Why Scammers Make Their Scripts Deliberately Stupid – The Hidden Economics of Phone Fraud

Optimization Model of Telecom Fraud

Herley (2012) models the scam as an optimization problem. Define: N – number of recipients who see the message. r – reply rate (fraction that respond). p – probability that a respondent is fully defrauded after the interaction. v – average monetary loss per successful victim. c – communication cost per respondent (time, resources).

The expected profit for the fraudster is

Profit = N * r * p * v – N * r * c
       = N * r * (p * v – c)

Because the attacker’s time per victim is limited, the optimal strategy maximizes the term p * v – c rather than simply increasing r. “Low‑quality” scripts that appear obviously fake act as a filter, allowing only the most gullible respondents to proceed, which raises p enough to offset the loss in r.

Law of Large Numbers

Even with a tiny conversion rate (e.g., r * p = 0.0001), sending a massive volume yields substantial profit. Example:

Send 1 × 10⁸ messages.

Conversion rate 0.01 % → 10 000 victims.

Average loss per victim v = ¥50 000 → total gain ¥5 billion.

Marginal cost of bulk messaging is near zero, so profit is essentially the gross gain.

This demonstrates that a low success probability multiplied by a large base (law of large numbers) yields stable returns, breaking the usual risk‑reward trade‑off of conventional crime.

Behavioral‑Economic Mechanisms

Scam scripts exploit well‑studied biases:

Loss aversion : messages threaten account freezing or legal trouble, prompting immediate action to avoid perceived loss.

Certainty effect : framing a prize as “already yours” (e.g., “you have won ¥500 000, pay tax to claim”) makes the offer feel certain, reducing risk perception.

Time pressure : deadlines (“transfer within 30 minutes”) limit deliberation, forcing decisions under stress.

Under acute stress, decision quality degrades for both experts and novices, making the victim’s judgment unreliable.

AI‑Generated Deep‑Fake Attacks (Post‑2024)

Since 2024, real‑time AI‑generated video and voice synthesis enable attackers to impersonate trusted contacts in live calls. Documented cases include:

A businessman transferred ¥4.3 million to a “friend” whose face and voice were AI‑fabricated.

A Hong Kong employee wired HK$200 million after a fabricated video conference with alleged executives.

Tools can create convincing faces from a single photograph and clone voices from short audio clips, rendering visual verification ineffective.

Game‑Theoretic Capture Inequality

Model the interaction as a game where the expected net payoff per scam is E[payoff] = p * v – c – q * L where q is the probability of being caught and L is the expected penalty (e.g., 10 years imprisonment ≈ ¥2 million). As long as p * v – c > q * L the activity remains profitable. With typical values ( p ≈ 0.01, v ≈ ¥50 000, c ≈ 0, q < 0.33, L ≈ ¥2 million), the inequality holds, explaining why cross‑border enforcement alone does not eradicate the crime.

Practical Decision Rules

Empirical testing of three realistic scenarios shows that consistently choosing the “verification” option reduces expected annual loss from tens of thousands of yuan to under ¥5 000, while adding a small positive expected gain (~¥7 500) from the extra verification step.

Three non‑negotiable rules derived from the model:

Immediate‑transfer requests (“right now”, “within 30 minutes”) are 100 % fraudulent.

Financial requests must be confirmed via an independent channel (e.g., call the official number, ask for information only the real party knows).

Transactions exceeding a self‑defined threshold (e.g., ¥10 000) must be delayed at least one night before execution.

These rigid rules bypass emotional judgment, which is unreliable under fear or greed.

Key Takeaway

Telecom fraud is a mathematically engineered system that combines reverse screening, the law of large numbers, behavioral‑economic manipulation, AI deep‑fake technology, and jurisdictional arbitrage to sustain profit. Recognizing the underlying incentives and applying the three hard rules provides the most reliable defense.