About the lab
Slotester Lab measures slot games on their official demo servers and publishes what the logs say. The rig plays, the editor checks, the data stays open. This page is who is behind it, how it is run and how it is paid for.
What the lab is
measurement first, opinion secondSlotester Lab does one thing: it runs slot games on their providers' official demo servers, records every server response, and publishes what the numbers say. A review here is not an impression of a game after an evening of play. It is a report on a fixed procedure: full logged sessions of ordinary rounds, a series of bonus purchases where the game sells them, one recorded session, and a set of statistics computed from the log with their uncertainty attached.
The testing is done by software. The judgement of what the numbers mean, and the responsibility for what is published, stays with a human editor. Both facts are printed on every page, because the reader should know which parts were measured and which were written.
Three parts
rig · editor · open dataThe rig
A test harness written in Python. It speaks the game server's own protocol, so a run is thousands of real server rounds rather than a simulation, and it drives a real Chrome window for screenshots and video. The rig also drafts the first version of each review from the log.
closed sourceThe editor
One person reads every figure in the draft against the log, checks the screenshots and the video, rewrites what is unclear and decides whether the run is fit to publish. Nothing goes live on the rig's word alone.
human sign-offOpen data
Every review links to its rounds file with a SHA-256 checksum and explains how the numbers were computed. Anyone can recompute the RTP, the hit rate or the biggest round from the same file.
downloadable logsPrinciples
the rules the lab holds itself to- 01No money from game studios. Providers do not pay for tests, cannot commission them and are not told in advance which game is next.
- 02Numbers are never edited. If a run looks odd, it is re-run or investigated, and the reason is written into the review. A measured 80% stays 80%.
- 03One sample, said out loud. Every review carries a confidence interval and the sentence that a different session can land elsewhere.
- 04Same procedure for every game. Same spin count, same purchase count, same bet logic, same checks. Comparisons across reviews mean something because the method does not move.
- 05Open data, closed rig. The logs are public; the software that produces them is not, because a published rig is a rig that can be gamed.
- 06No invented people. Reviews are signed by the lab, not by fictional experts with stock-photo faces.
Numbers to date
live from the published tests- slots tested 163 2026-09-07 to 2026-10-16
- rounds logged 2,820,000 base series only
- bonus buys 10,200 113 games sell the feature
- natural bonuses 18,807 triggered without buying
- mean measured RTP 95.9% declared average 96.38%
- hit rate range 9% – 43% longest dry run 114 spins
| Slot | Tested | RTP declared → measured | 95% interval | Buys | Log |
|---|---|---|---|---|---|
| Razor Shark | 2026-10-16 | 96.7 → 82.8% | 55.4–110.1% | no buy | 1eeaa33318… |
| Jammin' Jars | 2026-10-16 | 96.8 → 76.6% | 53.8–99.3% | no buy | 18b3b27c2e… |
| Rich Wilde and the Shield of Athena | 2026-10-16 | 96.3 → 93.8% | 76.8–110.8% | no buy | 6a65e98ea0… |
| Goldilocks and the Wild Bears | 2026-10-16 | 96.8 → 92.5% | 82.4–102.7% | no buy | a7d4398b4e… |
| Beat the Beast: Griffin’s Gold – Reborn | 2026-10-16 | 96.1 → 88.5% | 65.5–111.6% | 50 · 1.10x | 315651914f… |
| Diamond Mine Megaways | 2026-10-16 | 96.4 → 85.9% | 74.3–97.6% | no buy | 57866a63cf… |
| Kingmaker | 2026-10-16 | 96.7 → 88.7% | 76.7–100.6% | no buy | dc4f16040f… |
| Iron Bank 2 | 2026-10-16 | 96.1 → 88.3% | 68.1–108.4% | 50 · 1.28x | 5655220230… |
| Gates of Olympus Super Scatter | 2026-10-16 | 96.5 → 99.7% | 74.2–125.2% | 50 · 1.51x | 5714842227… |
| xWays Hoarder xSplit® | 2026-10-16 | 96.1 → 97.2% | 71.4–122.9% | 50 · 0.93x | 6fc1500aa7… |
| Rise of Merlin | 2026-10-15 | 96.6 → 103.9% | 81.0–126.8% | no buy | f5f7a5d46e… |
| Bonanza Falls | 2026-10-15 | 96.6 → 107.9% | 83.8–132.0% | 50 · 0.86x | bff74e4d95… |
Latest 12 of 163 tests shown. Full register with sorting on the home page
Timeline
how the lab got here-
Rig 0.1.0 and the first review
The harness learns the Pragmatic Play protocol end to end: session, spins, cascades, free spins, purchases, media capture. Gates of Olympus is the pilot review.
-
The standard is set
Full-length runs with a purchase series per game replace the shorter pilot runs. The pilot is re-run to the new standard before anything else is added.
-
Ten games in one night
Cluster games, respin features and 1000x editions force protocol fixes; a balance-reconciliation check is added and applied to every run retroactively. Ten reviews are live by morning.
-
The register and these pages
Reviews are cross-linked through a lab register, and the documentation pages are rewritten around live data instead of boilerplate.
-
Second protocol: BGaming
The rig learns the BGaming demo server: one JSON request per round, three purchase levels, the max-win odds the studio publishes in the game config. Bonanza Trillion is the first review from a second studio.
-
Third protocol: Relax Gaming
The rig learns the Relax Gaming demo server: a whole round, respins and bonus included, arrives in one response, and a bought feature is the same request with a price attached. Money Train 4 is the first review from the third studio.
-
Fourth and fifth protocols: Nolimit City and Quickspin
The rig learns two more demo servers: Nolimit City, where rounds travel as encoded frames over a WebSocket connection, and Quickspin, where a whole round, free spins included, comes back in one response. Test runs from both studios can now be queued.
-
Partner studios on the Relax server
The Relax Gaming demo server also hosts games from partner studios, so the same protocol now covers Big Time Gaming and Push Gaming titles. Danger High Voltage 2 is the first review from a partner studio, credited to Big Time Gaming with the Relax platform named in its method.
-
Sixth protocol: Play'n GO
The rig learns the Play'n GO demo server: a demo login over HTTP, then every spin, feature step and collect as short numbered commands over a WebSocket hub. The studio's demo sells no features, and a session that expires mid-series is reopened. Book of Dead is the first review from the studio.
-
Seventh protocol: Blueprint Gaming
The rig learns the Blueprint Gaming demo server, where every round, and every free spin inside it, is a pair of XML messages, and a feature purchase is a side bet of 99 stakes placed with the spin. Fishin' Frenzy Megaways is the first review from the studio.
-
Eighth protocol: Thunderkick
The rig learns the Thunderkick demo server: JSON service calls that open a round and continue it step by step until the server marks it over, with the feature-buy prices read from the game config. Midas Golden Touch – Reborn is the first review from the studio.
-
Ninth protocol: Hacksaw Gaming
The rig learns the Hacksaw Gaming demo server: one bet request per round, a confirmation after every win, and bonus games bought by their index in the game's own list. Le Bandit is the first review from the studio.
-
First Push Gaming reviews
Jammin' Jars and Razor Shark are the first Push Gaming titles in the register, played through the Relax Gaming protocol like the other partner studios.
How it is funded
the short versionThe site can carry partner links to online casinos under the demo of a game. If a reader signs up through such a link, the lab may earn a commission. That is the only revenue model. Partner links never touch the measurements, the order of reviews or the wording of a verdict, and they are marked as sponsored wherever they appear.
Why one name
no personas, one responsible editorReviews are published under the lab's name rather than under invented reviewer profiles. The measurements are the same whoever presses the button, and a fictional byline would add a face without adding accountability. Corrections, questions and requests all go to the same place, and the editor who signs off on a review is the one who answers for it.