HouseKnows methodology: where the data comes from and how we read it
HouseKnows logo HouseKnows
Reviews Blog Русский ქართული

Methodology

HouseKnows computes metrics from real play histories, not from providers' pages. Here is what that data is, what exactly we measure, its limitations and what we deliberately never promise.

What data we use

Real betting histories that players export from their casino accounts (JSON/CSV), plus our own manual sessions. We read the raw export — every spin, stake and payout — not a slot's marketing page. The data is grouped by account, provider, game, stake, payout and bonus events.

What we measure

Observed values on a given sample: the actual return (how much the slot paid back against total stakes), distances between bonus games, bonus frequency and size, distributions. These are observed, not theoretical numbers: we show what happened in the data, not what 'should' happen.

Sample and period

A number's reliability depends on volume: the more spins and sessions, the steadier the picture. Where relevant, the page states the sample size and period. A small sample is noise, and we treat it as such.

How we compute it

The calculations run on HouseKnows' own models. We do not disclose the method — that is what sets the platform apart and what catalogues don't have. What is verifiable is the rest: we don't take ready-made numbers or retell the provider — everything is computed from real histories you can check against the original export.

What we don't do (an honest frame)

A slot's RTP is below 100%, and over the long run the house is ahead — that is maths, not opinion. We don't sell 'systems' or 'how to win', and we don't promise any guaranteed result. Observed is not guaranteed; the past does not decide the next spin — the random number generator is independent.

Limitations and error

Observed RTP on our sample is not the theoretical RTP over an infinite distance. Rare big wins move averages a lot, so short samples are deceptive. Any conclusion holds for the described data and period, not 'in general'.

Data updates

Metrics are recomputed as new histories arrive; the date of the latest recompute is shown on analytical pages.

Author and responsibility

Author and data owner — ХАОС, creator of HouseKnows. Questions and corrections about the data — via the project's Telegram channel.

18+. The materials are analytical and are not a call to gamble. If play stops being entertainment — take a break and seek help.