All the talk about people restarting their RP systems after the expansion seems to be stimulating a lot of RP discussion. My personal topic of choice--how to design the upgrade policy in your zero-sum scheme.

For a long time now, we've been running a slightly complicated upgrade system where they key feature is that somebody looting two items in a slot pays a total number of points which is greater than the price of the more expensive item, but much less than the sum of the two prices. In general, you pay half price for things when upgrading. The idea was that this would have two advantages over a strict "pay the difference" policy:
1) Multiple items in the same slot do give your character more versatility than the person who'd only looted the most expensive one, and it would be nice to have this somehow reflected in the number of points a person has paid for their total complement of gear.
2) It discourages frivolous looting of intermediate-quality gear. There's still a cost to having looted many similar items. In fact, the way we've been running it, there's no priority in place to put items where they would present the largest upgrade--it's the perogative of the person with more points to decide whether to take something for half price, or pass it to a lesser-geared person. This would be harder to do if the person with higher RP would be getting his upgrade for very little.

Anyway, the problem arises when there are multiple (more than two) tiers of loot in close proximity. For example, if there are three successive items in the same slot which cost 400, 600, and 800, it doesn't work. There's no incentive to loot the middle item, because it does not serve as any kind of a down payment on future loot (which was the point of an upgrade system in the first place). It's even worse with the parallelism of BWL in AQ. In many cases, price progression in a certain slot might go 400-->600-->600, and nobody ever wants to possess both of the latter two items.

The way that Blizzard seems to be doing itemization, where loot is more of a continuum than a series of strictly stratified plateaus, probably means that this style of system is probably not going to work out in the long run.

---------------

So, I'll discuss "pay the difference" schemes (like the EJ system). My guess is almost everyone running an upgrade system will do it this way. One possible variant: instead of looting anything for free, there's a small minimum cost for any loot, no matter what you're upgrading/downgrading from. But I'll get into that later.

Once you've decided that your upgrade policy will be "1 + 2 = 2," there are two issues to discuss: pricing and priority.

Pricing + Upgrade paths
A huge advantage of this upgrade scheme is that balancing should be easy--you can look at someone's character sheet and estimate, almost exactly, how much they've paid over their lifetime. You can make sure that, after each tier of loot, the amount paid by each class is equal, but you need to have upgrade paths worked out.

The simplest way to do this is to ignore different items roles and just allow people to cross-upgrade between, say, healing and feral gear. But can you do any better than this? Or is even worth trying?
Say you made healing and feral gear (or tanking and DPS gear or something), into separate slots with no cross-upgrading, and priced things so the sum of the two items the Druid bought is the same as the one item another class is paying for. This seems like it would be difficult, due to the gray areas (which "slot" does Genesis go into?), but still, there's something unpleasant about lumping together completely different items and having people get what amount to significant upgrades for free.

How about resist gear? Lumping it into the main upgrade path doesn't seem like it's going to work. It will just be randomed for free, almost all the time, or else it will go to an undergeared person, who's probably a less active raider to begin with. But placing it outside the upgrade path makes cross-class balancing more difficult.

Ultimately, these questions are all related. In reality, gear is not one-dimenstional, and a person's total value (in terms of gear) is not determined solely by the most expsensive item he possesses in each slot (hybrid gear and resist gear are the best examples, but lots of classes fine-tune gear selection for different encounters). You have two options:
1) Ignore that fact.
2) Try to somehow account for the value of having multiple items, at least in the very stark cases of differing utility (but how do you draw the lines?).

Interested to see what people are planning.


Priority
This one has been discussed a bit more before. Who gets the drops first? Do you always go by point standings, or do you make a high priority of placing items where they're the biggest upgrade? In either case, there are a few chioices.

1) Always go by point standings.
Here, you have the obvious problem of people high on the list taking marginal upgrades, for very few points, over people who need them more. If your system allows for free downgrades, it seems like you at least have to draw a line there (so an item first goes to highest-RP person who's paying a nonzero price for it, then can be /randomed for free). In other cases though, can you make this work?

2) Put things where they're the biggest upgrade first.
This carries its own issues. First of all, after a while, people will be upgrading from a variety of different places, and the point standings will frequently be meaningless. Also, there comes a point where you don't want to give loot to someone just because they're the most undergeared person there. So you can wind up with a bizzare situation where the "middle" group get the best priority. For example, a T3 item drops, and one person present has a T2, one person has a T1, and one person has a blue. It's inefficient to give it to the person with the T2, but the person with the blue may well not deserve to have it defaulted to them, so your priority system might favor the person with the T1.
First of all, do you want this behavior?
Secondly, how you make this concept rigid when loot doesn't always come in clearly defined tiers?

Also, there are secondary classes to consider. For example, say all of your Warriors have Helm of Wrath, and Helm of Endless Rage drops. No Warrior is in a position to pay points for it, but some Paladins are. You'd still rather have it go to a Warrior. But if all your Mages have Netherwind and Mish'undare crops (and they cost the same), it's reasonable to want to give it to a Priest for points rather than to a Mage for free. How do you differentiate these two situations?

Again, interested to see what people think would work best.

This is why loot priority exists.

How can he get the damage items for upgrade over the mages/locks who are willing to pay full price?

My guild also uses a simple "1 + 2 = 2" "pay the difference" system (within a fixed value zero sum DKP system).

e.g. loot a tier 1 hat for 125 points, then when a tier 2 hat (priced at 150 points) drops, you can upgrade for 25 points.

Sidegrades and downgrades are free, but upgrades always take priority.

e.g. Nightslayer Cover drops. Rogue #1 has no epic hat yet, rogue #2 has Bloodfang but wants to grab NS too to mix'n'match set bonuses. Rogue #1 gets first priority regardless of their DKP standings.

There is no priority distinction between small and big upgrades.

e.g. Bloodfang Hood drops. Rogue #1 has Nightslayer now, rogue #3 has no epic hat. Priority is determined by their DKP, rogue #3 gets no boost for it being a bigger upgrade.

This has worked quite well, but here are some problems we have encountered.

1. imho our prices for different tiers of gear are too close together. A full tier 1 set costs 950 DKP, a full tier 2 set only costs 1150. So upgrading an entire tier 1 set to tier 2 costs roughly as much as looting your first two tier 1 pieces. Result: raider #1 could be only a small amount in front of raider #2 on the standings, from just doing a handful more MC raids, and then chain-loot an entire tier 2 set without dropping off the top of the standing. Solution: we should definitely have made bigger differences in pricing between tiers of gear.

2. Free downgrades and sidegrades (assuming nobody wants to upgrade and pay DKP for an item) encourages people to loot stuff they're never ever going to use, rather than have it sharded and add to our nexus crystal stocks. This is only a very minor problem. Solution: a minor tweak, like charging a minimal amount of DKP for all downgrades and sidegrades; alternatively, making people who downgrade or sidegrade donate a nexus crystal to the guild bank.

3. Upgrades taking priority over sidegrades and downgrades can mess up people trying to accumulate multiple gear sets. e.g. a warrior is trying to assemble a tanking suit and a DPS suit. We have Flameguard Gauntlets valued at 125, Gauntlets of Wrath at 150. If Flameguard Gauntlets drop for him first, he can loot them for 125, then next week, if Gauntlets of Wrath drop in BWL, he can upgrade to them for 25. However if loots Gauntlets of Wrath first, the Flameguards are now a downgrade, and he can't loot them unless there are no other warriors wanting to get them as an upgrade. Big differences in looting outcomes depending purely on the luck of what order items drop in = unhappy raiders. Solution: We instituted a rule that you can volunteer to pay full price for a downgrade or sidegrade, in which case you get normal DKP order priority. I'm not convinced this is a great solution, though.

Oh, actually, yeah, that's another problem we have encountered that I forgot to mention.

Raiding MC with new raiders = lots of epics looted, most of them taken by new people at full DKP cost = lots of DKP for everyone.

Raiding BWL with geared raiders = lots of epics looted still, but most of them taken as upgrades for a very small DKP cost = not much DKP for everyone.

An idkp-like approach we have used in the past (when we were first learning BWL) was to average out the DKP earnings for our MC and BWL raid for the week. e.g. MC gets mercilessly farmed, 2000 DKP is earned. Then we go to BWL, wipe a lot, kill a couple of bosses, loot stuff which is mostly upgrades. 400 DKP is earned. Rather than the MC raiders getting 50 each and the BWL raiders getting 10 each, we averaged it, called it 1200 per raid, and gave everyone 30 DKP.

In our system, it's simple: an upgrade is anything whose DKP price is higher than the highest value piece in that slot that you have looted already. A downgrade or sidegrade is anything whose DKP price is equal to or less than the highest value piece in that slot that you have looted already.

Drauk is suggesting that the priest with massive DKP lead is forcing the mages and locks to pay full price rather than wait for upgrade price. The priest is effectively leveraging his strong current position for future advantage.

The mages/locks are too shortsighted to stop wasting their points and now its becoming some sort of large problem apparently.

Makes me think of some strike breakers getting rid of the mage cartel on the waterfront. Depends on the system whether cartels and non compete agreements are desirable or not.

This seems like a people problem to me, the Priest should probably be told that it's better for the raid if he stops looting all the DPS pieces over Mages/Locks.

We use a zero sum system that has upgrade pricing built in too. Downgrades/Sidegrades cost 5 DKP, but you get last priority on the item, so that raiders who have not used DKP yet for that slot will get first shot at it. However, we made some exceptions, due to wonky item values from the point system that we adopted -- some Tier 2 items (such as Handguards of Transcendence) were less DKP than the Tier 1 pieces. (Gloves of Prophecy) We made it so that going from Tier 1 to Tier 2 was always considered an upgrade.

However, over time (9+ months of using this system) the following problems have slowly surfaced:

1) DKP inflation relative to new raiders. The long time raiders are all massively positive now. Our main focus is BWL, and new people who get in, have ZERO shot at getting any loot until every single person has the item who needs it. This is because we earn so little DKP each run due to almost everything being an upgrade price. I would say our average item price in BWL is ~80-100 DKP. In actuality, we're really only paying ~20-30 DKP or so for each item. So we have raiders such as myself, who have 400 DKP, and a new priest at 50 DKP just will never catch up to me, so I will end up in full Trans + all the bells and whistles (plus every other regular raiding priest almsot) before he even gets a sniff of it. The crux of the matter is you want your longtime and dependable raiders to have the best gear, but then new people effectively get shut out, and I'm sure it's very disappointing and frustrating to them. The short term solution is at least they have a MC still to go to in hopes of getting some Prophecy for the time being, but then they will really be in the hole more. (albiet at least better geared.) I'm not sure of an elegant solution to this though - new people into this system will always get the short end of the stick.

2) People at the top able to upgrade on anything they want, at the expensive of classes which are stuck at the bottom. I'll use myself as an example. Even as a shadow priest, fully specced for it, I've still been taking my healing gear (8/8 Proph and 3/8 Trans along with misc healing items.) I'm still ahead far enough of the mages/locks where I can pick up every single piece of DPS gear that I want. Someone else pointed out a situation very much like this. Solution? Have people who don't suck, and think of the raid. I still have yet to take a ToEP, or a mana igniting cord, etc. I repeatedly pass on them because the mages/locks need to be having a fair shot at it. Fortunately as the priest class officer, I can restrict what the other priests roll on too as there were several who thought MIC was a good priest belt. (it's fine, but wait for Girdle of Prophecy instead, and leave MIC to the mages/locks) You need people considerate and cognizent of the raid's needs. If they aren't, then they need a polite talking to about it, and the reasons behind loot priorities and restrictions. If they get really defensive and angry about it -- well, red flags go up for me that this person probably isn't going to be an ideal candidate for AQ or Naxx down the road as they are more focused on themselves than the raid.

3) People can't downgrade until everyone else has the item in question. We instituted a rule just as a previous poster wrote, in that you can override this rule if you are willing to pay full DKP for the item in question. For items that you truly want, this shouldn't be a problem. For small upgrades... it still isn't going to happen. Rogue A who has BF hood and wants NS cap still wont get it until all the other rogues who come on runs have the NS cap. IMHO this is fair until you are reaching the point where you are bringing new people along all the time simply so you stop excessive sharding. Solution: We have a minimum DKP earned eligibility requirement. It's not excessive; 2 full MC clears will bring them to eligibility. So the rogue who wants NS cap will get it over the new one. If nobody needed it as a downgrade/sidegrade, then the new rogue gets it. Generally the new people get 1 or 2 items off the bat anyways.

Overall our system was decent for a long time while MC was the main focus, and everyone still needed a lot of items. But I think this system starts to fail on several levels as the raid gets decked out. We're looking into instituting a completely new DKP system when Naxx comes out, as a trial run for the expansion. For the time being though, we're stuck with what we have, which requires a lot of oversight - such as the scenario that the OP wrote with Mish'undare, if it costs the same or less than Netherwind. IMHO Priests (at least this will be my call) will NOT be taking this item until all the proven A-team mages have had a shot at it. It's just common sense!


I'm not convinced there will ever be an elegant solution to awarding loot. There are always going to be special cases, and situations where someone is going to feel cheated out, or that the system is unfair. IMHO having good people both running the system, and good people in your raids go a long way in alleviating this.

An upgrade system is a terribly complicated thing to impliment, but is also necessary as skipping tiers is obviously out of the question if you want your guild to progress ahead of the nerfing curve.

My guild does not do this specifically, but I will be suggesting it come Naxx because the current system we have really hurts new people (and is inflated, but that's a different post). First, upgrades are treated essentially in the same fashion as pay the difference. The critical distinction is how new people get their gear, while retaining a sense for older members of "I'm a vet, and if I want a new item where I already have a purchased epic, that's my prerogative." Ashkandi drops and there are 2 sword specced arms warriors in the raid who could use it. Warrior A has 200 dkp and an Untamed Blade. Warrior B has no purchased epic weapon and 150 DKP. Some systems give Warrior A the sword and say "well he has the most points" but that does not strengthen the raid quite as much as Warrior B getting the weapon. Other systems give Warrior B the sword and say "it will benefit the raid more to have 2 warriors with epic swords than having one warrior with 2 epic swords." Unless Warrior A grows 2 more arms and starts dual wielding 2 handers, he's shit out of luck, completely unjustifiably. He has more points, has taken less from the raid than he has put in, and he really wants this somewhat rare drop NOW.

What to do?

I've seen some systems like the one I'm about to lay out, but there's a slight twist to prevent the "I'll loot anything cause I don't care about my points" people. If Ashkandi drops in the scenario above, valued at 150 dkp, Warrior A would be paying a 50 pt upgrade from Untamed (100 dkp), and Warrior B would have to pay the full 150. Warrior B gets first dibs if Warrior A is only willing to pay the upgrade price of 50. If warrior A really wants the sword however, as mentioned above, he can elect to pay the full 150 dkp, forgoeing the upgrade price, and excercising his superior point count. However, this is only viable if there is a limit to how often Warrior B can challenge Warrior A for the sword in the first place, otherwise the upgrade system will only work on trivial and situational items. If Warrior B cannot pay for the sword without going into negative points (ideally zero sum this is the point where he will have taken out of the raid as much as he has put in), then the whole deal goes out the door and Warrior A can pay upgrade price without worrying. This may encourage some light point whoring, but if Warrior A is hell bent on having Ashkandi, then sure as shit he's going to let Warrior B know in warriorchat that he should just back off and start making his character better and take the bloody upgrades. So the resolution to above is Warrior A has to either pay full points for the sword (which in turn makes the standings change dramatically for warrior loot order) or he has to relent to Warrior B who can pay for the sword without going negative. Now the system sucks a bit if Warrior A only kinda wants the sword and Warrior B has 149 dkp, but stuff like that can usually be worked out within reason.

That was a somewhat dramatic example, as no arms swords warrior would pass on Ashkandi if the price he pays suddenly tripples. But in the case of a warrior wanting to replace his 150 dkp Blessed Qiraji Bulwark with a 160 dkp Elementium Reinforced Bulwark (damn you chromaggus), he would have to think long and hard about taking the upgrade over a warrior using a draconian deflector with less dkp, but still >/= 160.

Wait, you price the Chromaggus shield above the AQ shield? :unsure:

Anyway, lots of interesting thoughts here. Various responses and comments:

New players: On some level, coming in as a brand new player once a DKP system of any kind has been rolling for over a year, with no gear to speak of, few systems are going to offer immediate reward, nor should they. I'm all for fairness and parity, but if you allow a newcomer to the system to come in and tag along at the top-tier instances that your guild is currently doing and take items over the veterans without whom that player would have no place even being in that instance in the first place, you're making a mistake. The inherent benefit to newcomers in a developed system is that they will get defaulted tons of loot. It took me 6 months of heavy raiding before I had epics in all eight of my main armor slots. Today I've seen newcomers accomplish that in 3 weeks. That's the advantage. Also, from a guild-efficiency perspective, you might argue that letting some guy go blue-->t2 is a bigger raid upgrade than letting someone else go t1-->t2. However, if you go the former route and then disenchant a t1 piece of loot because the formerly-blues-wearing guy has better and no one else in the guild needs it, that hasn't helped the guild at all.

That said, one of our main solutions in this area has been implementing caps on earning DKP from old instances. It's an administrative nightmare, but it helps people catch up. It slows the DKP growth of the high-end people, while simultaneously accelerating the gains of the new guys. These days, on an MC run there might be 10 people who are below the zone cap. That means that whenever anyone (even if it's someone not part of that 10) spends points to full-price an item from MC to replace their non-raid gear, those points are only split 10 ways. That makes MC runs much more lucrative than they were for the original crew that split the points 40+ ways. We had people who came into our system as brand-new raiders (either new additions to the guild or rerolled mains) in Aug/Sept, and today many of them are among the highest-DKP members of their respective classes, despite starting 6 months after the rest of us. Now, people who just started 3-4 months ago are still significantly further behind, but they're also very well-geared given the amount of time they've invested, so I don't really see the problem there.

Upgrade Priority: In MC and BWL, we gave people who had no epic in a given slot and were full-pricing strict priority over people who were upgrading. That made a lot of sense and was efficient in the long run. As we move forward into higher tier instances, we're moving towards something of a tiered upgrade system when it comes to certain items, since no one really has blues anymor. So if C'Thun drops a DEoI, a warrior with Ashkandi or Kalimdor's can't loot it over someone with a Rag weapon or Untamed. Also, with regard to the "brand new warrior takes Ashkandi over the vet" scenario, we place restrictions on one's ability to full-price loot if one lacks a certain amount of earned points -- it's a bit of an analogue to our cap system. In our system, you stop earning pre-Rag MC points at 1500, Rag at 2k, pre-Nef BWL at 2.5k, and Nef at 3k. On Nef, if you don't have at least 1000 Earned points, you get upgrade priority regardless whether you'd be paying full or not. The logic is that if you are that new, you probably have a dozen different upgrades in BWL. You don't need to be taking an item that might be the only thing a vet still needs from the zone. In AQ, that goes up to 1500. And so forth.

Resist Gear: Resist gear has its own separate path in our system, and usually costs 25 JP (where the average t1 set piece costs 100) regardless what someone has in that slot. The key point is that resist gear raises one's caps on zones, allowing someone to earn back the points they spend in the long run. Someone who spends 200 JP on FR and NR gear will keep earning JP from BWL up to 2700, whereas someone who doesn't buy any resist gear will cap out at 2500. It's an incentive system and works pretty well -- most people snap up resist gear when it drops.

Class Priorities: We probably have fairly tight priorities in place on our top-tier content, which then loosen as time goes on. Dam/heal stuff is exclusively the domain of mages and warlocks when we first encounter it, for example, which avoids the situation of the top-JP priest who snags all the caster loot unless told not to. Big 2h weapons go to warriors first, then others when warriors don't need them. AQ40 tokens are not allowed to be used for Oracle pieces, though once Naxx opens that will change. Boots of the Shadow Flame are now rogue/druid copriority since Deathdealer are out there now. Firemaw's Clutch goes to priests a lot these days. Claw of Chromaggus could go to priests or druids if the damn thing ever dropped, though originally it was mage/lock only. I think this system makes sense for optimizing guild progression while still allowing people to pick up the items they want for PvP, off-spec use, or whatever, in the long run.

Class Parity: Our current system screws melee classes. I disagree with Zellyn's suggestion of doing the normalization solely via weapons, because that just distorts incentives and the pricing structure within a given class's items. We're just applying band-aids to our system in the short run and hope to overhaul it for the expansion in this regard. In my eyes, the real problem is that melee weapons are simply better than caster weapons. They do more for the class. Going from an OEB to an Ashkandi is a much, much larger upgrade for a warrior than going from a Dominance to a Staff of Shadow Flame is for a caster, numerically. The flip-side of this, however, is that casters get more out of their armor slots. Spell damage is cheaper than attack power to produce the same DPS results. The proper way of balancing classes in the long run will be to add a small modifier to the price of all caster armor that adds up to the difference in expected costs of weapons between the two classes. So a warrior might pay 1000 DKP for a full suit of armor plus 400 more DKP on weapons, whereas a mage would pay 1200 DKP for a full suit of armor plus 200 more DKP for his weapon. As long as a full kit costs the same, it'll work out fine. Right now, that isn't the case however.

Hoarding: This can be a problem in any system, and while we want to allow people to pass when it makes sense to do so (passing to someone who could use the item more, someone who needs it for a set bonus, etc.), we don't want people to plan their loot orders to maximize their personal benefit at the expense of their peers. Yes, looting that Styleen's today instead of two weeks from now might mean that you lose priority on Ashkandi if it drops, but suck it up and loot it. No one is entitled to get any particular piece of loot first. If they stick around long enough, everyone will get pretty much everything they want eventually (or newer better items will come out and they'll get those). There is strong pressure not to pass on upgrades, and "yeah it's an upgrade but I don't want to spend that many points" is not a valid excuse for passing.

Well, in situations where individual pieces of Oracle are an upgrade (one case of Mantle of Prophecy --> Oracle shoulders) that's different, but by and large the Oracle set simply is not a raiding upgrade for a priest. Even the people who wanted some of the Oracle pieces admitted they wouldn't really ever use them in a raid zone, except for MC farming or something. Telling a rogue that he can't have his Deathdealer this week because a priest wants to farm more efficiently seems like an unwise course of action. A lot of our priests have 8/8 Trans or are very close to 8/8 Trans, and it's just too good to break up right now. Priests will reap the benefits when they get first dibs on most of the caster tokens in Naxx as a result, though.

Stormcaller for shamans is different -- the pieces may have a bit str/agi on them, but that really just replaces the resists on the old sets, and the pieces themselves are straight healing upgrades in every slot. With 5/5 Stormcaller I am a better healer than I was with Ten Storms in those slots, without question. Genesis is a great set for druids to wear when they need to offtank and to heal at various points during the same encounter. Only a couple of druids have been interested in it very much, but those who have make extensive use of it in raids.

Basically, I'm very sympathetic to nontraditional raid roles and such, and if someone has a reasonable argument for how something is genuinely useful to them in our high-end 40-man raids, I'm inclined to give them the benefit of the doubt. But even the people who wanted Oracle in this case had to admit that they weren't going to break up 8/8 Trans to wear a couple of pieces of Oracle pretty much anywhere in AQ or even BWL.