Among Us is one of the latest crazes to hit online gaming and the streaming platform Twitch. It has a simple premise: players complete tasks on a ship, but among them are impostors that attempt to sabotage and kill the other players while avoiding detection. The game is cross-platform across PC and mobile, lending to its viral success. However, this is not a game review…

在我们当中，是最受欢迎的在线游戏和流媒体平台Twitch之一。 它的前提很简单：玩家在船上完成任务，但其中冒名顶替者企图破坏并杀死其他玩家，同时避免被发现。 该游戏在PC和移动设备上都是跨平台的，这极大地促进了其成功。 但是，这不是游戏评论…

The idea of an “Impostor,” or malicious actor, among a group of individuals working towards a common goal is what makes one of the biggest problems in computer science and theory so complicated. This problem is what underpins the mechanics of leading blockchain technologies such as Ethereum, Bitcoin, Litecoin, and others. The concept of the game is an allegorical jab at developers everywhere working on consensus algorithms for distributed technologies. To understand this allegory, let’s take a trip back ancient Istanbul.

在为共同目标而努力的一群人中，“冒名顶替者”或恶意行为者的想法使计算机科学和理论中最大的问题之一变得如此复杂。 这个问题是支撑诸如以太坊，比特币，莱特币等领先区块链技术机制的基础。 游戏的概念在各地开发分布式技术的共识算法的开发人员中寓言讽刺。 要了解这个寓言，让我们回溯古老的伊斯坦布尔。

拜占庭将军的问题 (Byzantine General’s Problem)

In a fictional ancient war involving the Byzantine empire, there were two armies surrounding a fortified city. The city had substantial defenses, enough to easily defeat a single army, however not enough to fend off two at the same time. In order to coordinate the attack, the generals enlisted the help of messengers, you know, since cellphones didn’t exist back then. Because of this communication strategy, there was no way to tell if the other general received the message, let alone the correct one. For instance, the messenger could actually be an embedded spy that is attempting to throw off the attack, or the messenger could be caught and replaced with another messenger (a classic man in the middle attack). In the same vein, we could introduce a third army to the mix that is thought to be an “ally” but turns out to be in cahoots with the city under siege. The question becomes, how do the armies coordinate to attack the city?

在一场涉及拜占庭帝国的虚构的古代战争中，有两支军队围绕着一座设防的城市。 该市拥有强大的防御能力，足以轻易击败一支军队，但不足以同时抵御两支军队。 为了协调这次袭击，将军们寻求信使的帮助，因为那时手机还不存在。 由于采用这种交流策略，因此无法判断对方是否收到了该消息，更不用说正确的消息了。 例如，信使实际上可以是试图阻止攻击的嵌入式间谍，或者可以捕获该信使并将其替换为另一个信使(中间攻击中的经典人物)。 同样，我们可以向被认为是“盟军”的混合部队中引入第三支军队，但事实证明该部队与被包围的城市密不可分。 问题变成，军队如何协调进攻城市？

The essence of the story is that when attempting coordination of multiple agents, the method of detecting failure, compromise, or malicious intent of an agent or group of agents is difficult. In order to decide the right action to take, or who to exclude, the agents must come to a consensus. The path to consensus is not always clear or computationally inexpensive.

故事的本质是，当尝试协调多个代理时，检测一个代理或一组代理的故障，危害或恶意意图的方法很困难。 为了决定采取正确的行动或排除谁，代理人必须达成共识。 达成共识的途径并不总是很明确或计算上不昂贵。

分布式分类帐 (Distributed Ledgers)

I would now like to break from the hype terminology “blockchain”, “cryptocurrency”, or “bitcoin” and instead focus on the root component and goal. In the case of a typical “blockchain,” the goal is for multiple agents, or nodes, to maintain a single indelible ledger that cannot be modified, discredited, or misrepresented. This ledger should also prevent concepts such as “double spending” and other malicious behavior from occurring in the system. Relating back to the previous idea of consensus on accomplishing a singular goal, the distributed system should be able to maintain a “distributed ledger” of record that is impervious to the problems faced by the Byzantine generals. This attribute of being impervious to these Byzantine faults is known as “Byzantine Fault Tolerance,” and the more general terminology for the technology is “distributed ledger technology (DLT).”

我现在想打破炒作术语“ blockchain”，“ cryptocurrency”或“ bitcoin”，而是专注于根源和目标。 对于典型的“区块链”，目标是多个代理或节点维护一个不可修改的总账，该总账不能被修改，抹黑或歪曲。 该分类帐还应防止在系统中发生诸如“双重支出”之类的概念和其他恶意行为。 回到关于达成单一目标的共识的以前的观点，分布式系统应该能够维护记录的“分布式分类帐”，这对于拜占庭将军所面临的问题是不可渗透的。 这种不受拜占庭式故障影响的属性称为“拜占庭式容错” ，而该技术的更通用术语是“分布式分类帐技术(DLT)”。

回到我们中间 (Back to Among Us)

Whilst playing the game, a player might come across another player acting suspicious (“sus”), and has the ability to call an emergency meeting. Another call to action for all players is when a player reports a dead body.

虽然玩游戏，玩家可能会遇到可疑的行动(另一位玩家“SUS” )，并已召开紧急会议的能力。 所有玩家的另一个号召性用语是当玩家报告尸体时。

These meetings invoke a voting mechanism in which all players, both Impostors and Crewmates, will attempt to oust the malicious actors. Crewmates (non Impostors) will try to cast out the Impostor, and the Impostor will try to thwart and obfuscate the truth to remain in their midst and in the game. As seen in the game, a preferable outcome based on consensus of the players is not always accomplished, and this same method of trickery and deceit is what makes blockchain so difficult. Because of this propensity to failure, the game’s rules and voting mechanism means the game is not Byzantine Fault Tolerant. This shows how a simple voting mechanism among agents is not enough to ensure Byzantine fault tolerance.

这些会议调用了一种投票机制，在该机制中，所有参与者，包括冒名顶替者和船员，都将试图驱逐恶意参与者。 队友(非冒名顶替者)将尝试驱逐冒名顶替者，而冒名顶替者将试图挫败和混淆真相，以保留在他们的中间和游戏中。 从游戏中可以看出，基于玩家共识的可取的结果并不总是能够实现的，而这种欺骗和欺骗的方法正是导致区块链如此困难的原因。 由于这种失败的倾向，游戏的规则和投票机制意味着游戏不具有拜占庭式的容错能力。 这说明了座席之间简单的投票机制不足以确保拜占庭容错能力。

This voting mechanism echoes loosely the ideas of voting-based Byzantine fault tolerant (BFT) algorithms such as PBFT, where a combination of multi-stage voting and other processes are implemented. Let’s look at a case where we incorporate a proof-of-work (PoW) algorithm, not dissimilar to the one used by Bitcoin. In this scenario, we will attempt to make the game Byzantine Fault Tolerant, meaning the Impostor never wins.

这种投票机制松散地回响了基于投票的拜占庭容错(BFT)算法(例如PBFT )的思想，其中实现了多阶段投票和其他过程的组合。 让我们看一下我们结合了工作量证明(PoW)算法的情况，该算法与比特币使用的算法类似。 在这种情况下，我们将尝试将游戏制作为“拜占庭容错”，这意味着“冒名顶替者”永远不会获胜。

任务和更多任务 (Tasks and More Tasks)

In the normal game, crewmates are working to complete tasks to fill up a bar at the top. Once the bar is filled, the game is complete and the Crewmates win. Let’s equate this to “miners attempt to solve cryptographic problems, and the first one to solve the problem generates the valid block.” In order to make this analogy, we need to change the structure of the game a bit.

在正常游戏中，队友正在努力完成任务，以填补顶部的酒吧。 一旦酒吧被填满，游戏就结束了，而船友们也赢了。 让我们将其等同于“矿工试图解决密码问题，而第一个解决该问题的人会生成有效块。” 为了进行类比，我们需要稍微改变游戏的结构。

In our new version, Among Us BFT edition, Crewmates still complete tasks, but as they complete tasks they might come across a “golden ticket” task that will enable them to “win the round.” This task will randomly occur while completing the other tasks, and the difficulty of reaching this “golden ticket” increases each round. The impostor will attempt to carry out “bad tasks” in an attempt to sabotage the other players. The malicious tasks are quickly recognized by the rest of the agents as not conforming to the standard, and so are quickly dealt with as they do not align to the rest of the players’ output.

在我们的新版本“我们之间的BFT”版本中，船员仍然可以完成任务，但是当他们完成任务时，他们可能会遇到“金牌”任务，这将使他们能够“赢得一轮”。 该任务将在完成其他任务时随机发生，并且到达“黄金票”的难度每轮都会增加。 冒名顶替者将试图执行“不良任务”，以破坏其他参与者。 其余代理Swift将恶意任务识别为不符合标准，因此由于恶意任务与玩家的其他输出不符，因此可以Swift进行处理。

In addition to this, the number of players in the game is quite large, with more than 50% of them using their combined “task rabbit” power to get the tasks done. Therefore, there is a lot of power being used to reach the “golden ticket” task. Theoretically, if 51% of this total “task rabbit” power was acquired by the group of Impostors with them “killing off” the Crewmates, they could legitimize their actions and overthrow the Crewmates. This is known as a 51% attack. In order to subvert the dangers of a 51% attack, the number of agents has to be sufficiently large in order to guarantee a large amount of distributed power.

除此之外，游戏中的玩家数量很大，超过50％的玩家使用其组合的“任务兔子”力量来完成任务。 因此，有很多力量可以用来完成“金票”任务。 从理论上讲，如果这组“任务兔子”的全部力量中有51％是由“冒名顶替者”团体获得的，他们“杀死”了船员，那么他们就可以使自己的行动合法化并推翻船员。 这被称为51％攻击。 为了颠覆51％攻击的危险，必须确保代理数量足够大，以保证大量的分布式电源。

So now we have more than 50% of the players focused on reaching the “golden ticket” task with all of their output contributing to the common goal. This proof of work among all the agents towards winning the game is what makes Bitcoin and other PoW-based block chains Byzantine fault tolerant, and could make Among Us fault tolerant (and very boring) as well.

因此，现在我们有超过50％的参与者专注于完成“金票”任务，他们的所有输出都为共同目标做出了贡献。 所有代理商在赢得比赛中所做的工作证明是使比特币和其他基于PoW的区块链拜占庭容错，并且也可能使我们之中的容错(并且很无聊)。

However, what makes Among Us fun is its flawed method of consensus that is rife with distrust, lies, and deceit, which has the ability to ruin longstanding friendships and cause rage quits galore.

但是，使我们之中的乐趣变得有趣的是，它充满缺陷的共识方法充满了不信任，谎言和欺骗，它有能力破坏长期的友谊，并导致愤怒泛滥。

结论 (Conclusion)

Although Among Us isn’t a 1-to-1 correlation with the concepts of distributed ledger technology, it is a fun thing to muse about when playing the game. The game is an interesting view on the concept of consensus, and the implications of a malicious actor in a group of agents.

尽管“我们当中的人”与分布式分类帐技术的概念不是一对一的关联，但在玩游戏时却想着这很有趣。 该游戏对共识的概念以及一组代理中恶意行为者的含义是一种有趣的观点。

Try to think of ways that a more favorable consensus can be reached in Among Us. What about leader election mechanisms? Multi-stage voting? Staking based on the number of tasks that a Crewmate accomplished? Let me know what you come up with!

尝试思考可以在我们当中达成更有利共识的方式。 领导者选举机制如何？ 多阶段投票？ 根据Crewmate完成的任务数量进行投注？ 让我知道您的想法！