Jaybob wrote:aquasoup wrote:
My point has always been in very simple terms, what happens after that 7k, why flat stop, why not either have a law of diminishing returns so you get steadily more xp penalty upto and over the 7k, what's called a soft cap.
that's not a bad idea, but it may require so much work (i have no idea) for the team to implement, that it may be better to try the planned one out for a little while first. Although the planned change is being tested by us, there's really no way to see how it will work with our real characters and in the real-world Avlis until it gets implemented.
Just for Jaybob, here's the maths, it might look a bit complicated, but I'll try and explain it as clearly as I can - I'm not really a maths person so bear with me.
The aim is to create a system where instead of getting a constant value for the same kill until you hit 7k inside this week and it suddenly cuts off, instead you have a gradient that trails off to almost nothing but never quite reaches it.
okay lets say X is the normal value of a kill, this is the current figure that's produced depending on your party size and levels and CR's and all the other things that go intro producing how much killing that last giant was worth.
Lets say the giant was worth 200, i.e X=200.
Next up in order to have any sort of 7k cap, you have to have been tracking how much xp people have got inside this week in order to tell when they're got to 7k of course, so lets call this Y, i..e total xp got this week.
We're going to make a little formula to produce a value of P, basically a multiplier to adjust how much that last giant kill is currently worth to
you after all other calculations, solely depending on how much adventuring
you have done this week.
To produce our gradient, we'll set a midpoint of 3000 say, anything before 3000, gets a touch more xp anything after gets increasingly less. To do that, you use a fairly simple statement to produce the multiplier P.
P=1+(3000-Y)/10000)
Alright I'll try explain how that works, if you have say 1150 xp accrued so far this week, then 3000-1150=1850, divide 1850 by 10,000 gives you .185 then finally add that to the 1. So your end value for P is 1.185.
That giant that's worth a flat 200 is currently worth to
you 200*1.185=237 because
you have only got 1150 xp so far this week.
Going over that again, to produce the value of current kills you times X the normal kill value determined by current situation, by P the multiplier for how much powergrinding you've been doing so far this week

to produce XP.
Computers can naturally simplify the end product of sums for you, so you give the XP product the properties of a whole positive number, i.e. if the sum comes out as 1.253 the computer says that's a 1,
0.241 is a 1,
-398.25 is a 1,
1.49999 is a 1,
1.5 is a 2,
2.237 is a 2
and so on. The computer cleans it up to give you neat 1 2 5 23 whatever at the end.
Of course that last kill has changed your Y value, you've now got more xp this week, so if you kill another giant in exactly the same way, the new 1 will be worth 200*1.1613=232.26 which the computer will clean up to be 232.
If you want to look at it another way, you start out with a 1.3 multiplier at the beginning of the week, if you just got 237 xp for a kill, it lops 0.0237 off the multiplier giving your next kill value a multiplication of 1.2763.
The formula changeover at the 3000 we picked just means you get more XP before 3000 and less after it.
For example when you have 5600 xp so far this week, that giant kill is worth 200*0.74=148. If you get upto 12350 xp that giant is worth 200*0.065=13. If anyone gets to 13001xp in that week, the formula goes negative and everything beyond that produces 1xp as you have your XP value set to return positive whole numbers..
It naturally slows down your xp, making it harder and harder to powergrind the more you are doing inside your basis period, which is a week currently, but it never entirely stops you dead, because XP is a whole positive number, it's always going to produce a minimum value of 1, but I think people will get noticed killing nearly a million crop rats, if that bothers people give the XP properties of a whole integer ie. zero or above, so it can also produce zero on anything 4.9999 and below.
The 3000 was just an arbitrary pick, but it essentially puts you on track for your 7k target with a bit more before 3000 that week and steadily less after. You could go for 1000 for example if the fact that the soft cap allows progress, if increasingly slower progress beyond 7k, is an issue.
Additionally it's worth pointing out that when most kills are worth alot lower than 200xp to folk, you are going to see the 1xp per kill cutoff much earlier than the above examples. if you want to cutoff to 1xp or the zero option earlier - just subtract something from P before it's used to multiply, i.e. P=0.8+((3000-Y)/10000) so starting you out from a 1.1multiplier and dropping from there.
There's an example of the maths for one type of softcap, if it's any interest to folk or not.... cue the comedy pictures
