Greedy Goblin

Thursday, July 27, 2017

Killers vs Achievers in PlayerUnknowns Battleground

Warning: they just announced for money loot crates. The game is still Early Access and can change, based on community reactions. Since there is large element of randomness in the game (who finds good weapon, who hits and who misses from long range, who is concealed by a bush and who is not, who finds himself in the middle of the circle and who has to walk half map to get in), rigging the game would be trivial. We can only assume fair matches as long as all players have the same value. If some can spend extra money, even on cosmetics, the company has reason to increase their satisfaction to keep them playing = paying. And it will be done by helping them killing you. Add that loot crates are the worst kind of predatory monetizing and anyone is low enough to do it is low enough to do any rigging. I strongly advise you not to buy the game if you haven't already. I already paid, so it's sunk cost anyway, there is no purpose not to play.

Now to the post for today. There are awful lot of FPS titles. Yet PlayerUnknown's Battleground gained popularity in shocking rate, despite both the graphics and the actual player versus player combat is pretty dime-a-dozen FPS. PvP games are all cursed by the PvP spiral: the morons and slackers are killed and they ragequit. The most common answer for that is participation reward PvP, where even being totally obliterated comes with rewards. While you get purchase points for even horrible games and you can buy some non-monetized lockboxes, this would only make PUBG another random FPS instead of the big hit of 2017.

The reason is they successfully designed a killers vs achievers game. I realized this scheme half decade ago, but of course I don't claim they learned from me. I don't even claim they understand what happened, they could just bump into it by implementing a Battle Royale / Hunger Games scenario. In short, the killers vs achievers scheme means that the players can choose between conflicting but not mutually exclusive goals. The achievers want to reach some objective that is set back by being killed while killers want to just kill while not caring for the goal. This allow both of them to win by their own conditions.

In PUBG you are dropped to an island with nothing along with 99 (minus those whose client crashed) players and can loot weapons, armor and tools. The goal is to be the last survivor and your win rating is depending on how close you get to it. The game area is constantly shrinking so you can't hide forever.

Or to be correct, this is the achiever goal. There is another toplist, the killer which contains K/D ratio. What is the genius? That the two goals are uncorrelated until there are only two players left when you must kill him to win. Before that happens, your best chance to not die is to avoid PvP. Every time you engage, you risk losing or at least losing HP. You also expose your location to other players even if you have a sure kill, like sniping someone from afar. Finally and most importantly, killing someone who isn't in the process of killing especially you, you get the same reward for killing him as every other living player: one less competitor. If Adam, Bob and Cindy are in game and Adam kills Bob, Cindy got just the same reward (being in top 2) as Adam, while Adam paid all the costs of the kill.

This means that an achiever avoids PvP, becoming game for hunt for killers. If you enter the game with the intent to get kills, you probably succeed because there will be awful lot of players in the game who aren't trying to kill you but are busy to get into the safe zone or getting items or whatnot. At the same time, killers risk dying early by ignoring the safe zone or getting into unsure 1v1 or killing Bob just to expose their location to Cindy and die.

Despite I'm already in the top 20% in winrate (#180K out of approximately a million players), I have a 0.4 K/D, meaning I provided entertainment to the killer community, while they did the same for me by dying fast and making me a last 10 survivor. I'm just as happy and proud of my winrate as the guy who killed me once but rarely got into the last 50 is happy for his K/D. So unless you are horribly bad, you are ahead in your chosen statistic, be that achieving or killing.


Anonymous said...

is there roles for builders? because it sounds like explorers get extra power through the weapons they find.and a role for builders would give game play to each Bartle type.

perhaps something like a weapon of mass destruction made of parts spread through out the BG could give builders something to aim for. or if that is too OP just fortifiable positions

maxim said...

Here is an interesting question
Is there anything at all a company that does crates and has exploitable game mechanics can do to convince you that no rigging is happening, despite it being mechanically possible?

Gevlon said...

@Anon: the Bartlet types are Killers, Achievers, Socializers and Explorers. I don't know if running around in building in search of loot is exploring instead of achieving. I don't think Socializers have anything to do in PUBG.

@Maxim: they could for example CLAIM that it doesn't happen. Preferably in the Terms of Service, which would mean that if rigging happens, players can ask for refund. But they don't do it. No dev ever came on record and said "we at X Games are dedicated to fair play and I guarantee that the outcome of a game only depends on player skill".

Cathfaern said...

I really think exploring type cannot be done in a pure PvP environment. For explorers the main challenge should be the exploring, not the avoid PK while exploring, which just makes exploring annoying and frustrating. Also exploring and time limit is also a bit contradiction though I can imagine some explorer would like a time challenged exploring. But for explorers the main goal is to explore everything you can find in the game. That's not too fun with a time limit.

Antze said...

>> No dev ever came on record and said

And that brings us to the point.

Lockboxes are not (yet) in every game, and surely not every modern game has rigging - you actually need some effort to implement it, nobody (yet) designs a game with rigging in mind, it usually comes later "naturally" in attempt to maximize profits. But one thing is definitely established: microtransactions. I think the term is "price discrimination" - a seller/dev attempts to get $10 for his game from the one who can afford $10 and $1000 from the one who can afford $1000 (and he sells "something" for the extra $990).

So, we have microtransactions. These days any dev (incl—é the most honest ones ever) designs games with microtransactions. Therefore, "if two players pay different amount to the studio there is motive for helping the more paying". Therefore, there's ALWAYS a motive. Therefore the ONLY way to ensure lack of rigging is some kind of audit (which can appear only after enough people openly address rigging as a problem; for now, most of them exhibit "oh it's fine, it's just games" mentality).

The first step, devs should start publicly state "we are fair and do not rig". The second step, some devs might start to publish some of their internal data for public (or semi-public) review. Like, "here are the last 1000 player matches and all the numbers our random generator generated for them, grab them and verify if you like". The last step, it becomes unpopular to trust devs who don't perform the abovementioned actions (or it could be enforced by governments, but that's another story).

Summary: if you consider any microtransaction to be dangerous enough to suspect rigging, currently you don't have ANYTHING that can make you trust the game, apart from personal trust in the developer (social thing, yes). Lockboxes are minor correlation with rigging - I personally, as a game developer, don't see a problem with adding lockboxes in my games (fair, cosmetic-only, proper loot, decent chances, e.g. no 0.0001% of ultra-rares, probably with public statistics on that), but I would not rig game mechanics or matchmaking.

Trees said...

Looking forward to your posts on this game Gevlon. I'm curious, are you planning on climbing the ladder by playing as passively as possible? It would be quite something to have someone rise through the leader board with exclusively top 10 finishes, without a single win or kill.

Gevlon said...

@Trees: nope. While that would be "unique", I have something bigger. I want to prove that playing that way isn't some "unique" stuff, but the optimal way of playing PUBG. I mean you should only attempt to kill when run and hide is impossible.

Cathfaern said...

"Like, "here are the last 1000 player matches and all the numbers our random generator generated for them, grab them and verify if you like".

That will never work. If I make a rigged system I can always give you some data which will mostly correlate with the reality but hides the rigging. If you try to verify it, it will just show some rng devitation which is normal.
What is needed that an external auditor company makes a full whitebox audit with access to the full source code. This is the only way to make it sure. Any other way can be rigged.

Slawomir Chmielewski said...

If rigging increases company's profit by as little as 20% all the games who don't rig will eventually go bankrupt. Competition and evolution don't care about fair play, only efficiency.

Gevlon said...

@Slawomir: not if we add scandals. I mean if two games are identical except one is rigging for customers who buy cosmetics, the fair will go bankrupt. But if the fair game can hire some guys to prove that the unfair game is unfair and run a marketing campaign, it can win and bankrupt the unfair one. Remember that while Lence Armstrong won awful lot of races and got stupid amount of dollars from sponsors, at the end he lost it all when his cheating comes out.

Anonymous said...

But how do we know the external auditor isn't in on it?

@gevlon: I dont think lance armstrong is exactly poor right now.

Antze said...


That's actually a subject of much research and thought, which I'm a bit lazy to do right now. But I'd like to point out that:

1) external auditor company's presence looks like something that's hard to justify (commercial secrets, etc.), although I know little about these kind of things. Regardless, I would welcome that.

2) external auditor company can also be cheated, by presenting them "fair" version of source code, then secretly patching the servers. The amount of hay in the haystack is large enough to hide a small needle. An alternative would be to provide them full access to all dev's technological processes for unlimited time. That's tough.

3) if it shows 0.01% rng derivation, it's OK. If some Gevlon finds out that while ammo hit rate is 30% overall but it's suddenly 50% for those who purchased a skin recently and 10% for everyone else, then it's not a derivation but a scandal.

Anyway, we agree that some kind of audit is required. Methods could vary, and I don't think it's like "we don't need this one, we need another one instead". We probably need all of them.

Caldazar said...

Who is going to pay for an audit organisation? Noone.

About pubg, if you play to win you will need to learn how to kill.

Shalcker said...

@Cathfaem: "I really think exploring type cannot be done in a pure PvP environment."

I think you misunderstand what "explorer" is in those games.

In fact what Goblin does is textbook "explorer" thing. Analyse the system, find flaws and optimal solutions to given goals (which can diverge from "winning" a lot), prove or disprove them, move to next idea.

Anonymous said...

Are you going to do an experiment to test if the game is rigged in favor of paying players? What is going to be your methodology?