly prize.

        And any 'monopoly' of the front page is likely because the more known/experience users tend to pick and choose what they post, and only post deals of a higher caliber, and therefore they appear to get a higher vote count.

        It's like singling out the smartest kid in the class, and scoring their test scores differently to everyone else. Why would you do something like that?

      • 

        GameChanger on 07/05/2016 - 10:59

        @Spackbace: That deal was free, not really fair to use that example.

        Its not a matter of appearing of getting a higher vote; they do and its being acknowledged by the Mods- the term coined 'Ozbargain friends'.

        Again the class room example doesn't relate to this at all. There is countless examples where a less known user has received half the votes for posting a really good bargain and the more known users post 'bargains' that are very questionable. To be clear this isn't an attack on well known users as they can't control user behavior, but rather the algorithm that causes it.

        In my analysis I have noted that the Ozbargain algorithm, has made the min vote requirements redundant for known users. Once a known user reaches the front page with <25, there is 0 possibility that it falls off it. And to be fair this is a flaw for both known and new users because once the deal is on the front page herd behavior kicks in and the deal receives a good amount of votes.

• 

  gimli on 05/05/2016 - 22:08

  ,+