第二章-2
【Action Choice: Author's Deduction】
【Set Goals: Personal Survival Rate Simulation】
Personal Survival Rate:
Do not approach.
Refrain from making any unnecessary statements.
Immediately begin surveying the surroundings and adjusting position to ensure a physically viable retreat route.
Worst-Case Outcomes Deduction:
Outcome 1: Others (non-Eva) touch humanoid entities, triggering an instant-death crisis.
Outcome 2: Others' (non-Eva) approach is interpreted as hostility, sparking conflict.
Outcome 3: Malicious entities behind broadcast-induced chaos to divert attention and commence slaughter.
These represent only basic worst-case outcomes deduction.
Actual possible outcomes far exceed these, but further elaboration is omitted here due to length constraints.
Action Implementer Controllability Deduction:
The action implementer cannot control humanoid entities, groups, or broadcast malicious entities.
However, they can control their own words and actions to preserve minimal mobility.
Avoid placing oneself in the primary line of fire and prepare basic fallback options.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Threatens the set goal.
Is this controllable by the Action Implementer?
Possesses basic controllability.
——As the worst-case outcome threatens the set goal, yet the Action Implementer maintains basic control over it, this action choice constitutes a 【Gambling Strategy】
——————
【Perspective Swap Simulation Sketch (Author → Character POV)】
【Note】:
Perspective swap sketching ≠ speculating on the character's actual psychology and motives.
Perspective swap sketching = The author “places themselves” in the character's position and conducts quick simulations with the sole condition of “serving the current set goals.”
【Set Goals: Personal Survival Rate】
A murder scene-like setup, with what appears to be a corpse and a rifle.
An unknown voice—likely the abductor— uses the broadcast to lead us here.
No need for overthinking; an instant-death crisis could erupt any second.
I cannot do anything that might be interpreted as hostile.
I must immediately start looking for an opportunity to escape this place, while avoiding drawing the attention of others or the entity behind the broadcast.
This is almost impossible to guarantee, and I cannot predict how the group or the entity behind the broadcast will react next.
But right now, endless Deduction is pointless. Doing my best not to die in the next moment is the only thing I can do.
——————————
【Action Choice: Author's Deduction】
【Set Goals: Group Survival Rate Scenario】
1. Provide advance warning to the group, alerting them of the risk, and preventing others from approaching or interacting with humanoid entities.
2. Attempt to curb and calm the group's tendency toward suspicion and panic, maintaining control of the situation.
3. Plan and organize the group's evacuation.
Worst-Case Outcomes Deduction:
（1）- Group panic escalates, triggering internal physical conflict.
（2）- Group risk awareness declines, reducing defenses against potential threats.
（3）- Malicious entities are provoked by broadcasts, initiating elimination operations.
Any of these worst-case outcomes may cause the group to develop suspicion and hostility toward the Action Implementer, or result in the Action Implementer's immediate targeted elimination.
Action Implementer Controllability Deduction:
The nature of humanoid entities is uncontrollable, group reactions are uncontrollable, and the actions of malicious entities post-broadcast are uncontrollable.
However, the Action Implementer can control their own words and actions to attempt to influence the current situation.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Threatens the set goal.
Is this controllable by the Action Implementer?
Possesses basic controllability.
——As the worst-case outcome threatens the set goal, yet the Action Implementer maintains basic control over it, this action choice constitutes a 【Gambling Strategy】
————
【Perspective Swap Simulation Sketch (Author → Character POV)】
【Note】:
Perspective swap sketching ≠ speculating on the character's actual psychology and motives.
Perspective swap sketching = The author “places themselves” in the character's position and conducts quick simulations with the sole condition of “serving the current set goals.”
【Set goals: Group survival rate】
Murder scene-like, with what appear to be corpses and a rifle.
An unknown voice—likely the abductor—led us here after the broadcast.
If no action is taken, the group may face an instant death scenario.
Any action I take now risks making me the focal point for both the group and the kidnappers, even be interpreted as needlessly causing anxiety or harboring ill intent.
But it's better than putting everyone in danger of death.
Signaling the risk, calming the panic, organizing an evacuation... These are the most effective options available right now. I must choose one immediately—there's no time for hesitation.
——————————
【Key Supplement】
As mentioned earlier, a flawless strategy is virtually nonexistent in most situations, especially when resources and options are severely constrained. The same holds true for an “absolutely safe strategy.”
Strategy is never about finding the perfect answer, but about avoiding the more unacceptable option.
—————————————————————————————————
【Time Progression】
【Current Objective Reality and Author's Deduction (Eva's Perspective)】
New Information:
The watch's electric shock capability is sufficient to render a person completely incapacitated and unconscious (Grace's experience)
An entity claiming to be Tozu appeared, demanding the group engage in mutual killing.
An entity called Mara appears, armed with a firearm, and fires a shot that nearly kills Grace.
Tozu threatens immediate execution for any further boundary violations.
After Wolfgang's second statement, the group's emotions become highly charged.
The space has a single exit (the academic tribunal site). Departure is only possible by elevator.
The enclosed external space contains multiple areas, but details remain unclear.
Currently visible personnel include myself and 18 others, with 2 (Tozu and Mara) exhibiting clear hostility.
Objective Enemy/Friend Assessment:
Enemy 1
Me (Eva): Physical condition stable, unarmed, restrained by an electric shock bracelet.
Enemy: Estimated at least two individuals, possessing absolute firepower superiority and no restraints.
I (Eva) possess no intelligence beyond the adversaries' basic appearance, self-reported names, demands, and confirmed execution capability.
The adversaries can monitor my (Eva's) every movement in real-time via surveillance.
Enemy 2
Me (Eva) 1 individual
Enemy (worst-case scenario) 15 individuals
No controllable variables exist beyond myself.
Author's Deduction:
At least two kidnappers confirmed. Possession of lethal instant-kill weapons confirmed. No constraints from civilized norms confirmed.
Chaotic situation with unpredictable outcome.
Immediate risk of death.
No place to escape.
Eva possesses no military assets. Physical safety is entirely dependent on the personal will of the armed hostile entity and the group's self-restraint threshold.
【Actual Action Choices of the Action Implementer (Eva)】
After Mara fired a warning shot and Tozu threatened execution for crossing the line, when the group's emotions were running high following Wolfgang's speech, the action implementer made stigmatizing speech about the group, stigmatized the trust, and made unfounded accusations that Wolfgang had murderous intent.
【Set Goals—Personal Survival Rate (Based on Assumptions)】
Worst-Case Outcomes Deduction:
Short-Term:
The group descends into catastrophic rioting, with Eva becoming one of the first threats targeted for elimination.
Provoked armed kidnappers, they target Eva for public execution.
The situation descends into complete chaos and violent conflict. Trapped in a confined space, Eva cannot escape, leading to irreversible consequences.
Long-Term (if the situation does not collapse immediately):
Wolfgang is a malicious individual or spy, viewing Eva as the primary target for elimination.
The group descends into moral slide and a chain of suspicion, civilized boundaries dissolving. Eva, identified as a “liar” and having engaged in open aggressive behavior, becomes the primary target for elimination within the group.
Eva is flagged by the armed kidnappers as an unstable element and a target for elimination.
The above represents only a few basic worst-case outcomes deductions.
The actual possible trajectories far exceed these, but further elaboration is omitted here due to space constraints.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Destruction of set goals.
Is this controllable by the Action Implementer?
Completely uncontrollable.
——As the action choice fundamentally contradicts the set goal and exhibits no controllability, it constitutes a 【Self-Destructive Choice】 and does not constitute a strategy.
Therefore, 【derivative actions and Deduction based on this self-destructive choice】 simultaneously lose the strategic significance.
————【Set Goals—Group Survival Rate (Based on Assumptions)】
Worst-Case Outcomes Deduction:
Short-Term:
The group descends into catastrophic rioting, with casualties and death tolls spiraling out of control, leading to total annihilation.
Armed kidnappers become provoked and begin executing members of the group.
The situation devolves into complete chaos and violent conflict. Trapped in a confined space, the group cannot escape, resulting in irreversible consequences.
Long-Term (if the situation doesn't collapse immediately):
The group descends into moral slide and a chain of suspicion, eroding civilized boundaries. The most vulnerable members become the first targets for murder and assault.
The armed kidnappers escalate their control, permanently eliminating any possibility of resistance or escape for the group.
The above outlines only the most basic worst-case outcomes deduction.
The actual possible trajectories far exceed these, but further elaboration is omitted here due to space constraints.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Destruction of set goals.
Is this controllable by the Action Implementer?
Completely uncontrollable.
——As the action choice fundamentally contradicts the set goal and exhibits no controllability, it constitutes a 【Self-Destructive Choice】 and does not constitute a strategy.
Therefore, 【derivative actions and Deduction based on this self-destructive choice】 simultaneously lose the strategic significance.
————————
【Note: To avoid redundancy, the term for armed kidnappers will henceforth be referred to as ‘the Mastermind’.】
——————————
【Key Supplement】
Yes, if the true objective of the Mastermind is solely “to make students kill each other,” then they cannot—or rather, cannot easily—shoot the students.
But how can Eva be certain that this is the Mastermind's sole objective?
Or rather, even if the Mastermind's goal is indeed solely “to make students kill each other,” how can Eva ensure that the Mastermind is sufficiently rational, possesses strong judgment, remains singularly focused and constant, and won't commit self-destructive acts or abandon their own objectives?
As stated earlier, this is the distinction between 【set goals】 and reality. “The Mastermind wants students to kill each other” is a set goal.
To avoid decision paralysis, we must hypothesize a “set goal” for the Mastermind and analyze their words and actions accordingly—but we cannot equate this directly with the world's truth.
————
【Action Choice: Author's Deduction】
【Set Goals: Personal Survival Rate Simulation】
Remain silent.
Refrain from any actions that might attract the attention of the Mastermind or emotionally charged groups.
Whenever possible, physically distance oneself from the conflict center (away from firing lines and Wolfgang).
Worst-Case Outcomes Deduction:
Violent conflict erupts. The Mastermind begins eliminating resisters, plunging the situation into chaos and collapse.
Action Implementer Controllability Deduction:
Eva cannot control how the crowd or the Mastermind interprets this situation or what actions they might take.
However, Eva can control her own words and actions, as well as her movement and physical position. By maximizing and prolonging her survival rate and time, she can gain the opportunity to escape the trial ground.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Threatens the set goal.
Is this controllable by the Action Implementer?
Possesses basic controllability.
——As the worst-case outcome threatens the set goal, yet the Action Implementer maintains basic control over it, this action choice constitutes a 【Gambling Strategy】
——————
【Perspective Swap Simulation Sketch (Author → Character POV)】
【Note】:
Perspective swap sketching ≠ speculating on the character's actual psychology and motives.
Perspective swap sketching = The author “places themselves” in the character's position and conducts quick simulations with the sole condition of “serving the current set goals.”
【Set Goals: Personal Survival Rate】
Two sides are currently present on the scene: the armed masterminds and the group.
Wolfgang's speech and the group's emotions are now established facts.
The situation is highly likely to spiral completely out of control. I must not draw the attention or hostility of either side, much less provoke either. If conflict erupts, it will inevitably center around Wolfgang and the masterminds first. I must immediately leave the conflict zone.
The only exit is the elevator. Though I don't know if it can be activated or how quickly it might respond, I must reach it as fast as possible. Even if I can't leave now, I must at least create an opportunity.
Even if I could safely return to the ground level, this is an enclosed space. I'm wearing an electric shock watch—there's nowhere to run.
But it's better than dying here immediately. If I seize the chance to survive this moment, there might still be a chance.
————————
【Action Choice: Author's Deduction】
【Set Goals: Group Survival Rate Simulation】
Action Choice:
Follow Wolfgang and group members, amplify the group's emotional state, and further focus it on external threats to prevent dispersion or inward explosion of emotional energy. Trap the Mastermind into a strategic deadlock.
Worst-Case Outcomes Deduction:
The armed Masterminds begin suppressing resisters, including the Action Implementer, by lethal force.
Action Implementer Controllability Deduction:
Eva cannot control how the group or the armed Masterminds interpret this possibility or what actions they might take.
However, Eva can control her own words, actions, and choices.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Threatens the set goal.
Is this controllable by the Action Implementer?
Possesses basic controllability.
——As the worst-case outcome threatens the set goal, yet the Action Implementer maintains basic control over it, this action choice constitutes a 【Gambling Strategy】
——————————
【Perspective Swap Simulation Sketch (Author → Character POV)】
【Note】:
Perspective swap sketching ≠ speculating on the character's actual psychology and motives.
Perspective swap sketching = The author “places themselves” in the character's position and conducts quick simulations with the sole condition of “serving the current set goals.”
【Set goals: Group survival rate】
Two sides are currently present on the scene: the armed masterminds and the group.
Wolfgang's speech and the group's emotions are now established facts.
Attacking the group carries a high risk of triggering internal collapse (as demonstrated by Eva's original action choice), while openly supporting an external enemy would plunge the situation into even more unpredictable chaos.
Therefore, I should avoid introducing uncontrollable variables and instead strengthen existing factors that could potentially benefit our side (the group).
If the gun fires, the illusion of a “mutual killing game” will shatter, reverting the scenario from a “game with free choice” to a “kidnapper slaughtering hostages.”
The raw violence of the scene will force the group to confront reality, shattering the “game” brainwashing.
If the Masterminds proceed this way, it will self-destruct their own set goals.
Yes, as stated above, if the Mastermind's set goals are not—or not entirely—a “mutual killing game, or if the Mastermind misjudges the situation, or if the group retreats and collapses too quickly, this move will lead to my—and even the group’s—annihilation.
But this is likely the only gambling opportunity.
A closed space, electric shock watches that can incapacitate people, an absolute disparity in firepower, an information black box, no institutional protections, no guaranteed outside aid, a group easily incited and manipulated... This is a complete, irreversible disadvantage.
If the Mastermind successfully implants its framework of mutual killing, the group will likely spiral into a snowballing, self-destructive, vicious cycle.
Not to mention that even if the group were to obey perfectly, the Mastermind would still possess the full capability and opportunity to suddenly execute everyone at any point thereafter.
Slow death versus quick death, passive death versus active death—the outcome is death in essence. There is no difference.
Only when the Mastermind reveals a flaw can we operate under the working assumption that their set goals are “to make students kill each other, not to slaughter them personally.” Force the Mastermind into an either/or dilemma (either abandon their objectives or make temporary concessions) to reclaim a sliver of initiative and room for maneuver.
The success rate of “lying low and waiting for an opportunity” hinges entirely on every group member possessing extraordinary psychological resilience—the ability to maintain self-discipline under prolonged, high-pressure, high-risk conditions.
One person, two people, even three might manage it. But can I guarantee everyone will? Impossible.
The Mastermind's sudden conscience or behavioral patterns are not things I can gamble on. Entrusting the success rate of the set goals entirely to an entity confirmed to possess absolute malevolence is strategic suicide.
Every second spent clinging to wishful thinking and illusions of salvation pushes the current situation further toward the Mastermind's control, ultimately leaving our side with no possibility of counterattack.
Therefore, whenever an opportunity arises to seize control, not a single second should be allowed for the Mastermind to maintain their complete dominance safely.
In the most extreme scenarios, whether attempting to gamble or delay in hopes of gaining potential opportunities or advantages, one must be prepared to sacrifice some members due to uncontrollable factors or to use some members as leverage to gain a breakthrough advantage.
And this includes oneself. One is also a resource; there is no reason not to use it.
—————————————————————————————————
【Time Progression】
【Current Objective Situation and Author's Deduction (Eva's Perspective)】
New Information:
The kidnappers demand that the group engage in a “mutual killing game” (the killer’s process includes murder and a class trial) in exchange for an escape opportunity.
The kidnappers possess instant-kill lethal force and have demonstrated their willingness to use it. They have also threatened execution for crossing boundaries.
The group has demonstrated aversion to Eva's aggressive words and actions, along with unease about her “liar” identity.
Toxic substances are present in the pharmacy. Knives are present in the kitchen.
Personal basic information of group members appears on the watch.
“Mutual Killing Rules” appear on the watch.
General Physical Environment:
Enclosed space with multiple distinct areas.
Currently visible personnel: 18 individuals, including myself. Two (Tozu and Mara) exhibit clear hostility.
Objective Enemy/Friend Assessment:
Enemy 1
Me (Eva): Physical condition stable, unarmed, restrained by an electric watch.
Enemy: Estimated at least two individuals, holding absolute tactical superiority with no constraints.
Me (Eva): No intelligence beyond their basic appearance, self-reported names, demands, and execution capability.
Enemy: Can monitor my (Eva's) every move in real-time via surveillance.
Enemy 2
Me (Eva): 1 individual
Enemy (worst-case scenario): 15 individuals
Author's Deduction:
The kidnappers are confirmed to be at least two individuals, confirmed to possess lethal instant-kill capabilities, and confirmed to operate without any constraints from civilized norms.
The situation has temporarily stabilized, but there are no actual safeguards or protections. Chaos, collapse, and attack could resume at any moment.
Risk of losing physical mobility sovereignty (electro-shock watch) at any moment.
No privacy protection.
No security guarantees.
No place to flee.
Eva has potential access to weapons, but so do others. Eva remains without genuine physical safety assurance.
No controllable variables beyond oneself.
——————————————————————
【Key Deduction】
The “mutual killing rule” promulgated by Tozu does not contain a clause stating that only the first killer counts.
This implies that if a group genuinely accepts the framework of a mutual killing game, there will be no constraints from rule-based cost-benefit calculations (since someone is already dead, killing again offers no benefit) to restrain them.
The worst-case outcomes would not be “one member killed and one murderer executed,” but rather multiple members consecutively killing each other, triggering a cycle of internal collapse and potentially sliding directly into collective internal slaughter.
Under these conditions, any factor capable of triggering a chain-reaction collapse and mass panic carries a high risk of igniting the fuse for snowballing destruction.
————————————————————————
【Action Implementer (Eva)‘s Actual Action Choice】
Upon observing the group members’ personal basic information appearing on the watch, immediately declare that “the watch and its information place the group in a life-or-death crisis.” Use this to prohibit the group from viewing the information, thereby concealing her true talent (mathlete) and maintaining the self-proclaimed title of “Ultimate Liar.”
【Set Goals — Conceal True Talent (Based on Eva's Actions)】
Worst-Case Outcomes Deduction:
The group disregards Eva's demands and independently accesses the watch information.
The Mastermind proactively reveals Eva's true talent.
Action Implementer Controllability Deduction:
The action implementer cannot control others' interpretations of her actions or their subsequent responses.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Destruction of set goals.
Is this controllable by the Action Implementer?
Completely uncontrollable.
——As the action choice fundamentally contradicts the set goal and exhibits no controllability, it constitutes a 【Self-Destructive Choice】 and does not constitute a strategy.
Therefore, 【derivative actions and Deduction based on this self-destructive choice】simultaneously lose the strategic significance.
————
【Set Goals—Personal Survival Rate (Based on Assumptions)】
Worst-Case Outcomes Deduction:
The group descends into a spiral of anarchic riots/paralysis due to panic outbreaks and the reignition of suspicion chains. Eva is incapable of controlling the group, leading to irreversible consequences.
The group discovers Eva's true personal objectives, marking Eva as a dangerous element and potential traitor, making Eva the primary target for internal elimination.
Eva is marked, exploited, or targeted for early elimination by the Mastermind due to actions interfering with its operations.
The following consequences may include, but are not limited to:
Serial massacres.
Violent disposal.
Physical assault.
Lynching.
Action Implementer Controllability Deduction:
Eva cannot control others' interpretations of Raven's actions or their subsequent responses.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Destruction of set goals.
Is this controllable by the Action Implementer?
Completely uncontrollable.
——As the action choice fundamentally contradicts the set goal and exhibits no controllability, it constitutes a 【Self-Destructive Choice】 and does not constitute a strategy.
Therefore, 【derivative actions and Deduction based on this self-destructive choice】 simultaneously lose the strategic significance.
————
【Set Goals—Group survival rate (Based on Assumptions)】
Worst-Case Outcomes Deduction:
The group descends into a spiral of anarchic riots/paralysis due to panic outbreaks and the reignition of suspicion chains. Remnant civilized consciousness and moral boundaries collapse, regressing society to a law of the jungle. The most vulnerable suffer casualties first, evolving into a collective self-destructive cycle and internal slaughter.
The Mastermind decides to initiate physical clearance prematurely due to the group's complete loss of control.
Action Implementer Controllability Deduction:
Action Implementer cannot control others' (the group and the Mastermind) interpretation of her behavior or subsequent actions.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Destruction of set goals.
Is this controllable by the Action Implementer?
Completely uncontrollable.
——As the action choice fundamentally contradicts the set goal and exhibits no controllability, it constitutes a 【Self-Destructive Choice】 and does not constitute a strategy.
Therefore, 【derivative actions and Deduction based on this self-destructive choice】 simultaneously lose the strategic significance.
————
【Action Choice: Author's Deduction】
【Set Goals: Conceal True Talent-if Deduction】
Abandon this goal.
The moment student records appeared on the watch, Eva lost any guarantee of keeping her true title hidden.
As stated above, Eva cannot ensure control over the group's actions.
Even assuming Eva successfully manipulates the group to permanently block members from viewing records—the Mastermind remains uncontrollable. They could personally leak the data.
At this stage, attempting to forcibly guarantee control over the Mastermind's actions and the existence of the data would require instantly possessing the resources to fulfill the strategic requirement of “physically eliminating the external threat (the Mastermind).”
Ultimately, ensuring the achievement of the “100% guarantee to conceal true title” objective generally points to two approaches.
1. Instantly physically eliminate all group members and the Mastermind.
2. Instantly seize control of the facility systems, erasing all relevant data and information.
Under the current objective realities of the story, the requirements for the Action Implementer to execute either of these approaches fall into the realm of supernatural and psychic fantasy.
If, within a semi-realistic world setting, achieving a Set Goal necessitates fantasy or supernatural power as a prerequisite, then that Set Goal need not be considered.
————
【Set Goals Adjustment: Mitigating Negative Impacts After Exposure of Real Talent (if Deduction)】 (1)
【Set Goals—Personal Survival Rate (Based on Assumptions)】(2)
【Set Goals—Group Survival Rate (Based on Assumptions)】(3)
【Action Choice: Author's Deduction】
Before the group reviews the materials, proactively confess the true talent and apologize for previous lies, requesting the group's forgiveness.
Worst-Case Outcomes Deduction:
The group rejects Eva's apology.
Eva is henceforth labeled an untrustworthy high-risk individual and a potential priority elimination target.
The group descends into a chain of suspicion triggered by the confirmed deception, leading to chaos and a self-destructive cycle.
The Mastermind exploits this incident as ammunition to launch attacks.
The above represents only basic worst-case outcomes deduction.
Actual possible outcomes far exceed these, but further elaboration is omitted here due to space constraints.
Action Implementer Controllability Deduction:
The Action Implementer cannot control how others (the group and the Mastermind) interpret her actions or their subsequent responses.
However, the Action Implementer can control her own words and actions, striving to minimize and mitigate potential losses.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Threatens the set goal.(1)(2)(3)
Is this controllable by the Action Implementer?
Possesses basic controllability.
——As the worst-case outcome threatens the set goal(1)(2)(3), yet the Action Implementer maintains basic control over it, this action choice constitutes a 【Gambling Strategy】
——————
【Key Supplement】
This is the so-called “chain reaction.”
A choice or action does not exist in a vacuum, nor do external circumstances. This is a continuous, ever-changing dynamic reality.
Had Eva made different choices at past junctures— such as not concealing her talents—then scenarios such as “true talents being exposed,” or the failure of set goals like “Personal survival rate” and “Group survival rate,” would have no conditions for existence.
When Eva simultaneously acted on two self-destructive choices—“declaring herself a liar” and “accusing the group”—the potential consequences became inevitable. Thus, the only recourse is to attempt 【damage control】now.
A 100% guarantee of concealment is impossible; exposure is merely a matter of time. Since a perfect solution cannot be found, the only option is to avoid the more unacceptable choices.
Such as the completely uncontrollable “being exposed by others,” “actively exposing and offering more vulnerabilities,” “actively increasing the risk of systemic collapse”...
Because the potential risks were already embedded in previous action choices, the past cannot be altered out of thin air. All that remains is a gamble-like stop loss—betting that everything won’t be destroyed because of this.
—————————————————————————————————
【Time Progression】
【Current Objective Reality and Author's Deduction (Eva's Perspective)】
The Mastermind's Promise: Murderers who successfully pass the class trial will be released.
Credibility of the Mastermind's Promise:
The Mastermind's promise lacks explicit contractual guarantees or constraints backed by institutions (such as law).
The power/status/position gap between the promised party (students) and the promisor (the Mastermind) is absolute.
The Mastermind's identity is that of armed, violent kidnappers coercing groups into inhumane, self-destructive violence.
The Mastermind has a history of breaking promises. (Using wordplay to lure the group into participating in the Mock Trial)
The Mastermind's interests are fundamentally at odds with those of the murderer (e.g., if the target is escape).
——Once the murder is committed, the Mastermind's interests no longer align with the murderer's.
Macro-physical environment: Enclosed space.
Author's Deduction:
The Mastermind holds no credibility.
There is no guarantee that murderers who pass the class trial will be released, nor even that the trial itself will occur.
Confined within the space, Eva has nowhere to flee.
——————
【Action Implementer (Eva)‘s Actual Action Choice】
Participate in the mutual killing game, attempting to escape by “winning the class trial after killing.”
【Set Goals—Personal survival rate (Based on Assumptions)】
Worst-Case Outcomes Deduction:
1. Eva's murder plan fails, permanently destroying her social identity and potentially leading to the group stripping her of her right to survive.
2. Eva is selected as the murderer during the class trial and loses, resulting in her execution.
3. Eva wins the class trial, but the Mastermind reneges on their promise and eliminates Eva.
4. The class trial never occurs. The group descends into total collapse and a cycle of anarchy and riots due to the murder and the shattering of institutional illusions. Trapped in an inescapable enclosed space, Eva faces irreversible consequences.
The above outlines only the most fundamental worst-case outcomes deduction.
Actual possible outcomes far exceed these, but further elaboration is omitted here due to space constraints.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Destruction of set goals.
Is this controllable by the Action Implementer?
Completely uncontrollable.
——As the action choice fundamentally contradicts the set goal and exhibits no controllability, it constitutes a 【Self-Destructive Choice】 and does not constitute a strategy.
Therefore, 【derivative actions and Deduction based on this self-destructive choice】 simultaneously lose the strategic significance.
—————————————————————————————————
【On the Threat of the Mastermind】
Regarding the Mastermind threat that ‘if you don't kill after accepting the gift package, you die (no time limit set),’ Eva said, ‘therefore I can only kill.’
“Killing others then escaping use class trail” is, from the perspective of Personal survival rate, a completely self-destructive decision. The worst-case outcomes destroy Set Goals, and the Action Implementer has no control over them at all.
Although, as Desmond says in the story, she could have completely refused the gift package. But for now, let’s assume she must accept it.
So if she doesn't kill, what other possibilities exist to handle the threat from the Mastermind?
【Possible Action Choices: Author's Deduction】
【Set Goals: Personal Survival Rate if Deduction】
【Special Constraint: Address the threat from the Mastermind while adhering to the premise of “not killing”】
Action Choice 1:
The Action Implementer pretends to agree, but never acts alone again.
If the Mastermind dares to approach and kill the Action Implementer at close range, they will expose themselves to being witnessed by other group members, thereby violating their own rules of mutual killing and causing their own power to collapse.
Worst-Case Outcomes Deduction:
The Mastermind uses long-range methods to kill the Action Implementer.
The Mastermind distorts the Action Implementer's future actions into “violations” to justify eliminating them.
The Mastermind incites other group members to attack the Action Implementer.
The Mastermind disregards its own established rules and kills directly.
——etc.
Action Implementer Controllability Deduction:
The Action Implementer cannot control the Mastermind's decisions.
However, the Action Implementer can control their own response methods, using the situation as leverage to attempt to restrict the Mastermind.
[Determining the Nature of Action Choices]
Severity of harm to set goals from the worst-case outcomes?
Threatens the set goal.
Is this controllable by the Action Implementer?
Possesses basic controllability.
——As the worst-case outcome threatens the set goal, yet the Action Implementer maintains basic control over it, this action choice constitutes a 【Gambling Strategy】
————
【Perspective Swap Simulation Sketch (Author → Character POV)】
【Note】:
Perspective swap sketching ≠ speculating on the character's actual psychology and motives.
Perspective swap sketching = The author “places themselves” in the character's position and conducts quick simulations with the sole condition of “serving the current set goals.”
【Set Goals: Personal Survival Rate】
【Special Restriction: Counter threats from The Mastermind without killing anyone】
The Mastermind threatened me, saying, “If you accept the gift package and don't kill anyone, you'll die,” but no explicit time limit was set.
Alright. Then I will never act alone again.
The Mastermind's “illusion of legitimacy” and the power built on it stem from his apparent adherence to self-imposed rules, pretending to be a neutral host.
If the “host” personally intervenes to kill and is detected by others, this “illusion of legitimacy” collapses. Once the rule illusion shatters, the group will likely refuse to obey any further demands, causing his goal of “promoting mutual slaughter” to self-destruct.
Therefore, I can exploit this vulnerability to limit the Mastermind's close-range direct attacks.
However, if the Mastermind employs long-range or indirect methods—such as poisoning, directly labeling one of my actions as a violation, or inciting the group to turn against me—these tactics are harder to guard against and more susceptible to narrative distortion.
But if the Mastermind's goal is to “foster mutual killing,” why must all attention be focused solely on me?
Employing any of the above tactics to attack me would consume the Mastermind's time and resources, while also introducing unnecessary risks through such actions.
The situation isn't static; other group members will also undergo changes and take actions. If the Mastermind keeps focus on a single point, they'll inevitably lose the bigger picture over minor gains, affecting the entire situation.
Of course, this strategy assumes the Mastermind genuinely intends to persist with mutual killing, their assessment of the situation is entirely accurate, and the group won't prematurely target me or suddenly collapse.
At its core, it's a gamble, because there are too many uncontrollable factors.
But I am a person in a desperate situation: zero military strength, zero authority, no one trustworthy, nowhere to escape. I'm also threatened that if I don't kill, I will die. And even if I succeed in killing, my life or death doesn't depend on myself.
In other words, because I have nothing more to lose, I fear the gamble less. Gambling, instead, might offer a sliver of control.
————