That makes more sense, thanks.

Actually, yes it most probably may be a problem. Quite the same (though, I suppose, less severe) than the one we actually have with badges and PvP rewards killing the incentive to progress in many, many guilds, and destroying a lot of them in the process.

It's a common human trait : if we can get it easily one way, the "hard way" lose a lot of its appeal. Reserving "signature items" for downing select bosses (I suppose the tier items will be one for each final Wing boss, and one for Kel'Thuzard, as it ends up being a nicely even distribution of the five pieces) give a strong motivation to kill them.

Really, seeing the damage that "farming badge rewards" has done to the game, I'm not too thrilled to see anything close getting too much part in the gearing process. Let's keep the symbols out of the market, and keep them only for rewarding progression.
That's more or less the idea I have. Badge items not being on par with the "real" items you get from the bosses (hence keeping the "bought" items less valuable and less prestigious), but allowing to catch-up in case of bad luck.

One token may be a bit boring (I don't really have issue with it, personally); but it's considerably less boring/bland than buying the set piece at the vendor, honestly. It keeps it being a prize, and not "grocery gear".

This is kind of how we ended up in TBC.

I'm in a guild that was raiding BT when I hit 70 a few months ago, and has just dropped the first boss in Sunwell this week.

With diligent work in the PVP realm, heroics and 10 mans I've equipped myself well enough for T5 raiding and entry-level T6 raiding. IE, while the badge items are good, they're still 1-2 tiers behind the progression raiders. It means I don't get to raid with them - only with alts doing beginning content that I currently overgear. It's hard to actually get player raid skills with an environment like this.

There is a fine line to be walked here. Instance-specific badges risk becoming rapidly irrelevant. Consider Honor Hold marks of honor, Soul Gems, Argent Dawn marks....if the scope of the badge is too narrow for the time it takes to earn enough to be useful, they're kind of a waste of effort to create in the first place.

So it appears we have NAX10 badges in both heroics and NAX. That's fine while most everybody is working on NAX. But lets translate that to BC and see where that gets us.


KZ10 - found in heroics and Kharazan
Gruul/Mags25 badges
Serpentshrine/Eye25 bdges
Hyjal25 badges
BT25 badges
ZA10 badges
SW25 badges

The problem here is KZ loot didn't progress you to being able to really do ZA, but they'll have 10 man versions without the big gap, so it may not be so bad.

A new 70 coming in late to BC you had two kinds of people to play with - KZ raids with maybe a little Gruul/Mags or BT/Sunwell raids. There is a big hole in the middle and when you cap out on the KZ stuff you have nowhere to get the next tier. The PVP rewards and badge loot can bridge the gap - I'm geared well enough now because of them to fill in in all the tiers except Sunwell and Black Temple, and could probably do BT in a pinch.

When we're starting out, this isn't an issue. There will be people working all the tiers and progression. As the game matures though, gaps may emerge. If there isn't some way to bridge the gap on content people are actually doing then progress can stall out for newer players.

I think the badge+drop scheme will work ok. As long as the final boss in each instance drops something everybody wants it's all good. The prior arguments that these items shouldn't be badgeable for motivation reasons I find pretty convincing. You'll kill through 12 bosses if the 13th contains something like the warglaives for each class.

It may be though that as WOLK matures, Blizzard will introduce again something that lets people skip content to "Catch up" to where the raiding guilds are playing, as the gap opens up between the guys who still are working on a Moros strategy vs those who are within a dungeon or two of the top end content.

This would more accuratley be shown as:
KZ / Heroics = BoJ
ZA / Gruul / Mag = Badge
SSC / TK = Badge
MH / BT = Badge
SW = Badge

This is because so far the understanding is:
T7-10 / Heroics = Badge
T7-25 / T8-10 = Badge
T8-25 / T9-10 = Badge
T9-25 = Badge

I never understood why the loot system in WoW works as it does.

One one hand, if a guild clears an instance for a few specific drops and gets them they won't come back. On the other hand, if a guild never gets the drops they want and new content comes out, they still won't come back. Even considering they're going to put an item like DST on a badge vendor is just kidding yourself.

I think the fundamental flaw in purchasing loot with badges is the fact that a system like badge loot even needs to exist in a tiered format. There are a ton of solutions Blizzard could implement into their loot system to remove the need for badge systems, even while retaining the relatively random nature of loot distribution. Burning Crusade's original implementation of Badge Loot was a fairly decent system but after ZA came out and they started adding everything they could think of to the badge vendor is when the idea went down hill.

I always looked at the badge vendor as an outlet for off-spec gear, or items that people never want to see drop. Off-set gear for classes that have very limited representation in raids (Boomkin gear, for example) should have been the direction the Badge vendor went, not "Oh, here are the best Rogue pants and boots in the game until 3/6 Sunwell. These will help you do Heroic Mech faster."

Badge gear is a good idea for filling in the holes that limited instance itemization creates, not compensating for the streaking and questionable RNG the instances have.

Because people who play Moonkin really enjoy having half as many drops as everybody else?

Why the guild and not the individual character? If you're going to restrict it in this manner, then certainly the logical thing to do is to restrict it to characters who have actually been present for a kill.

Care to elaborate why you think fixed price is wrong?

From my experience in my old guild (that was just over a year I guess, first ZG/AQ, then Kara and Gruul) we've had auctioning DKP and I've since then sworn to never again advocate that. I have to say it wasn't that bad in the 20man instances, we had hardly any loot drama there, but when we hit Kara, it was total madness. In a 10man instance we had such a change of people every day, every week it was totally random if you paid 10 DKP or 70 DKP for the same piece, just how the raid composition was that day. The healers and tanks generally got their stuff for very low prices, melee sometimes had to fight and casters had it worst of all.

With the "fixed price" I've been part of for the last year the only "problem" is that people at the top of the DKP list can take minor upgrades over new people with less DKP. But then again it would be the same with bidding. Except that maybe someone would offer his total 100 DKP to get a single piece, because he's not getting much loot anyway and the guy with 300 wouldn't pay 100 for a single piece, but that's just a wild guess.

I just don't see any typical situation where "auction" over "fixed price" wins to make everyone happy and still gives the overall guild gear level a boost to progress.

My issues with any Auction-class format are that it's gameable, and it encourages competition between, rather than cooperation among, the raid. It can work fine, just like most loot systems can, for sufficiently enlightened or cooperative groups, but I think it has a higher threshold of enlightenment to remain drama-free. Players getting greedy and trying to screw each other out of prize drops has a hugh drama-bomb potential, as does cartel-style behavoir for people not in on the cartel. Fixed-price DKP has the advantage of letting people plan for their upgrades by knowing the price, and how many people are ahead of them in line, in a deterministic fashion.

The main advantage of auction systems is psychological: items get priced by their desiredness, dynamically and probably accurately. Some people prefer this. Personally, I think modelling prices on a free-market system is a logical fallicy when the supply is absolutely fixed, but I will admit it allows prices to decay naturally as the drops get saturated among their optimal distributions, so long as people aren't being excessively retarded.

I'm curious with the Teired badge loot coming out for PvE what will happen with PvP gear. I'm not sure it's fully known yet, but PvP gear was always brought out comparable to average "progression" in PvE.
If there isn't Teired PvP gear then what will it be based on?? 10 man progression or 25?? If it's 25 I can see the same problems with PvE people having to PvP to gain rewards above their PvE progression level.

The argument against fixed price was that it results in sharded loot when no one wants to pay for a minor upgrade. When useful loot is sharded, the raid loses. An auction system never has this problem as long as it allows a bid of zero (with /random to break ties).

Obviously there are ways to solve the sharding issue in a fixed price system (officer overrule, whatever).

DKP is a rough measure of "Attendence - Reward". It's pretty hard to keep track of that kind of information for 30+ people without explicitly recording it somehow. It also has the advantage of being open and objective, compared to a loot council.

Another problem with fixed price is inflation. In my first raiding guild, we used fixed price but it can be pretty hard to find the right balance between incoming and outgoing dkp (you need to tune the prices just right). At some point, people who have been longer in the guild had a lot of dkp and it would take them a lot of items before a new recruit would be able to get anything. With auctioning, this is less of a problem.

There are so many types of DKP systems, but the only system that is used nowadays that in theory is really fair, is loot council. The reason for this is that current DKP systems do not consider "future drops". If pr0 item X drops for the first time, you cannot say what the value of this item is or will be. If some veteran member burns all his dkp to get this item, and then it happens to drop week after week, obviously he has paid too much for it because the item turns out to be not that special. Fixed dkp can solve this problem, but this can lead to "being lucky" on being on top. I remember our first Nefarion kill where Lok'amir dropped. I wasn't particularly planning on getting this mace, but I happened to be on top at the time and took it. If it were a bidding system I am sure someone who wanted it more would've gotten it. Too bad for them that since that first kill it never dropped again - I paid about as much as any Nefarion item was going for.
With loot council, the members of the council can use their own insights ("oh he had gotten pr0 item X but now half the guild has it so he deserves pr0 item Y"). You could design a dkp system that rewards players that already have pr0 item X to get some of their dkp back if it drops again (e.g. first time it drops again you get 50% of your dkp back, with or without a time limit and so on) and have both benefits of an open, "non-corrupt" system but still make sure you will not be screwed over by RNG. But it seems most people are pleased with the current systems that are being used.

First, I recommend using /random for 10 man raids and an auction for 25 man raids. The overhead that goes into handling updating points simply isn't worth the trouble if you are running two or even three 10 man raids a week. You're right that prices can fluctuate wildly if you let people use auctions in a 10 man setting, but that's arguably from one of two problems. Maybe the raid was poorly stacked from a loot perspective (one 10 man group has four people who want item X, the other 10 man group has zero people who want). That's partially the fault of the four participants and partially that of the raid leader. The other option is that people aren't correctly valuing passing their bid. They think of it as "50 points for the item or 0 for not-item", when it's really "50 points for the item this week, or 1 point for the item the next time it drops". Two stupid people will definitely cause one person to waste a lot more points than they should, but that's not the fault of the system. That's the fault of the top bidder, and they pay for by getting fewer good drops later.

Basically, /random is a fine system of arbitration when you generally have 1 to 3 people who want any given drop. Something more exotic like a virtual currency system (eg. points) only helps when you commonly have 5+ people competing for the same item.

Back on the subject of a fixed price system, the system is flawed on its very premises. Here are the three flaws, and a bunch of economic explanations for why they break down. Be warned, this is kind of long, but not that complicated.

Prices can't be fixed:

Fixed price systems are designed with the belief that each item has one fair price and you need to get it "just right" to avoid a point inflation (or deflation) scenario. That's not true. In basic economic theory, two things control the price people will pay for an item-- the supply and demand. An item's supply is its drop rate and its demand is the number of people in the raid who want to loot that item. The theory is that different people will have different maximum prices they will pay for an item. For example, a retribution paladin will probably pay more for [Shivering Felspine] than a Hunter will, because the hunter just wants the item as a stat-stick. When fewer copies of the item drop (supply decreases), the fair equilibrium price increases (to a price point that the ret paladin is willing to pay but the hunters are not).

See the problem with fixed price? Optimal price depends on the drop rate, and drop rate depends on your luck. If your guild loots a [Warglaive of Azzinoth] every week for three months but never sees a [Dragonspine Trophy], then the glaives really are worth fewer points than the trophy. The stat boost of the item isn't the only factor in pricing an item.

Additionally there's the problem of similar drops. Maybe you really want some offset helm, but it rarely drops. If you can use your T6 helm instead, demand for the offset helm will be much lower even with a low drop rate. But if the item isn't very replaceable, like a trinket, expect the price on that item to go through the roof the few times it does drop. Prices change over time, and fixed price systems fail when they ignore this basic economic fact.

There's one other reason prices can't be fixed. There is value in looting the first drop of an item. Think of loot as a consumable with a long duration, not a permanent upgrade. If I'm the first person to loot a piece of T6 gear, I might use that item for 6 months before upgrading it. One of the last people in my group might only use it for 3 months before it gets replaced with another item. Even if we upgraded from the same old piece to the T6 gear (same stat increase), I'm getting twice the value from the item, so the amount he's willing to pay is significantly less than my price.

Zero sum points aren't zero sum:

As you mentioned before, inflation and deflation scenarios are dangerous in a fixed price system. With inflation (points gained > points spent), the people on top of the ladder just take everything and positions don't change. With deflation (points gained < points spent), a lot of genuinely useful drops get sharded because people can't afford to pay for them. The typical solution is to "zero sum" the points. In other words, if someone pays X points and there are 25 people in the raid, everyone in the raid gains X/25 points for the item. (Or maybe if you have 5 people benched, X/30, or whatever.)

Having raided with a fixed price zero sum system for two years, I can explain exactly how the system warps in that time. You will bring in trials, not all of which will pass, and they will loot stuff. They'll end up in the negatives, adding more points to the core raiders, thereby inflating the average point value to something above zero. Similarly if someone is planning on leaving the guild, they have incentive to point dump, creating the same effect. People are far more likely to quit the guild with negative points than positive ones, which makes it really good to be in the raid where they spent their points.

Over two years time, our point system was almost perfectly stratified by the total time the person had spent in the guild. The long term members were always on top, new people always on the bottom. When you average the points each raider brings to a raid and end up with a number above 0, you are effectively in an inflation scenario after all. We tried adding exponential point decay (which actually gained points for people in negatives) and higher prices for people who had more points banked (so highest points paid 125% of price, lowest paid 75%), and it still didn't change anything. Zero sum points never summed to zero.

Rare items can't be evenly distributed with fixed prices:

In my mind, the biggest problem with fixed price priority system isn't so much the fixed price, but the priority. A priority system is fine for most items, but for the high demand items, it tends to give all of them to the same 20% of people at the top of the ladder. Here's why:

There are two basic kinds of items that drop in raiding: common and rare. A common item is something that everyone in the raid who wants it will get it, given a reasonable amount of time. Rare items are things that not everyone can guarantee they'll get no matter how long they raid. For example, a T6 token is common and [The Skull of Gul'dan] is rare. In 10 weeks of raiding, you'll get 30 copies of a given T6 token, pretty much gearing up the entire raid. In that time, you'll average slightly under 2 skulls, even though 6 to 10 people in your raid will want it. Furthermore, due to random variance, it's possible you'll spend a year raiding Illidan and still only get 2 skulls to drop in that entire time. (That's exactly what happened to my guild, incidentally.) A rare item is one for which it's not reasonable to farm until everyone has it.

Common items are largely irrelevant in a loot system. Since everyone will get the item, all that really matters is if you get it first or last-- the actual price you're willing to pay is low in either case because the item is a certainty. The whole point of a good loot system is to evenly distribute the rare drops.

Here's the problem. I fundamentally believe that rare items should be evenly distributed by contribution to the raid (eg. attendance, or whatever else you're measuring). If three rare items drop that two people want, and the first person has double the attendance of the second, the first person should get two rare items and the second person should get one. They should not be split three to the first person and zero to the second, and that's exactly what a priority system will do.

This occurs because fixed prices never account for the rarity of the item (low supply), only the stat upgrade. And you can't even say "oh, the pricer should have realized Skull of Guldan is low drop rate, high demand item and priced it at 10x what common items cost". That's because the actual price depends on whether or not you get an artificially low drop rate (like 2 drops in 52 kills instead of 8 or 9 like expected). It's simply not possible for a fixed price system to be fair.

Its possible to fix this problem, at least, by changing the way you calculate points. Assign each item a value in item points. Each raid gives a certain number of attendance points. Then, you can calculate the value of an attendance point by dividing the total value of items looted by the total number of attendance points in the system. When someone leaves the guild, you just delete their row (both item and attendance points) from the spreadsheet.

We use straight DKP with offspec /randoms.

We chose this primarily to minimize drama. If you want more loot, make more raids. We once had to threaten to charge half the DKP cost for any clear upgrades you passed on to limit DKP hoarding.

In the example above, if two people want three rare items and one raids twice as much, from an individual standpoint, yes, the person that raids twice as much should get two items and the person with less attendance should get one.

However, from a global view, it is better for the raid to have those three rare items being used in every raid helping you progress than to have them logged out in IF for half your raids.

Getting back to the original question, I'm starting to think that a DKP penalty for any clear upgrades passed up may be the way we go. If you are able to justify passing, the DKP penalty can be lifted, otherwise, there's a DKP price to be paid for hoarding.

As much as I hate to contribute to this thread, I feel the need to point out that this post is stupid.

The criteria on determining "a clear upgrade" is completely subjective. If an item has 1 stat point upgrade, is it a clear upgrade? Most would say no. If it is higher itemlevel and itemised similarly, but with a few points in a wasted stat (say, hit when you're at the cap already) is that a clear upgrade? This is subjective because you don't know if you are going to get x item which will make that stat useful again.

You're still reliant on someone drawing a line to say whether an item is a clear upgrade or not (or coming up with some criteria that defines this), effectively making it DKP-with-loot-council. This may penalise someone for not picking up something that really isn't good for them, just because, by your criteria, it satisfies the point at which you should be penalising them, or because whoever is making the decisions is also affected by the ruling and exploits it to gain an advantage.

I respectfully disagree with you, and though I don't agree with you, I would hardly come out and call your opinion "stupid". You complain that the "clear upgrade" criteria is subjective, but there is very little that is more subjective as determining what is "stupid" or not.

Fact is, you make your case. I'm not talking about sidegrades. If you are in blues and a piece of your set drops, then yes I need you to tell me why it isn't an upgrade for you. This isn't a government audit. It's a group of people who chose to raid together trying to kill stuff. My raiders just talk to me. If you have a rational reason then great. If the reason is that you are passing to get something else off the end boss then that's not a valid reason to pass.

However, your opinion is duly noted.

I thought I'd explain the thought process behind our loot system or lack thereof. Start with the assumption that any loot system can be gamed; even with /random, an unscrupulous raider can convince others to roll or not roll to his advantage. If a loot system can't be perfect, at least it can be easy. We'd rather address the problem of why a raider is trying to get these items in the first place than just preventing the symptoms.

So we just /random most items, and people choose to roll or not roll as they see fit. It's not a perfect system - we've had a few items looted that in retrospect could probably have gone better to other people - but even those mistakes have not caused any substantial amount of drama that would justify the overhead of an actual loot system.

TL;DR Deal with loot whores by dealing with loot whores, not by hiring an accountant to run your DKP

tedv: Just because you have items being supplied, does not mean that you have "supply" in the true economic sense. In order for economic theory to apply, the supply has to be able to increase to meet demand. While we use the same terms, pricing is just determining the order that you hand out items, not the amount being handed out (except in the edge case of having enough raiders to run an extra instance). Economic concepts should be applied with caution, although they can apply.

The price decreasing over time can be handled without changing the literal value of the weapon. If later bosses give more DKP, and later loots cost more DKP, then the "real value" of the fixed DKP cost degrades over time. Since the real cost of the loot is the opportunity cost of what else you could have picked up instead, this isn't just coincidence, it reflects the lessening of quality.

My biggest issue with auction systems is that they reflect desire, but not utility. A well-founded auction system relies on people thinking objectively, and using enlightened self-interest, but rules-based systems in general tend to promote selfish greed about maximizing loot per points rather than maximizing attendance. This is true of DKP as well, obviously, but an auction system puts people in conflict (not just competition) with each other, which I find tends to be detrimental for morale (read: drama-prone). If your guild group is awesome enough not to fall to such levels... well, in that case rules-based systems themselves are redundant. The entire reason to have a system for apportioning loot, instead of just handing it out, is to put on a veneer of objectivity to prevent intercene conflict. If you can do that without the system, there's no need to give yourself administrative overhead.



Tangent: If you want a theoretically well-founded method of apportioning loot, do some work into the fair division problem. In short, it's about partitioning a finite amount of resources amongst a finite number of players with differing utility functions for the resources at hand. It would necessitate having some datamined drop rates and everyone stating their (relative) preference for various loots, and then having normalized and from them generate either a loot assignment that gets updated with each drop, or a continuous partitioning of loot that basically weights the rolls people are allowed to make (ie a 2:1 partioning designates a /random 150 vs /random 100 for 2:1 victory odds). I don't think there's been much research done into probablistic variations, but it's a good place to start. Of note are envy-free and efficient (Pareto) divisions.

All of these threads end up being e-peen contests eventually and this one is no different. Whatever loot system works best for your specific guild/group is up to your specific group. If you are unhappy about your current loot system, look up the other standard loot systems used and try one on for size.