tonify-sims

*🇬🇧 English 🇷🇺 Русский*

English translation of CRITIC.ru.md; the Russian original is the source of truth, numbers are identical.

CRITIC.md — the red team against our own model (v0.2 → v0.3)

A pass over the entire vault. Format: accusation → verdict → action.

1. “Status quo = 0.42 donations/year” — REFUTED, taken off the board

Accusation: the P6 inversion was built on SoundCloud’s 29% FPR, but their 29% is the superfans’ share of ROYALTIES (their subscription money under user-centric), not of donations. A category error. Verdict: the critic is right. The 0.42 figure has been removed from the results; in its place — measured rate-k benchmarks from adjacent industries (see RESULTS §2). (Throughout, rate k = payments per superfan per year; “kef” in the Russian original.)

2. “Plays/listener = 21.21 per year” — CORRECTED

LFM-1b measures user×artist pairs over the panel’s entire history, not per year. 21.2 is an upper bound. Sensitivity added (8–21). The direction of the error is AGAINST the direct economy (the “till”, касса in the original) — conservative: with fewer plays the pool needs even more listeners, and the direct economy’s breakeven drops to 0.38.

3. fig2 “the tail is too smooth” — CONFIRMED, fixed

v0.2 produced deterministic fractional superfans. v0.3: binomial sampling — an artist with 30 listeners now has an honest ~60% chance of zero direct-economy income.

4. fig3 “too pretty a straight line” — CLARIFIED

It is the analytical dilution curve F/(1+F) for a fixed pool with no detection, not a simulation — and it is labeled as such. A caveat has been added: the direct economy has losses of its own (chargebacks, stolen cards), but they are not smeared across the innocent — the qualitative difference stands.

5. “$1000 Loud & Clear is the artist’s pocket” — CAVEAT

It is royalties to the rightsholder. The comparison is valid only in the independent mode (artist = rightsholder). In the signed mode the in-pocket threshold is even harsher.

6. The $5 ticket — CORRECTED TO A RANGE

The only music-specific PWYW experiment: mean €3.10, ≈87% of fixed-price WTP, and the share of outright refusal RISES (24.4% vs 17.3%) — the social norm crushes the variance. Ticket sensitivity: $3.1 / $5 / $6.9.

7. Superfans 1.7% — CORRECTED TO A RANGE

Under direct measurement the 90-9-1 rule turns out to be 97-2-1 (0.6–0.75% active). Sensitivity: 0.6% / 1% / 1.7%.

Red team bottom line

The central conclusion SURVIVED, but became a range: the direct economy’s breakeven against the independent pool — 0.38 … 1.25 … 6.31 payments/superfan/year (min/median/max across 18 combinations). All three axes of the range are exactly the quantities the MVP measures. None of them requires a miracle: Twitch’s recurring mechanics give a paying fan rate k = 12 automatically.