Corion, good point about the "many nasty adjustments." Whatever system is developed, it should be simple enough that a newbie can looks at it and say "oh, yeah." There are many different factors that can be taken into account:
- Votes received
- Votes possible
- Total number of Monks
- Total number of Monks logged in at time/day of vote
- Level of Monk that voted you +/-
- Whether or not vroom likes you ;)
I'm sure there are others and I don't think that we should take all of these into account. But how do we consider these without making the XP algorithm a candidate for the Obfuscated Perl Contest?
Perhaps we should acknowledge the the algorithm will be complicated, but keep the XP system? People would understand a simple XP system, and might accept that the allocation of XP is a mysterious thing (anyone remember their gaming days?). The algorithm can be there for them to delve into if they're really curious (and any true Monk would be). | [reply] |
Good point.
Some further points:
- We should make votes from higher monks more weighty, and
if a monk is above a certain level (monk, for instance), they
should be able to see a post and the score of the post before
voting to ensure that the post is not getting inflated
uneccesarily.
- If the XP based levels ever need to be changed, vroom
should probably change the XP for each level by a set
percentage, and then increase each existing monk's XP by
that same percentage.
- Questions, depending on the level of the monk posting,
should be limited
to a certain reputation, but higher level monks can vote on a
node regardless of reputation of the node.
- Posts in the Snippets, Code, Craft, and Obfuscation
nodes should have more bearing on the XP of the poster than
questions.
Any other ideas?
J. J. Horner
Linux, Perl, Apache, Stronghold, Unix
jhorner@knoxlug.org http://www.knoxlug.org/
| [reply] |
jjhorner said:
- We should make votes from higher monks more weighty,
and if a monk is above a certain level (monk, for instance),
they should be able to see a post and the score of the post
before voting to ensure that the post is not getting
inflated uneccesarily.
I like this idea. There seems to be a tendency for some
people to vote up a post simply because of the reputation of
the author -- even if the post was rude, inflammatory or
just plain incorrect. Some posts could receive negative
votes for similar, non-quality-related reasons. Higher
level monks are expected to contribute to the quality of the
site and to help regulate the workings of Perl Monks. This
would definitely help them fulfill a "balancing" role...
Russ
| [reply] |
Somehow I like the idea of black-box-XP, but the problem still remains, there must be a way, obvious, public or cryptic and hidden, to distribute XP.
Maybe a system based on the potential votes would work indeed. Let's keep the XP as they are, and you still get XP for logging in. No XP just for creating a node. XP are then given via the ratio of votes for this post out of the total votes for each day. That would mean, if a post gets (say) 1% of all positive (resp. negative) votes on a day, the user gets 1 XP (resp. loses 1 XP).
This system would be untouched by the number of posts/votes, and the only anomalies would be on really lazy days, where a post that gets one vote already gets 1 XP, but the number of idle days should diminish with the number of active monks.
| [reply] |
XP based on percentages don't sound attractive to me, because, when we have 10,000,000 posts a day, percentages will be pretty low. Say, why not calculate the average post's reputation, then derive the number of XPs the user receives based on that?
| [reply] |