Score Trending to PR Formula (in General)
Score trend to pr formula
I posted all of this in another thread, but not all in one post as one cohesive idea so wanted to do that here.
Here is the formula to be used each night at cache flush:
If score is less than pr and pr is less than 4,000,000 then new score=(1/10*(pr-score))+old score
At 1/10th this would be the daily changes to score based on a 2m score and 3m pr:
It would take over a month for score to get very close to pr and the daily increment if someone continued farming a character would only be about 100k score increase. I kinda like this ratio as it isn't a big chunk of score change at any one time and i think that would make it less exploitable as again the n*b growth rate would likely carry them past any target before score could regen enough to matter.
Both the cap on pr and the ratio could still be changed, but this is just to show how it would all work. I copied and pasted from several posts and if it reads a little weird that is why. Hopefully everyone will understand what it is doing though.
August 4 2010 11:11 AM EDT
That is way too much score regen and it is also too wide spread. If you want a more workable solution using that method it would much more likely have to be 1/10 every week at sunday cache flush and only for characters in the 9 and 10 ba regeneration rates.
nat, perhaps after we recover some of the score that was drained when tournaments were sapping players then we can change the percentage and the cutoff. i do think that for one month we should let it go as written for a one-time fix though.
instead of using a hard-coded mpr though it probably would be wise to have that based on a percentage of the top mpr though just as ba regen is now.
i see implementation of this being a two-stage process. first stage is to give back some score to inactive players drained from tournaments to help the early game and realign score to something near pr all the way up to where we have active players.
the second stage would then be readjusting the percentage and cutoff to maintain things for the lowest levels of the game.
once the formula is in place, tweaking things would be very easy!
August 4 2010 11:42 AM EDT
I think initially we would see a lot of score 'produced' under this suggestion.
However, over time, I think it would reach a much lower and more stable rate of score production.
The concern I have is with 'gimped' chars or just plain bad strats getting propped up to be farmed over and over.
At the same time I don't think a reasonably managed farm shouldn't be allowed to keep some score to give newbies:
a) valid targets
b) strat tips
c) a better feel for CB
A simpler version would then to just reset score back to PR for all the lower regions and be done with it.
Might as well do everyone...
we could make it a monthly thing maybe.
That would give me over 30% CB...
active players can gimp characters now and when they build them back up drain much score from the system. it is much easier to just re-equip items or remove the mage shield from the mage and fight the score back up than it would be to wait for the daily or weekly trending.
down low you could leave an inactive gimped mage with a mage shield on it and let it auto trend i guess. wouldn't our auto retire code take care of those issues though and if not shouldn't it?
The auto retire would not work with a daily score trend. However it might still work with a weekly score trend of 1/10.
i don't think we should do everyone. i really consider us having about 70 really active players. look at the weekly clan mvp's and see how many fights are done by the top 70 and you will see why i say that we only have about 70 very active players now.
i set my threshold at 4m pr as that is where the top 70 pr's fall. i do realize that some of the most active will be n*b's and not in that zone but figured it was close enough.
i do kinda like it being based more on the percentage of top pr though, so maybe make the 6/20 zone not regen if we are going to leave this in and not do a two stage process.
i think those details would likely be up to ns though but would love to see anything tried at this point.
August 4 2010 11:58 AM EDT
a one time reset of the 9 and 10 regen brackets might be a good start... they would have suffered the greatest impact of tourney drain
For reference, there are 10,810 parties that are not retired and not in a tournament that have not fought a battle in the last 168 hours.
I guess I should filter by parties with actual minions...
So looking at parties that:
haven't fought in the last week
aren't in a tournament
That brings the total to 10,154.
I could dump that into a csv and plot score vs power.
Ok, well.. google's excel isn't as great as MS excel, even older versions. You can't do for instance logarithmic scales. But basically, the data shows that there's not so much of a disparity.
I can make the data available to anyone that wants to do further analysis.
i cm'd with gl and got some of the major aberrations he had seen over the last few months. it would appear that they have since been auto-retired.
i guess the issue is taken care of by the auto-retire system, i just hate to see teams go bye-bye with our population dwindling as well.
it adds to the ghost town feeling i have been getting! ; (
August 7 2010 1:21 PM EDT
Double kill! Adrenaline full! Booster!
I think the major problem, is that while teams have avoided auto retirement, they have be drained of the excess Score they would normally have, from fighting up.
If you want, all the stagnent teams down low are 'lean' and there aren't enough 'fat' targets for newer teams to get high Bonuses off of when starting.
But, this can be addressed by maintained NPC teams. ;)
August 7 2010 6:06 PM EDT
These NPC teams, are they manually trained by admins or just spawned by the system?
They would be manually trained by Admin, but it's a WIP.
when a problem comes along, you must wip it! ; )
Whoops my bad. Read that as What is WIP
This thread is closed to new posts.
However, you are welcome to reference it
from a new thread; link this with the html
<a href="/bboard/q-and-a-fetch-msg.tcl?msg_id=0034yZ">Score Trending to PR Formula</a>