Being the place to discuss stuff of value, i would like to share here, what we came across as a guild, in hopes that someone might find it usefull, suggest improvements or predict problems.
A bit about us as a guid, to understand from where the needs for such system have grown.
We raid 5 days a week, for 4 hours, with 35-50 online in the evenings.
HC DKP or as it was originally named Hardcore Attendance Bonus, is something that came up when we started continuous wiping on Ragnaros. It was the first boss who required not a few tries but weeks of wipes. It was obvious to us that getting 40 DKP from killing farm bosses in one night run up to Domo, and getting 0 from wiping to Ragnaros for endless hours was fundamentally wrong - someone who missed that one night run clear, but attended all the wipes and did his best for guild progress deserves the purple rewards more then someone who comes to only farm runs.
Thus in it's simplest form, 2 HC DKP was given for an hour of effort (compare it for example to 5 constant DKP from Lucifron kill, yes in the beginning we didn't have zero sum either).
This was a bit better, but still not enough. We looked upon the situation and wrote the following:
Boss on farm status = chance for an item, 2-3 gold, X DKP.
Boss on wipe status = minus 20-30 gold.
Obviously, there is the emotional part of beating the encounter, but DKP as we saw it was there to give perfect practical reflection of player's effort for guild progress.
Thus we balanced ourself accordingly, halfed the DKP gain from bosses, and re-wrote the HC DKP gain formula to :
HC DKP = time spent + progress made
Later on we decided to switch to zero-sum system for deciding DKP gain from boss kill, but the HC DKP remained.
A good example for this system would be our Huhuran attempts :
One night we started with 14% as best attempt over a few weeks of progress, to 1% after 4 hours of trying, and then instead of finishing, re-supplied ourself with potions, repaired, and kept trying for another 1.5 hours just to hit that same 1%.
For that increased time spent, and the incredible progress made, we awarded 20 DKP.
Next day, she died after additional 2-3 tries. The reward for her was just 8 DKP, some as zero-sum loot and some as a bit of HC DKP.
HC DKP comes from a point that says defeating hard encounter is not a one night stand of performance.
It's continous progress you make as a guild, and rewarding that is of extreme importance.
We as a guild know our limits. We know we have some better players, and some worse. Some who will attend every try, and those with lower attendance. And we want our awarded DKP to reflect each and every person's investment uniquely.
Ideally you don't need any extra incentive for people to come beat your content.
The problem with awarding time spent is that it encourages sloppy behaviour once things are farm-status.
There's nothing all that unique about your system. Tons and tons of guilds give DKP for wipes. We used to. The Conquest zero-sum DKP system upon which we originally modeled our system had IDKP (Incentive?) awarded for wipes and zeroed out of the system periodically.
People who are consistently and mysteriously absent for wipes and always there for farm runs shouldn't be rewarded, that's true. And awarding learning DKP is obviously one way of doing it. For us, I prefer to use invite pressure. If you don't show up for the wipes, you go to the end of the line when it comes to the first several weeks, at least, of farming the newly-mastered boss. And we reimburse repairs on pure wipefests in the learning stages so that there's no downside for showing up. The problem with wipe DKP is that it's inflationary. In a non zero-sum system or bidding system, it's a fine solution with few downsides. Under zero-sum DKP, however, it's very much a rich-get-richer sort of thing. The undergeared new recruit who's trying to catch up might really want to come to the hardest new fights to help learn them, but he isn't geared enough to be useful, so he doesn't get to go. And those who already have the best gear and most DKP get even more. Also, if new raiders join your system, they never had the chance to get those one-time rewards while something was being learned.
Invite Pressure works very well in a guild that is large enough to have significantly more people that show up for a raid then there are spots. However, in smaller guilds that don't have a problem with a large number of excess people that really isn't a valid incentive. If you are using a zero sum system than I think its appropriate to award DKP for wipping and then periodically balancing the system.
I know in an earlier thread Gurg said that one of his main dislikes with rebalancing is that it's possible for a new recruit to join right after a balence went into effect so he doesn't get the minus DKP while everyone else does. I think the best solution to that problem is instead of doing a huge balence once in a while using small balences that are often done.
Also, the undergeared recruit really isn't competing with the long time members for gear. It's more of a he goes to MC and takes all the loot he wants kindof deal.
Well, you have to think in the long-term. Obviously a guy who's in all-blues today isn't going to be up for Kel'Thuzad loot in 4 months, so who cares?
But when we were learning, say, Vael, there absolutely were people who were undergeared to help but today are among our most active and best-geared raiders. The point is that inflation has hardly any short-term effects, but as a long-term policy the effects can be significant when stretched out over a year or more.
That's exactly the problem my guild is having now. We don't use sum-zero. The old officers gave boatloads of DKP for learning attempts and caused horrible inflation. We're having steady deflation nowadays, but new members (particularly warriors) will only catch up to our oldschool people sometime in 2012 at the current rate.
We're highly considering adapting the EJ upgrade system, but it doesn't eliminate our biggest problem of having such a massive disparity in points banked. For example: I have roughly 260 dkp and I've been around since day one with about 89% lifetime attendance. A new rogue who joined six months ago and has 95% lifetime attendance has -100 dkp.
Or consider Himeko, one of our warriors who joined about 9 months ago and has maintained 98% attendance the whole time. We consider him the United States of dkp because at this point, the concept of not being in debt is just silly to him. When we adjust using the upgrade system, he goes from like -150 to about +46. That's more in line with where he should be. It's obscene that such a reliable, contributing member could never dig himself out of a hole like that.
Things we're considering:
# Adapting the upgrade system and applying it retroactively.
# Adjusting the cost of equipment sets for classes that need less gear. (E.g. increasing the cost of all rogue armor by 30%)
It's a shitty situation to be in that I have no idea how to adequately fix without flushing it all away and starting over.
Yeah, I'm a strong proponent of using quick hacks and inelegant band-aids to fix symptoms for the next few months and then just scrapping everything and starting over from scratch for the expansion, with all of one's accumulated knowledge about implementing a DKP system in WoW raiding.
Somebody in our guild came up with the idea that we call "priority points," basically just learning DKP per hour that doesn't factor into the real zero-sum system. It allows people who hang around for learning to buy items out of order by spending their priority points. Since it doesn't change any numbers in our zero sum system (which got way out of control before we reset it) it's kind of a band-aid fix for learning DKP.
I've been giving some thought to going into the code for eqdkp and adding in some logic to account for point capped instances, upgrades, and sidegrades. Are other guilds considering the same thing, or are you going to write something new from the ground up?
The other, more complicated, thing we're considering is doing a two pool system where class specific items earn and subtract points in a different pool. Everyone could get raid gear competitively withing their class, but not be at any disadvantage for cross class items if their class gets a few extra drops. We have problems with people holding out on spending points so they will be first in line for a claw of chromaggus or drake fang talisman, even at the expense of not improving their raiding gear. Unfortunately with our current rules this possible because most of the raid is violently opposed to forcing items on people.
The other, more complicated, thing we're considering is doing a two pool system where class specific items earn and subtract points in a different pool. Everyone could get raid gear competitively withing their class, but not be at any disadvantage for cross class items if their class gets a few extra drops. We have problems with people holding out on spending points so they will be first in line for a claw of chromaggus or drake fang talisman, even at the expense of not improving their raiding gear. Unfortunately with our current rules this possible because most of the raid is violently opposed to forcing items on people.
We considered this but eventually scrapped it because
1) It's a major hassle to track.
2) Token systems complicate it even further than it already is. Because certain gear is "cross-class" but really can only be bought by 2-3 specific classes.
3) The "cross-class" pool is really just off-spec people buying PvP gear the majority of the time. Sure there's that Fury warrior buying rogue gear, but the majority of the time it's just warriors buying PvP gear. Of course you could always prioritize by class even in the cross-class pool. Like I said though, it's pretty annoying to track everything in this kind of system. How do you decide between priority on Deathdealer (a pure upgrade for rogues) and Oracle (PvP gear for priests)?
We're working on something from the ground up, but we'll see where that ends up. We have some talented php/database guys in the guild, so I'm optimistic about it. EQDKP makes me want to stab myself in the face every time I enter raids.
Are you using eqdkp with the CT_RaidTrackerImport plugin? Once I got that installed and got people to get the version of raid tracker that works with it everything was much quicker. But we aren't capping max earned or doing upgrades (though upgrades wouldn't be terrible with it).
We want two pool because we do things like shard primalist legs for no reason other than everyone has ten storms and wants to save points for lifegiving gem/herald of woe/nac/whatever. Or we shard spineshatter because the dagger rogues are saving for something else and don't believe in spending points on a SS weapon they'd only use occasionally. I've tried to argue for an upgrade discount and sidegrade /random system but it was uniformly shot down by the other officers because it's "too complicated."
If you do write your own thing, please, please open source it. (And comment it.)
We're working on something from the ground up, but we'll see where that ends up. We have some talented php/database guys in the guild, so I'm optimistic about it. EQDKP makes me want to stab myself in the face every time I enter raids.
That would be like 3 visits to the hospital a year!
Gurg, do you use the CT_RaidTracker plugin? We have a setup where we basicaly just have to hit copy and paste and then paste the data into a text box and hit submit, works very well.
Yeah, I've been using Freddy's modded CT_RT and plugin for a while now. That part is easy.
Harder is tracking the 10-15 people in our standby channel as they come and go, culling people who are above instance caps, redoing the math each time, and then manually adding adjustments every time someone upgrades.
We recently came up with an interesting Solution to the problem of awarding dkp for wipes in a zero sum system. Over the course of each weeks raids (we have 4 official raids a week) we record all the dkp spent, then at the end of the week (Tuesday night for us EU raiders) we divide up the total spent evenly between each of them. What this means is that, if our week consists of: farm AQ up to C'thun, full BWL clear, C'thun attempts and another set of C'thun attempts, then the BWL clear is worth the same as a night of attempts at C'thun (even though no dkp was actually spent on the night of wipes).
This works really well in terms of persuading people to come to nights of wipes (and not giving such high reward to the people who just farm content). Though it does have one further change to the way points work, which is that people only gain their dkp from official runs once a week, in one lump sum (this isn't a huge problem but we had some complaints in the first week or so when people were still adapting).
I use a system somewhat similar to what Log described.
We use a null system, but instead of being null over one raid, its null over a week of raiding. Each raid is assigned an arbitrary weight (we use a complex system to determine raid weight, its beyond the scope of this description), and at the end of a week, a raid's DKP total is not determined by adding up all the items that dropped for just that raid, but instead adding up all the items that dropped during the week, and then allocating a share of the weekly DKP based on the raid weight. This lets you prioritize first kills, wipe-sessions, great raids, whatever. And you can also adjust things as you go along to make sure you earn the most DKP at the learning attempts, and you earn less at farming-sessions that net a lot of loot.
Very quick example, 4 raids.
MC Raid 1000 pts of items
BWL Raid 1500 pts of items
AQ Raid 1 500 pts of items
AQ Raid 2 0 pts of items
We might weight the raids as follows
MC : 1
BWL : 2
AQ1: 2
AQ2 : 5
So the weekly total is 2000 pts of items, and the weight total is 10. Therefore the raid values are:
MC : 200
BWL : 400
AQ1 : 400
AQ2: 1000
Then the raid pts are divided among the participants as usual. This ensures a perfectly null system, and allows for easy bonuses like first kills, great raids, on-time bonuses, etc. If you set all the weights equal to each other, you would have the system Log described.
Regarding the required weekly update and the "lump sum" problem that Log describes, I handle it like this: As each raid occurs, it gets entered into the system using traditional "null over one raid" dkp, and these raids are designated as "estimates" Everyone can see their estimated DKP after each raid, which gives a good approximation of your true dkp. At the end of the week, I "finalize" the week and apply my weighting scheme and re-value the raids from that week, and update everyones DKP. I previously did all this in EQdkp, which was a nightmare, but then I wrote a custom mangement solution myself and now I am a very happy player.
That's exactly the problem my guild is having now. We don't use sum-zero. The old officers gave boatloads of DKP for learning attempts and caused horrible inflation. We're having steady deflation nowadays, but new members (particularly warriors) will only catch up to our oldschool people sometime in 2012 at the current rate.
We're highly considering adapting the EJ upgrade system, but it doesn't eliminate our biggest problem of having such a massive disparity in points banked. For example: I have roughly 260 dkp and I've been around since day one with about 89% lifetime attendance. A new rogue who joined six months ago and has 95% lifetime attendance has -100 dkp.
Or consider Himeko, one of our warriors who joined about 9 months ago and has maintained 98% attendance the whole time. We consider him the United States of dkp because at this point, the concept of not being in debt is just silly to him. When we adjust using the upgrade system, he goes from like -150 to about +46. That's more in line with where he should be. It's obscene that such a reliable, contributing member could never dig himself out of a hole like that.
Things we're considering:
# Adapting the upgrade system and applying it retroactively.
# Adjusting the cost of equipment sets for classes that need less gear. (E.g. increasing the cost of all rogue armor by 30%)
It's a shitty situation to be in that I have no idea how to adequately fix without flushing it all away and starting over.
We, and I'm sure many guilds, experienced the same problem as you guys before AQ was released. Inflation given from the MC era / early BWL stuff threw any balance between old raiders and new raiders out the window. Our top DKP people had 600 DKP, and our recruits were easily in the negatives.
What we decided to do was pretty simple. We went back and removed all the inflation from the system that we could (On Time bonuses, First Kill bonuses etc). If your system is anything like ours was, we had bonuses that were more than an entire MC run, and in a game like WoW with member turnover being so high, it just doesn't work in the long run. The only inflation we kept was what we deemed "Necessary Inflation", for example On Time bonuses for bosses that the guild struggled on (Ragnaros for us). Other than that, even our old school high DKP members agreed that the fact that they were still benefiting from killing Gehennas 1.5 years ago was ridiculous.
Our simple but inelegant solution has been to introduce a new raidpoint database with every dungeon. Along with bids, it has minimized the effects of inflation, since new loot is given to whoever has put in the most work in that particular dungeon. In older dungeons, older members have an advantage in finishing out their gear setups because of the inflation, but there's still tons of almost-free loot going to new members.
Of course, it's clumsy, and will become even more clumsy as the number of raid dungeons moves beyond 4.
Our system accounts for a lot of the "concerns" people have pointed out.
In addition to DKP, we have incentive DKP which is awarding points for anything not related to actually killing a boss. For example, being on-time, any first boss kill, time spent learning encounters and homework if needed. Our homework is done in such a way that you can't do more and get more bonus out of it, anyone who wants to can do it and recieve the same IDKP and it's only available for say, one week.
At the end of every week we use an IDKP adjustment which subtracts IDKP from everyone and it's based on the average amount of IDKP actually earned that week. People who didn't raid would lose DKP, and people who raided would gain some IDKP.
This is to help combat inflation which can be controlled simply by changing the amount of the adjustment. It basically means someone who was inactive for a week will come back with less IDKP then they left with. It's highly possible if someone takes an extended break, they can come back with negative DKP.
Our DKP for buying items is your total DKP + IDKP so the IDKP is equally important as the DKP. To buy items in our system, they have set prices and the person with the highest dkp gets the item at that price. This is to prevent people from "saving" on auction systems and possibly having one person pay way too much while the next couple of people get the item for almost nothing. With set prices, everyone has to pay the same amount regardless and whoever is available to purchase it first is the person with the highest dkp. If two players are really close in DKP, like within 10, they do a /random between those 2 based on their dkp values. For example, /random 1-100 is person 1 and 101 to 205 is person 2 if the two players have 100 and 105 dkp respectively.
The items are priced in such a way that everyone through a zone is paying the same amount of DKP. This way you don't have one class paying 300 dkp and another paying 40 dkp to upgrade themselves in the same zone. If one class has to buy multiple items to be effective, each item might be cheaper then items for another class so the total DKP spent throughout the zone is the same.
We also provide discounts if a player chooses to upgrade or sidegrade their gear. For example a tier 2 breastplate would cost less if you previously had tier 1. If you have tier 2 and choose to buy the AQ sets, they would be discounted etc.
Non-dkp items are awarded based on attendance and we use a weighted system to track it. The past 30 day attendance holds the most weight, and decreases as you go farther back. It keeps track of every raid all the way back to 1999, but anything over 6 months is going to provide an extremely low value towards the total. The only reason it is done like that is so that if two players have equal attendance then the player who has been there longer has precedence. If someone has higher attendance in recent months, they would have a higher value over someone with lower attendance who has been around longer. It's used to award items that are beneficial to players with higher current attendance. For example: New spells, Keys for raid zones and related key quests. Things that we want in the hands of our recent high attendance members. It doesn't help us to award a key to a zone and then the player doesn't show up when that key could've gone to another player with higher attendance much more likely to show at the raid. That's why it is done by attendance and not DKP since DKP is more of total contribution and attendance tracks recent contributions.
The last thing we do, in terms of being fair, is that if a member shows up to a raid and can't join the raid because there are too many people then as long as they are available to raid if needed(we can contact them to join) they will recieve DKP/IDKP for the raid as if they were on it.
I'm sure I am forgetting some other stuff, but that's the jist of it.
I find it remarkable how often people try solve problems by adjusting their loot system - problems that have no business being addressed by the loot system.
Generally, you want your loot system to be as simple as possible - so that people trust it and understand it easily (not to mention you won't spend hours figuring out the administration).
Problems such as people looting items that they shouldn't, not turning up for learning raids, etc are *people* problems, they have nothing to do with your loot system. As such they should be solved by looking at the underlying problem and adressing it directly.
If folk aren't showing up for your learning encounters, instead of dusting off the DKP rules and writing new incentives, you need to figure out why. If it's because they hate farming for 5 hours during the week to afford the cost of the repairs, then do what Gurg and others have suggested - pay their repairs from the guild bank. If you have people that don't show up to anything that isn't farm status, ask them to leave the guild. Seriously. The proper solution to this problem is to post on your guild forum that you are recruting to fill raid spots for learning raids. Once your learning raids are full regularly, those that show up will get spots for the farm-encounters by virtue of their attendance record. This means the freeloaders are relegated to getting spots only when there aren't enough folk available - i.e. they will either quit or pick up their attendance. The long-term benefit to your guild is enormous. If it turns out that, after understanding the core problem, the right choice is to award dkp based on time spent on guild scheduled activities, rather than bosses killed - consider a system such as Log and Sirloin described. I can attest to the fact that their approach works really well in practise.
Early on our officers decided we needed loot priority - restrict loot to specific classes. This was a fairly contentious point and, as early as the start of BWL, it was abandoned for a couple of fairly important points. It got in the way. People had matured beyond the 'zomg it's purple' phase and had learned what appropriate looting was - and often it was counter to the written priority rules. Looting priority rules basically are a way of saying that the person who designed the rule knows now more about proper looting than you will in a year's time - this is totally arrogant and your players will come to despise you for it. You can never encode smart looting behavior into a loot system - don't even bother trying. Instead solve this problem by educating your players. When you recruit a new inexperienced player, buddy him up with an experienced player who can tell him what to loot. Watch how they take feedback and mature - if they cause loot problems and refuse to learn, you don't want to promote them to a core raider - you want to find someone else who will learn how to loot in the context of guild progression. Trust your players - your mature players are quite possibly smarter than you are - and they're definitely vastly smarter than your loot system.
If you find yourself seriously considering an amendment to your DKP rules to institute, hypothetically, a 50DKP penalty for getting punted into the whelps, stop for a second and consider what's really going on. If you're yet to defeat Ony, you probably just need more practise. If you have Ony down, consider getting better players, or telling players to pay attention and focus. And if you're squashing AQ already, ridicule the person mercilessly and taunt them on TS and leave them to kill all the whelps by themselves while the other 9 members of your raid finish her off :). The point is, none of this requires polluting your loot system with yet another complicated clause - the analogy being appropriate to more realistic scenarios.
Your loot system is simply one part of running your guild efficiently. Along with other aspects such as raid spot allocation and guild recruitment it can have a fairly large impact on the kind of guild you build. Ultimately, though, it is the people that make the guild, not the loot system. Pay attention to fixing the people problems and your loot system will become much less important. A good group of people can make *any* loot system work - even a broken one. However no loot system, however good, can compensate for bad people.
Acknowledge that loot decisions are made by people and that your loot system is simply a tool to automate the process in most cases. I'm a great fan of doing all looting communication in /ra - many guilds do it with silent tells - however the open communication allows people asking for loot to see who else is interested. This gives them exactly the information they need to decide if they'd rather pass or not, regardless of the players' relative current dkp.
I have to admit, I'm rather proud of the loot system we designed. It's incredibly simple zero sum - the details aren't really important, because the important part is the people. After more than a year of playing, people understand how to work with the system and it's limitations and get on with the game. I have been loot master for runs without access to the guild DKP database - and it simply didn't matter. Every item that dropped people knew who would get it, or decided among themselves, so that in the end the only job I had to perform was recording who looted what (CT_RaidTracker) and saying 'grats' :)
I absolutely 100% disagree with the post above. If you do not have an incentive-compatible loot system, then when your guild hits hard times (and you will at some point), then people will start to get lazy and try to exploit the rules in some manner, consciously or not.
One of the amazing things with wow (and to a lesser extent eq) is that there's a huge amount of "tier two" guilds with some of the stupidest dkp systems ever contrived. They'll defend it to the death because when doing MC or vex thal, incentive-compatibility really is no issue since it's pretty hard to wipe to golemagg or aten ha ra more than 3 times in your life. Thus it "works for them". However once they come to a mob that requires multiple days or multiple weeks of wiping, then bad dkp systems will quickly become apparent.
The fleixible smart guilds are the one that are doing tweaks and/or preparing for large changes with the expansion. Everyone who designed their own system knows that wow itemization threw us a curveball that we didn't design our systems for. It would be stupid to sit on it unchanged and think somehow it's the player's fault and not the system's.