Everything I ever needed to know, I learned from World of Warcraft: Incentives and rewards

Link post

This is the sec­ond in a se­ries of posts about les­sons from my ex­pe­riences in World of War­craft. I’ve been talk­ing about this stuff for a long time—in fo­rum com­ments, in IRC con­ver­sa­tions, etc.—and this se­ries is my at­tempt to make it all a bit more leg­ible. I’ve added foot­notes to ex­plain some of the jar­gon, but if any­thing re­mains in­com­pre­hen­si­ble, let me know in the com­ments.

Pre­vi­ous post in se­ries: Good­hart’s law.


“How do we split the loot?”

That was one of the biggest challenges of raid­ing in World of War­craft.

We’ve got­ten 40 peo­ple to­gether; we’ve kept them fo­cused on the task, for sev­eral hours straight; we’ve co­or­di­nated their efforts; we’ve figured out the op­ti­mal strat­egy and tac­tics for tak­ing down the raid boss; we’ve ex­e­cuted flawlessly (or close enough, any­way). Now, the dragon (or de­mon, or sen­tient colos­sus of mag­i­cally an­i­mated lava) lies dead at our feet, and we’re star­ing at the fab­u­lous trea­sure that was his, and is now ours; and the ques­tion is: who gets it?

The prob­lem of the indivisible

“Why not di­vide it 40 ways? That’s only fair!”

If only it were that easy! But the loot can’t be split 40 ways, be­cause it’s not just a gi­ant pile of gold coins; it’s (for in­stance): a magic warham­mer, a magic staff, and a magic robe. Three[1] items; each quite valuable and de­sir­able; each of which can be given to one per­son, and can­not be sold or traded there­after. We have to de­cide, here and now, which three out of the 40 mem­bers of the raid will re­ceive one of these re­wards. The other 37 get noth­ing.

What to do?

It would be difficult to over­state how much thought went into an­swer­ing this ques­tion, among WoW play­ers; how much effort was spent on de­bat­ing it; how much ac­rimony it spawned; and how crit­i­cal was a good an­swer to it, in de­ter­min­ing suc­cess in the most challeng­ing en­deav­ors in the game (high-end raid­ing). Disagree­ments in mat­ters of loot dis­tri­bu­tion broke friend­ships, and ill-ad­vised loot poli­cies cracked guilds in half.

Nor should this be at all sur­pris­ing, for World of War­craft was hu­man civ­i­liza­tion in micro­cosm. The task of or­ga­niz­ing and co­or­di­nat­ing a raid was pro­ject man­age­ment; figur­ing out how to take down a raid boss was strate­gic plan­ning, and prac­ti­cal episte­mol­ogy, and math­e­mat­ics; mak­ing money at the Auc­tion House (in or­der to pur­chase perfor­mance-en­hanc­ing po­tions) was eco­nomics. But the dis­tri­bu­tion of loot—that was poli­tics.

Pie-slic­ing as pro­ject management

To the ques­tion of “how to split the loot”, there is at least one sim­ple an­swer: a roll of the dice.[2] This method was used of­ten—in “one-off” raids, wherein a group of play­ers would come to­gether, for this one oc­ca­sion only, to defeat a challenge (or a con­nected group of challenges), and then dis­perse, en­ter­ing into no longer-term re­la­tion­ship with one an­other, pos­si­bly never to co­op­er­ate or in­ter­act again. In such cases, it was un­der­stood that your par­ti­ci­pa­tion comes with the promise of a chance at win­ning a re­ward. If the group hap­pens to find some piece of trea­sure which you covet, you will have an op­por­tu­nity to throw your name into the hat, and—should the Ran­dom Num­ber God smile upon you—to win the prize. That’s all you can ex­pect, and it’s no more than any­one else is get­ting. In a tem­po­rary col­lec­tive, made up of strangers, one can hardly ask for more.

But such one-off groups are nowhere near effec­tive enough to tackle the most difficult challenges that are available—that is, to do “pro­gres­sion raid­ing”[3]. For that, you need for the same 40 peo­ple to as­sem­ble, week af­ter week, month af­ter month; they must learn to work to­gether smoothly, and they must all, to­gether, learn the raid en­coun­ters, and the strate­gies and tac­tics for defeat­ing them; and, just as im­por­tantly, all mem­bers of the raid must “get geared”—must ac­quire bet­ter and bet­ter equip­ment for their char­ac­ters, in or­der to im­prove their perfor­mance and be­come bet­ter able to han­dle the next challenge, and the next. This is a sus­tained, col­lec­tive effort, and it can only be man­aged by a per­sis­tent or­ga­ni­za­tion: the raid guild.

The challenges of run­ning a raid guild are le­gion; there is much to say about them—enough for many more blog posts. For now, the key point is this: for a raid guild, the ques­tion of loot dis­tri­bu­tion is, at once, both a se­ri­ous and thorny prob­lem, and a pow­er­ful tool which may be ap­plied to many other as­pects of guild man­age­ment.

There were many “loot sys­tems”. Some were com­mu­nis­tic: the raid leader (or a “Loot Coun­cil”, com­posed of the raid leader and a small hand­ful of oth­ers) would sim­ply de­cide which raid mem­ber would re­ceive each piece of loot—“to each ac­cord­ing to his need”. Others were at the op­po­site ex­treme—“free mar­ket” sys­tems, where one ac­cu­mu­lated “points” via raid at­ten­dance and con­tri­bu­tions to the raid’s suc­cess, and the al­lo­ca­tion of loot was de­cided via bid­ding.

Nei­ther ex­treme ever sat well with me (for rea­sons that should be ob­vi­ous enough to any­one with any pass­ing fa­mil­iar­ity with the real-world eco­nomic sys­tems to which I al­luded). When the guild to which I be­longed de­cided to get se­ri­ous about pro­gres­sion raid­ing, and it came time to for­mu­late a policy for loot dis­tri­bu­tion, I ad­vo­cated for a loot sys­tem which, to this day, I con­sider the most ideal, of all the sys­tems I’ve en­coun­tered. The sys­tem was adopted, and it served us well for years to come. That sys­tem was known as the “Effort Points /​ Gear Points” sys­tem, or EP/​GP.

Solu­tion: EP/​GP

The idea of EP/​GP was sim­ple. There are some ac­tions/​be­hav­iors that you don’t want your mem­bers to en­gage in at all; those you ban out­right, and set what­ever pun­ish­ments you see fit. Those aside, how­ever, there are two cat­e­gories of things that aren’t dis­cour­aged:

First, there are things that you want ev­ery­one to be do­ing as much as pos­si­ble—things that are un­bound­edly good. Se­cond, there are things which are good for peo­ple to be do­ing, healthy, ex­pected, cer­tainly not dis­cour­aged—but you don’t want any­one do­ing them too much, and you don’t want there to be a se­ri­ous im­bal­ance in who is do­ing those things.

The sec­ond cat­e­gory con­sists of things which peo­ple just want to do, of their own ac­cord, and don’t re­ally need to be in­cen­tivized to do; they’re their own in­cen­tives. The first cat­e­gory con­sists of things that you do gen­er­ally need to in­cen­tivize peo­ple to do, even if peo­ple “want” (or want to want) to do them. (In WoW, the first cat­e­gory is “help the raid kill bosses” and the sec­ond cat­e­gory is “get gear, thus mak­ing your char­ac­ter more awe­some”.)

The idea of the EP/​GP dis­tri­bu­tion sys­tem is that you use the first cat­e­gory to rate-limit the sec­ond. Each mem­ber has two quan­tities as­so­ci­ated with them: EP (effort points) and GP (gear points). Both start at 0; each goes up as a con­se­quence of ac­tions/​be­hav­iors in that cat­e­gory. Do un­bound­edly-good proso­cial thing? Your EP goes up. Do self-in­cen­tiviz­ing in­di­rectly-proso­cial self-benefit­ing thing? Your GP goes up. And when­ever there is any scarce re­source that peo­ple want, it is al­lo­cated ac­cord­ing to EP/​GP ra­tio; who­ever has the high­est such ra­tio gets the re­source (and their GP goes up ac­cord­ingly).

So, the more EP-gen­er­at­ing things you do, the higher your pri­or­ity in the al­lo­ca­tion of re­wards; the more GP-gen­er­at­ing things you are al­lo­cated, the lower your pri­or­ity sub­se­quently. As long as you can define those cat­e­gories, and place rele­vant be­hav­iors/​ac­tions into them, EPGP works to al­lo­cate your scarce re­sources and in­cen­tivize mem­bers’ con­tri­bu­tions.

The EP/​GP sys­tem has a num­ber of an­cillary benefits:

First, new mem­bers im­me­di­ately get the in­stant grat­ifi­ca­tion of be­ing top pri­or­ity for re­sources, as soon as they con­tribute any­thing what­so­ever (as EP and GP start at 0, any con­tri­bu­tion makes EP pos­i­tive, and pos­i­tive/​zero = in­finite pri­or­ity!) Hav­ing now given and got­ten some­thing of value, they are drawn in, at which point their pri­or­ity goes down to be­low that of reg­u­lars/​vet­er­ans; it fluc­tu­ates greatly at first, then sta­bi­lizes. This in­cen­tivizes early con­tri­bu­tion, but doesn’t make peo­ple “pay dues” ex­ces­sively to get any­thing at all.

Se­cond, be­cause EP can be as­signed for any­thing, and rel­a­tive val­ues set to what­ever the ad­minis­tra­tion wishes, the sys­tem makes it easy to de­sign in­cen­tive struc­tures that en­courage what­ever you like

Third, there is tan­gible benefit to sus­tained con­tri­bu­tion, with­out lock­ing out new­com­ers.

A con­trast: DKP

“Dragon Kill Points”, or “DKP”, was, once upon a time, the most pop­u­lar loot sys­tem; it long pre­dated EP/​GP.[4] In DKP, each mem­ber of a raid would re­ceive some num­ber of points (the titu­lar “dragon kill points”) upon suc­cess­ful com­ple­tion of a raid en­counter. To re­ceive a piece of loot, a raid mem­ber had to spend some of his points (the amount usu­ally de­ter­mined by a bid­ding con­test among all raid mem­bers who wanted that item).

DKP had many faults, and waned greatly in pop­u­lar­ity as WoW aged; bet­ter sys­tems (such as EP/​GP) had come along. With DKP, if you were a new­comer, join­ing a raid full of vet­er­ans, the only way you were go­ing to get any­thing was if no one else wanted it—oth­er­wise you had to toil through raid af­ter raid, con­tribut­ing effort but know­ing in ad­vance that you weren’t get­ting any­thing for your efforts (ex­cept scraps from the vet­er­ans’ table, as it were).

What makes a good loot sys­tem?

The twin needs, in any group that de­pends, for suc­cess, on a bunch of peo­ple all con­tribut­ing as much effort as pos­si­ble, are:

  1. You have to pull in good peo­ple;

  2. You have to get your peo­ple to stay, and keep con­tribut­ing.

DKP was bad at #1, be­cause the prospect of slav­ing away for weeks or months be­fore you had ac­cu­mu­lated enough to have a shot at the good stuff was daunt­ing (and then you could be out­bid by a vet­eran who’d been hoard­ing his points for longer; and even if you won, you might’ve just spent all your points on one thing; etc.). DKP was also bad at #2, be­cause af­ter you had ac­cu­mu­lated a cer­tain large pool of points, the in­cen­tive to keep con­tribut­ing dropped off.

EP/​GP, on the other hand, is good at #1, be­cause your first re­ward is ba­si­cally guaran­teed, as soon as you con­tribute some­thing of value. And EP/​GP is good at #2, be­cause go­ing up in pri­or­ity is easy at the start, and bounc­ing back from get­ting some gear is easy, and as it gets harder, well, ra­tios equal­ize; and as long as you keep con­tribut­ing, you stay at a good ra­tio, and mean­while, the higher your EP gets (if you’re a vet­eran), the faster it drops when you get some­thing.[5]

Is EP/​GP re­ally the best way?

Later in WoW’s his­tory—when it be­came pos­si­ble for high-end raid en­coun­ters to be tack­led by smaller raid groups—there arose, within some raid guilds, the prac­tice of hav­ing mul­ti­ple raid groups, in­clud­ing some that were more ‘elite’/​ex­clu­sive than the guild’s main raid group. (Mem­bers of such smaller groups typ­i­cally par­ti­ci­pated in the sub-group’s raids in ad­di­tion to tak­ing part in the guild’s pri­mary raid­ing ac­tivi­ties.) Where a good raid guild might’ve been in the 99th per­centile of com­pe­tence and achieve­ment, among the over­all player pop­u­la­tion, it might have within it a smaller raid group which was much, much fur­ther to­ward the right tail of the raid con­tent achieve­ment dis­tri­bu­tion.

Such smaller sub-groups usu­ally did not use the EP/​GP or other al­lo­ca­tion sys­tem of the main group, but had their own, sep­a­rate, loot poli­cies. Th­ese poli­cies typ­i­cally skewed closer to “man­aged com­mu­nism” than to “reg­u­lated cap­i­tal­ism” on the spec­trum of loot sys­tems; and I do not think that this is a co­in­ci­dence. The mem­bers of these smaller, more ex­clu­sive groups—which, in virtue of their greater se­lec­tivity for com­pe­tence and perfor­mance, al­most always performed bet­ter and ac­com­plished more difficult goals than a guild’s pri­mary raid­ing group—ex­hibited a higher de­gree of sub­li­ma­tion of per­sonal in­ter­est to group in­ter­est, than did mem­bers of a guild’s main raid group; they were more will­ing to make sac­ri­fices “for the good of the raid”.

If you’re try­ing to main­tain a raid­ing guild of 100 peo­ple, keep it func­tion­ing and healthy over the course of months or years, new con­tent, peo­ple join­ing and leav­ing, sched­ules and life cir­cum­stances chang­ing, differ­ent per­son­al­ities and back­ground, etc., then it’s im­por­tant to main­tain mem­ber satis­fac­tion; it’s im­por­tant to en­sure that peo­ple feel in con­trol and re­warded and ap­pre­ci­ated; that they don’t burn out or de­velop re­sent­ments; that no one feels slighted, and no one feels that any­one is fa­vored. You also have to re­cruit new mem­bers, to keep up with in­evitable mem­ber turn-over. All of these things are more im­por­tant than “be­ing max­i­mally effec­tive at tak­ing down this raid boss right now, and then the next five bosses this week”. If you fo­cus on the lat­ter and ig­nore the former, your guild will break and ex­plode, and peo­ple on WoW-re­lated news web­sites will place sto­ries about your pub­lic melt­downs in the Drama sec­tion, and laugh at you.

On the other hand, if you get 10 play­ers to­gether, and you say: “OK, dudes—we, these par­tic­u­lar 10 peo­ple, are go­ing to show up ev­ery sin­gle Sun­day for sev­eral months, play for 6 hours straight each time, and we will push through ab­solutely the most challeng­ing con­tent in the game, which only a small hand­ful [or some­times: none at all] of peo­ple in the world have done”—that is a differ­ent sce­nario. There’s no room for “I’m not the tank but I want that piece of tank gear”, be­cause if you do that, you will fail.

What a group of the lat­ter sort promises—which a larger, more skill-di­verse, less elite/​ex­clu­sive, group can­not promise—is the in­cred­ible rush of push­ing your­self—your con­cen­tra­tion, your skill, your en­durance, your co­or­di­na­tion, your in­ge­nu­ity—to the max­i­mum, and suc­ceed­ing at some­thing re­ally re­ally hard, as a re­sult. That is the in­trin­sic mo­ti­va­tion which takes the place of the ex­trin­sic mo­ti­va­tion of “get­ting loot”. As a re­sult, the ex­trin­sic mo­ti­va­tion is no longer a re­source which it is vi­tally im­por­tant to al­lo­cate. In that sce­nario, your needs are the group’s needs; the group’s suc­cesses are your suc­cesses; there is no sep­a­ra­tion be­tween you and the group—and con­se­quently, the need for equity in loot al­lo­ca­tion falls away, and ev­ery­thing is al­lo­cated strictly by group-level op­ti­miza­tion.

This was ev­i­dent in the re­ac­tions peo­ple had, to see­ing other peo­ple get loot. In a larger, some­what-more-ca­sual, raid group, it went like this:

Alice gets [awe­some piece of gear].

Alice: yay! :D

Bob: grats

[Bob is happy for Alice but also jeal­ous, Bob wanted that thing too.]

In tighter-knit, more “hard­core” groups, it was more like this:

Alice gets [awe­some piece of gear].

Every­one in the raid: F***| YEAH!! :D

[Every­one is ec­static that Alice got that thing; no one is jeal­ous.]

In the lat­ter case, it was not only un­der­stood, but viscer­ally felt, that ev­ery thing that any­one in the raid gets, is in­creased perfor­mance for the group as a whole—which is all that mat­ters.[6]

[1] It wasn’t always three, of course; some­times one, some­times four, etc.

[2] WoW came equipped with such a fea­ture; one would type /​roll 100 into the chat win­dow, the server would gen­er­ate a pseudo-ran­dom num­ber in the range [1,100], and would out­put the re­sult into the chat, for all raid mem­bers to see. Thus, ev­ery­one could roll for a piece of loot, and the per­son with the high­est roll would re­ceive it.

[3] See the pre­vi­ous post in the se­ries for a defi­ni­tion of this term.

[4] DKP is one of the old­est loot sys­tems; it was used even be­fore World of War­craft—in older MMORPGs like EverQuest.

[5] Most EP/​GP im­ple­men­ta­tions also in­cluded a “de­cay” fea­ture—which pe­ri­od­i­cally (ev­ery week, or ev­ery month, or similar) re­duced all EP and GP val­ues by some fac­tor—which helped even more.

[6] Of course, that sort of thing doesn’t scale, and nei­ther can it last, just as you can­not build a whole coun­try like a kib­butz. But it may be en­tirely pos­si­ble, and perfectly healthy, to oc­ca­sion­ally cleave off sub­groups who fol­low that model, then to meld back into the over­group at the com­ple­tion of a pro­ject (never, in­deed, hav­ing truly sep­a­rated from it—the sub-groups’ mem­bers con­tin­u­ing to par­ti­ci­pate in the over­group, even as they throw them­selves into the sub-pro­ject).

No nominations.
No reviews.