I am aware of the monster "Best practices" post that just got locked. I posted about this topic in the last page but it was probably too late to get any meaningful discussions going. I've searched the forum for EPGP and Relational and have not found any threads so I hope this isn't a dupe topic.

EPGP: http://groups-beta.google.com/group/epgp/
Relational DKP: http://www.curse-gaming.com/en/wow/a...ional-dkp.html

I recently came across these two mods using a loot system different from traditional DKP. Instead of having one value, DKP, that represents one's contributions to the raid as well as the value of gear, EPGP and Relational DKP have three different values.

1. Effort Points (EP) or DKP
Represents one's contributions. EPs or DKPs can be rewarded for attendance, boss kills, learning new content, being on time, staying till the end of raids, etc. It is up to the raid/guild leaders.

2. Gear Points (GP) or Gear Rewards (GR)
GPs or GRs represent gear that one gets from raids. Each raid drop is assigned an appropriate GP or GR value by the raid/guild leaders. Each time one receives an item from a raid, the GP/GR value of that item is added to one's GP/GR.

3. Priority (PR)
PR is the quotient of EP and GP, i.e. EP/GP. When an item drops, the person with the highest PR who wants an item gets it. His GP increases and thus his PR decreases.

Both EPGP and Relational DKP work this way.

Difference between EPGP and Relational DKP: Raid Window
EPGP calculates one's PR only from EP and GP earned in the last X raids where X is configurable by the officers. X is known as the "Raid Window". Relational DKP doesn't have a Raid Window and calculates PR from one's entire raiding history. There is a lengthy discussion between the authors of the two systems about the purpose and effects of the Raid Window so I will not repeat the arguments here. Read the discussion here:
http://groups-beta.google.com/group/...30773e7c34e399

I don't know anything about the author of Relational DKP, but I know EPGP's author is running the system in a 3-raid/week "casual" guild that runs ZG/AQ20. I'm really interested in what hardcore new content raiding guilds think about such a system.

Questions:
Is EP/GP = PR a fair way of distributing loot? In its basic form without the "Raid Window", is it better than traditional DKP or Zero-Sum DKP?

Are inflation and hoarding still big problems in EP/GP's basic form without the "Raid Window"?

What are your thoughts on the "Raid Window"? On the one hand, it is friendly to newbies because they become truly equal to any veteran member after X raids. On the other hand, it heavily encourages people to attend every raid because that maximizes one's EP and thus PR.

The "Raid Window" seems to solve a lot of problems like inflation and hoarding. Would we get equal results if we apply the concept of Raid Windows to traditional or zero-sum DKP? Is the brilliant(?) idea here EP/GP or the Raid Window?

Is the fact that the "Raid Window" wipes out contributions made more than X raids ago really abhorrent to people? Or is it not a big deal?

To conclude, are any new content raiding guilds considering such a system for the TBC?

That's still an option with EPGP, right? You simply stop giving out EPs and GPs for a particular instance and the Raid Window will take care of the rest. Ex. stop giving points for BWL while the guild is working on Naxx and AQ40 in earnest.

What i mean is have a loot system which tracks everything but provides multiple current totals, think:

MC: mc+bwl
BWL: mc+bwl+aq
AQ: bwl+aq+naxx
Naxx: aq+naxx

Or even more simply just delete MC when it doesnt matter anymore, deleting all points earnt and spent. The drawbacks to that sort of setup is it breaks upgrade systems, but simply deleting one shouldnt.

To make upgrade systems work what id like to try is price items roughly (just a basic representation of an ilvl based pricing structure with exponent):

MC: 10
MC: Rag 20
BWL: 20
BWL: Nef 40
AQ: 40
AQ: C'thun 80
NAXX: 80
NAXX Kel'Thuzad 160

But once MC and BWL are on farm people coming into the guild after this point would buy mc items for 10 then be rebated 10 points in an individual adjustment with the whole system rezeroed. Or if they buyin on a bwl item for that slot they would pay 20, be rebated 10 in an individual adjustment then the system rezeroed.

Could be used either once an instance was dropped, or once a whole next tier was available or once gearing reached a certain critical mass. Its a more risk prone anti stratification method but i think it softens loot streakiness, dropping farm content to work on progress for a few weeks and so on.

In one of the other threads on here someone posted a link to smoothed earnings which i would also really love to see implemented into a system but could potentially also pose some stratification risk. I wonder if only smoothing the earnings on actual boss drops but not learning would have a large enough effect to prevent new people paying full and receiving less from veterans when buyin reductions are not used.

Traditional DKP is basically EP/GP but using EP - GP as the equation. What does changing the equation do exactly? Well, nothing really. Take any 2 EP/GP combinations, and they would have the same order under either system and thus would never affect looting order. The result for 400 EP, 300 GP is smaller than the result for 450 EP, 330 GP under both systems for example. It's basically just a fancy way of displaying the same data so it's exactly as fair or unfair as traditional DKP.

Now perceptions can be very important, and I couldn't tell you exactly how people would react to the different perspective given them. It may psychologically reduce hoarding because there isn't a big stack of points to accumilate, and accumilating "stuff" is an activity potential hoarders delight in. Would making a ratio go higher be as tempting?

The raid window would drastically reduce hoarding, but it would also devalue the work your raiders put in, and would make poor dropping streaks hit pretty hard. There are a lot of guilds where practically no <insert class here> loot drops for a period. Fortunately the streak is evened out because they have all their accumilated points to spend on items when they do drop. Personally I'd prefer either zone caps or using different pools for different instances or tiers of content, both allow new players to catch up over time without hurting people for bad loot streaks.

Ah, good point, I was using points totals that were fairly even to test the maths behind it, and hadn't considered the impact of wildly different points (it's 3:40am here btw). I'll take another look when I've slept.

I am pasting here my reply on why EP/GP is different than EP-GP (and I urge you to discuss this on the Google Group of EPGP - hopefully pointing out problems with EPGP and making it better!):

No. This is a major confusion that most had when I tried to explain
the system to them. There is a big difference between EP/GP and EP -
GP.

1) If you have EP/GP, EP and GP does not need to be the same quantity.
Think if EP being meters and GP being time. You can divide the two and
have something meaningful in the end. Whereas if you are subtracting
GP from EP then both must be values of the same quantity. You can't
really subtract seconds from meters can you? Now meters and seconds
aside, what does this mean for this EPGP? It means that there are two
totally different scales to balance. EP and GP. EP is all about
balancing effort. You can measure it as time being in raid. You can
measure it as boss kills. You can measure it as member X wrote up a
tactics guide for the rest of the guild and put it on the website. And
so on. GP is loot. It doesn't have to be epics. You can put prices on
anything and balance them. Put GP on epics, put GP on enchanting
materials, put GP in mere gold. You can argue that this doesn't change
much. You can say all the points in the system are DKP (or Foo Points
or whatever) and you balance them all together. My point is that
separating the two makes it all that easier, since you have comparable
quantities to compare.

2) There is a big difference between subtraction and division. In the
classical model of subtraction you have the problem where the active
members end up getting all the loot but the more casual ones get
nothing (or leftovers just by luck). In order to show this let's take
an example:

A: rogue, hardcore attends all raids
B: rogue, casual, attends half the raids

Assume average EP (DKP) per raid 10. Assume each item costs 20 GP (DKP).

At some point in time the standings are as follows:

A: 400 EP (DKP)
B: 200 EP (DKP)

And one rogue item drops every 4 raids. If you use subtraction, member
B will get NO loot unless A passes. Here's the progression with B not
coming every other raid and both getting 10 EP (DKP) when they come.
When the item drops both A and B are in the run:

A / B
400/200 - start

410/200
400/210 (A loots)
410/210
420/220
430/220
420/230 (A loots)
430/230
440/240
450/240
440/250 (A loots)
...

It is clear that B will never see a piece of loot unless it is
breadcrumbs off the table.

Now let's see how EPGP solves this problem (assume they start with the
same EP (DKP) and loot has same price but this time is GP):

A / B
400/1 (400) / 200/1 (200)

410/1 (410) / 200/1 (200)
420/20 (21.0)/ 210/1 (210) (A loots)
430/20 (21.5)/ 210/1 (210)
440/20 (22.0)/ 220/1 (220)
450/20 (22.5)/ 220/1 (220)
460/20 (23.0)/ 230/20 (11.5) (B loots)
470/20 (23.5)/ 230/20 (11.5)
480/20 (24.0)/ 240/20 (12.0)
490/20 (24.5)/ 240/20 (12.0)
500/40 (12.5) / 250/20 (12.5) (A loots)
510/40 (12.75)/ 250/20 (12.5)
520/40 (13.0) / 260/20 (13.0)
530/40 (13.25)/ 260/20 (13.0)
540/60 (9.0)/ 270/20 (13.5) (A loots)

If you go on an on, the loot sequence will be:

A, B, A, A, B, A, A, B, A, A, B, A, A, B ....

This looks much more fair to me than A getting everything until she needs nothing else :-)

So there is a huge difference by EP/GP and EP - GP. The bottom line is
that in a EP - GP system if a member A can get EPs equal to the price
of most expensive item in the loot table compared to B for the frequency
an item drops, then B gets no loot at all. Whereas in EP/GP both get the
share equivalent to their effort.

EDIT: deleted for redundancy

I think DKP systems are all relatively equal. As long as your system gets gear to the right people who are helping right now, its great. I think the issue becomes when you are awarding disproportional gear right now based on effort 5 months ago that you have issues.

New people, especially, are helped by systems that have a shorter "window" because they catch up faster. New people, ime, are the lifeblood of the guild. If you bleed out your new people, you'll be in a perpetual catch up cycle, whereas if you retain your new people, you'll continually push forward. Keeping long term members is not difficult, if the loot system is fair to them, keeping new members is also not difficult, if the loot system is fair to them also.

This system, if used with the right sort of raid window(4 weeks, perhaps) will get the gear into the hands of your current 40 raiders. Thats what I want from a loot system, that it gets loot, evenly distributed, into the hands of the 40 people who currently show up.

My current system does it, and this system, configured right, sounds like it'd do it.

I've finally compiled my model that I used to track a few different DKP systems and compare the variations.
I've posted them at both my curse-gaming site and sourceforge site.

I believe it shows that when comparing gear to effort (Gear vs. Raid attendance) that a comparative system is more fair (compared to at the very least fixed-price maxDKP wins system), be you the hardcore vet, the hardcore new player, or the casual veteran.

Please let me know if you have any suggestions/questions about the model:

http://www.curse-gaming.com/en/wow/a...ional-dkp.html
or
http://sourceforge.net/projects/relationaldkp/

Look for the lootmodel.zip in the downloads.

EP / GP sounds like a disaster. As soon as a new person starts, they have 0 GP, so infinite priority. Or even when they have their first item, their GP is very low, making it very easy for them to get the next item, which as a new member they don't deserve.

EP - GP works fine so long as you put in a zero sum, say
X = EP of all members
Y = GP of all members
N = X / y (normalising constant)

then:
my real EP = EP / N
my real GP = GP * N

which is exactly the same way it happens in traditional EP - GP systems. The difference lies in the relational curves - the amount your buying power (PR, DKP) changes when you change either EP or GP by a fixed amount, say killing a raidboss or buying an item. As was said above, looking at the quotient or the absolute difference between two people's EP and GP doesn't really affect loot order. What's affected is the relative ease with which that loot order changes.

In a traditional system (EP - GP), this is linear, and stays linear all the time. If A has 500 points more than B, B would have to save up for exactly those 500 points to get precedence on that shiny first boss drop they're both after, regardless if they're at 100 and 600 or at 5000 and 5500 points. At say 200 points a drop, it would mean B taking nothing while A takes 3 drops, given equal attendance.

In a quotient system this can radically change depending on the history of both. Say A and B are both around balanced EP/GP ratios, there can be a huge difference in the way their ratios change due to equal changes in one of the two numbers. If A is at 5000/5000 points, B at 5000/5500 points, it still takes 3 items looted by A for B to get before him in EP/GP, assuming no change in EP. If B is a newcomer though, and maybe took one item on his first raid day, he'll be at something like 100/200 points. If he then decides he wants some drop before A, all it will take for him to get before A in EP/GP is to wait up for 101 EP to come in. Those 101 EP wouldn't really change anything in the first example, though.

Whole point of the example is, that with growing numbers of points earned, the system get's increasingly static. This might not be so bad, but the difference to the ease with which one's ratio is shifted around at low point values might be quite unintended.

Issues like this can probably be tackled by adjusting the general income/price ratio. Giving 10 times as many points as people spend for items would make saving up a bit more viable for those with a long history. The 'raid window' mechanics would make a final normalizer, preventing your point ratio to get too static, up to a point where nothing you do in the next two months would affect your ratio from the last 12 months very much.

All in all seems like a pretty nice thing, being mathematically more sophisticated and allowing for sleeker tuning. Feel you got too much inflation, veterans have too high a ratio? Increase the points earned and it will even out quicker and smoother than with an additive system. That complexity is the downside as well though, it's a bit harder to understand what's going on in this kind of system, and what this change or that would mean, for the average member not willing to really go into it.