#30 Alternate Priority math: Average DKP & GR out by raid attend

[Alex Bender]

From: <clipped for privacy>
To: Shazear <shazear@hotmail.com>
Subject: Observation (and possible solution)
Date: Tue, 14 Nov 2006 17:56:02 -0500

Greetings again Shaz,

Still evaluating RDKP here, and still running lots of numbers. After approaching you about the initial GR values, I spent some more time stretching things out and seeing how RDKP would look a year down the line. Here's what I got:

1 Person attends 12 raids per month at an average of 30DKP earned per raid. A year later this person has earned 4320DKP. Lets make it 4000 for the sake of simplicity. Lets say he's grabbed a dozen items (an item per month) with an average of... 40DKP. That's 480GR... lets make it 500 for simplicity.

His priority is now 4000/500 = 8.

This is where I see the problem starting. Think about how hard it will be for him to increase his priority by a point. Without taking anything else, it'll require 500DKP to increase his priority by one point. As a matter of fact, the longer a person raids, the more and more difficult it will be to increase their priority... though having their priority decrease will always be easy *grin*. If he spends 50DKP he loses almost a whole point in priority, yet it takes 3600 earned DKP to make that point of priority back.

*wince*

As much as I was initially against the raid window concept, the raid window does completely resolve this problem by giving more or less a cap to the GR value... and having it diminish over time. BUT! The raid window is not the only solution *grin* Keep reading...

As you can see, only veteran raiders would suffer from this. New raiders would not be affected, allowing them to easily overcome veteran raiders in priority, and fall back down below them, and easily gain on them again, until eventually they are in the same situation as the veteran raider.

One possible solution is instead of using a window... use a divisor. For example... Divide both earned DKP and GR by a number that grows over time. I don't know what a good number would be just yet... but lets say we used the number of raids. This is how it would look using the above example of the 4000/500 guy (Who ran 12 raids a month for 12 months, or 144 total raids)

(4000 / 144) / (500 / 144) == 27.778 / 3.472 == 8

The loot priority remains the same, but the GR value is something much more manageable. Of course you couldn't use the same divisor for everyone, so perhaps you'd need to track raids per person, and each person would have their own divisor. If you did it that way... lets say you had a guy that put in half the effort (ran half the raids), and took half as much loot. He should be equal in loot priority.

(2000 / 72) / (250 / 72) == 27.778 / 3.472 == 8

Yep. As a matter of fact, the GR's work out to be the same *grin*, which means the full-time raider and half-time raider can increase their loot priority with the exact same amount of effort!

At this point, what you'd really be working with is averages for each person instead of exact values. But, I hope I've proven that this may indeed be more fair. Extra work? I don't know. You'd need to store raids per player (including Alts). So you'd be storing one more piece of data. And you'd be doing two more divisions... of course, it always sounds more simple on paper than it sometimes can be in code.

I'd be interested to hear your opinion on this.

Malagul

[Alex Bender]

Logged In: YES
user_id=605887
Originator: YES

For now, down-grading this option. We have the new threshold method and we'll see how that works.