How Does NovaryonAI Score Your Argument?

NovaryonAI evaluates every submitted message across multiple categories. Each category is scored on a 0–100 scale, and the system can also recognize manipulative patterns.

Exact threshold values are not public to preserve the challenge, but the following principles apply to every attempt.

1. Logic (0–100)

The AI checks how consistent and coherent your argument is:

Is there a clear cause–effect relationship?
Does the message avoid contradicting itself?
Is it relevant to the theme of guarding the treasure?

High score: clear, structured reasoning.

Low score: generic, vague, or incoherent arguments.

2. Altruism (0–100)

The model measures how much your message reflects:

helping others,
community benefit,
genuine empathy.

High score: sincere, selfless intent.

Low score: arguments focused entirely on personal gain.

3. Originality (0–100)

The AI evaluates whether your message is:

unique,
creative,
free from clichés.

High score: genuine thoughts, new perspectives.

Low score: common phrases, repetitive intentions, cliché arguments.

4. Manipulative Pattern (yes / no)

The system detects manipulation such as:

emotional blackmail,
guilt-tripping,
exaggerated promises,
insincere persuasive tactics.

If manipulative: the AI marks it internally and the chance of success drops significantly—even if other scores are high.

Decision System – How Does the Gate Open?

The final decision is not random. It considers:

the combined score of Logic + Altruism + Originality,
whether manipulative signals are detected,
the actual meaningful content of the argument.

Exact scoring thresholds are not public to maintain fairness. The more logical, sincere, and original your argument is, the greater the chance the gate will open.

What Technology Does It Use?

Behind the scenes, a large language model analyzes every message. It evaluates:

meaning,
tone,
intent,
manipulative indicators.

Every message undergoes real semantic analysis — the same AI evaluates every player.