Topic: Bumping changes on some boards - page 13. (Read 8877 times)

I have found a bug:

I created a sockpuppet account with zero merit and activity, testing321, created a thread in service discussion, and moved it to service announcements. I am not sure if it got superbumped or not, but I presume not, and after moving the topic, it is the 5th highest thread. I am not sure how it is being sorted, but I don't think it is by bump score.

Great change.. This will make it possible to actually find topics in the Altcoin sections that at least have some working teams/ intellect behind them rather than simply being bumped by "Reports" and such.
 
I stopped looking at the Altcoin sections because they are all the same. Now if the topics show some real intellectual possibility, they may become easier to spot. Whether the Altcoin section falls to farming or not will now depend on how many of the honest users actually spend time their. If the majority of honest users are maximalists or have given up completely on Alts (which seems to be the case presently at reddit and twitter), they will still fall back to the farms.

I don't think this is correct.

If you assume someone has a starting base bump score of 100:
On month 1, a forum member can make 30 superbumbs, and collectively add 660.68 to bump scores.

If the forum member made one superbump per day, and continues doing so, on the 31st day, the additional superbump will add an additional 1.056 points. The same number of points will be added on days 2-60 because she will continuously have made 29 superbumps in the last 30 days immediately prior to making that day's superbump. The total collective bump score from superbumping this way is ~31.68 from day 31-60. If someone makes 30 superbumps on day one, waits 30 days + one second, the first superbump on day 31 will be worth 100.

---

I think this sorting should only be applied for topics last posted in the last x amount of time, and if not, to sort by the default last posted date. As a test, I was able to bump a topic last posted in 7 years ago, and someone else bumped a topic last posted in 2013. By allowing very old threads to be bumped like this, you may be inadvertently be encouraging non-relevant threads to be nerco bumped.

I'm not concerned that much about my own bumps, just looking for loopholes that farmers could exploit. It looks like if they have ~30 high-merit accounts they could keep one or two threads super-bumped to the front page. Likewise if they have a bunch of high-merit accounts they could also mini-bump multiple threads. That might be more feasible since mini bumps don't cost anything (other than maintaining earned merits) but they would need a lot of accounts to overcome the 7-day limit.

Those sound like very costly scenarios so likely not a big deal.

Will these changes be extended to some boards such as Digital goods,  Invites & Accounts, Services, and local boards (Announcements (Altcoins)/tokens)?

Are there any changes to the old topics in the discussion boards? Those with dozens of pages.

By "effect" I meant "total bump score contributed to good topics". By that measure, your two scenarios are equal assuming the same set of superbumps.

If your goal is to be able to occasionally try to ensure - with highest probability - that some topics are at the top for a few specific days (before your contribution decays), rationing like that could make sense. On the other hand, if your goal is to try to continuously keep some topics as high as possible, you would want to spread it out as evenly as possible due to decay.

But really, people should just superbump good topics whenever they think to do so, and not worry about this sort of strategizing.

Hmmm... equally spacing out super bumps over the course of the month seems to result in a lower total bumping effect than e.g. sending a bunch of superbumps in one day, waiting 30 days and 1 minute, and doing it again.

You can do both mini bumps and super bumps -- they don't interfere with each other. If you try to ration your super bumps, then you're just throwing some of that potential effect away.

Does it always increase though? It seems to top out at ~6 * 162, but if I bump e.g. 30 times a month my super-bump value is only ~3 so that's only ~90 per month total, unless I messed something up in my math. The mini bump seems more valuable at that point.

Because it's a rolling 30-day period (like source merit), I think that eventually people will end up with a fairly consistent super-bump power depending on how often they tend to do it. Note that super-bumping as often as possible always increases your overall effect, even if your one-off effect gets limited.

It feels a bit backwards that my most recently super-bumped thread gets the least bump power. Maybe I just need to get used to it.

If I got the formula right:

First super bump ~364 "points"
Second ~309
Third ~263
Fourth ~224
etc

Another opportunity for few users to start complaining about this change. By the way, this is a realistic move which was needed to take control of these boards. Good job theymos.

There will be no more newbie accounts with first post saying that, "This is a great project and I am selling my kidney to join it" :-P

Maybe later. If it's shown somewhere, I'm not sure where to put it.


Around that. It varies slightly depending on how the periods line up. It's as if you signed up exactly 365 days ago.


Correct.


I don't think that people should worry about this. If you're exposing problems, then it's good for it to be bumped for a while. (And if it's self-moderated, then the poster has some incentive not to delete your post, since doing so undoes your bump.) If the poster wants exposure more than anything, such that they're happy you posted to criticize, then - once people get used to this new system -, the topic will probably already have a much higher bump score than what you're contributing.

You just used your new-found powers to bump this to the top:

https://bitcointalksearch.org/topic/m.52420105

I would have thought more active topics with more posters would have higher scores than a topic with 2 posters, one of them a newbie.

And of course the fact that critique posts bump the topic is an unfortunate side effect. Most of my own bump power will be wasted on various scam coin threads.

I think it will still encourage certain users to sell their bumping power. Like if one dubious member realizes that he/she has good bumping power. Nothing can stop him/her from PMing different project thread starters with screenshots that they can help them bump the thread for a price.

I think it's for the better it remains mysterious

Those screams you're hearing are bump-farmers grieving their "business". Great news.


Can you at least make their own bump power visible to each user? Might help with debugging.


So this basically means maximum possible bump power is 364?


Mini bumps don't reduce one's bump power, right?

---

The biggest takeaway for me so far is that zero merits == zero bump power.

There's long been a problem of spammy bumping being used to keep topics near the top, but lately it's become unbearable. Therefore, on these boards, bumping has changed:
 - Service Announcements
 - Announcements (Altcoins)
 - Tokens (Altcoins)
 - Bounties (Altcoins)

Topics are no longer ordered by last post on these boards. Instead, each topic has a bump score which is calculated based on recent "bump events". Posting in the topic once per 7 days increases the topic's bump score. Clicking the new "bump" link (near "watch", "print", etc.) once per day increases the topic's bump score up to 100x as much, but it also decreases the power of your future bump-link-clicks for 30 days. Neither action is guaranteed to bring the topic to the top, since it only increases the topic's score.

I can see several things about this which may not be ideal, so adjustments will likely need to be made, but something is probably better than nothing.

---

What you need to know

 1. Don't worry about it too much. If you behave as normal, you'll do OK.
 2. If you like a topic and you see a "bump" link near the "watch", "print", etc. links in a topic, click it.
 3. If you've been spam-bumping, knock it off, since it's not going to work anymore.

---

Technical details

Base bump power

Your base bump power is the smaller of these two values:
 - The activity you would have if you'd signed up 1 year ago.
 - The merit you've earned in the last year.

If it's zero (as it will be for newbies), then you don't bump at all.

Mini bumps

If you haven't posted in a topic in the last 7 days, posting in it adds 1% of your base bump power to the topic's bump score. Deleted posts count against the 7-day limit, but not the topic's bump score.

Super bumps

Once per 24 hours, you can click a topic's bump link. When you do this, a modified bump power is calculated as something like your base bump power minus 15% for every super bump you've done in the last 30 days, and this is added to the topic's bump score.

The exact contribution is b*(1-0.15*^s/_b), where b is your base power and s is the total you've contributed to bump scores via super bumps over the last 30 days.

Bump score

A topic's bump score is affected by the above two actions over the last 36 hours. Each action's effect decays linearly over the 36 hours, so an event which contributed 1 at the beginning will contribute 0.5 after 18 hours.

Topics with equal bump score (such as topics which haven't been posted in for 36 hours) are sorted by last post time.

Non-default sorting

On the targeted boards, sorting under this scheme is the default. You can go back to the older sorting by clicking the "last post" column header twice.

On the non-targeted boards, you can see the sorting under this scheme by using a link like https://bitcointalk.org/index.php?board=1.0;sort=bump;desc . I'm not sure if this is at all useful. (It's also possible to super-bump topics on these boards, though there's no real incentive to do so.)

Non-publicity of data

For now, I am intentionally not making it public who bumped each topic, since that way you can't prove that you bumped anything, and this may prevent people from selling their bumps as easily. I don't particularly object to making the approximate bump scores or bump powers public, but I didn't do this yet.

---

If you find bugs, PM me.