Methodology specification
How We Test
mobile casinos
Every number in our tables comes from a repeatable procedure, and every gap is printed as a labelled reason instead of being filled with a guess. This page documents the procedure exactly — including its limitations — so you can judge how much weight the numbers deserve.
Test bench
The measurement environment
Lobby measurements run in a controlled Chromium browser configured as a phone: a 390×844 viewport at 3× pixel density, a mobile Safari user-agent, touch enabled. The network is throttled at the browser-engine level to 10 Mbps down, 1.5 Mbps up and 40 ms round-trip latency — a deliberately conservative 4G profile, so the differences between casinos come from their engineering, not from network luck.
Each casino's mobile lobby is loaded three times from a cold cache. We record the largest-contentful-paint time of each run and publish the median, and we count the real bytes that crossed the wire using the browser's network instrumentation — not the resource-timing API, which under-reports third-party assets.
This is device emulation, not a rack of physical iPhones. Rendering times on real hardware will differ by device; relative differences between casinos measured under identical conditions are the meaningful signal.
Fixed parameters
- VIEWPORT: 390×844 @3× · MOBILE UA · TOUCH
- NETWORK: 10MBPS DOWN / 1.5MBPS UP / 40MS RTT
- RUNS: 3 PER OPERATOR · COLD CACHE · MEDIAN
- LOAD METRIC: LARGEST CONTENTFUL PAINT
- WEIGHT METRIC: TOTAL ENCODED BYTES ON WIRE
- LAST RUN: 2026-07-10 10:36 UTC
Store data
App data comes from the stores
Download sizes, minimum OS versions and player ratings are read from the official App Store and Google Play listings at each measurement run — the same figures you would see on the listing that day. We publish them as-is and date them. Store figures move with every app release, so treat the exact megabyte counts as a snapshot and the differences between operators as the durable finding.
Where a brand has no native app, app columns read BROWSER-ONLY and the operator is measured on its browser lobby only. Where our automated matching cannot confidently identify a brand's store listing, the field reads STORE UNLISTED rather than a guess against a lookalike app.
Known limits
Limitations we won't hide
Some operators geo-restrict their websites and refuse connections from our test location. When a lobby serves a block page or a thin gate instead of the real product, our tooling detects it — a block URL, or a page transferring under 0.3 MB — and discards the run. That is why some brands show GEO-FENCED or NOT MEASURABLE in the load and weight columns while their app-store data is present: we would rather publish a gap than a number measured against the wrong page.
Measurements come from one test location, one emulated device class and one network profile. They are honest, repeatable and identically applied — and they are not a substitute for how a specific casino feels on your phone and your connection. Use the columns to shortlist; the block-page detection, run counts and retrieval dates exist so you can trust what the columns claim to be.
What we never do: accept operator-supplied performance figures, rank by bonus size or commercial relationship, or list an operator without a verifiable licence — UKGC, the Kahnawake Gaming Commission, or one of the seven US state regulators. Register links sit beside every table.
Disclosure
How this site is funded
Visit-site links may be affiliate links, which means an operator can pay us a commission when you sign up through them. That funding model is standard for comparison sites; what differs here is that commissions cannot buy a position — rankings are computed from the measured columns you can see, and an operator with no affiliate deal ranks exactly where its numbers put it. 18+. Gamble responsibly.
Query log
Frequently asked questions
Why do you throttle the connection to 4G speeds?
On an unthrottled fibre connection every casino looks fast, and the engineering differences disappear into the noise. Capping bandwidth and adding realistic latency spreads the field: a lobby that ships five times the bytes takes visibly longer, which is exactly the difference a player on mobile data experiences.
Why three runs instead of one?
Single page loads are noisy — a slow ad auction or a cold CDN edge can double one run. Three cold-cache runs with the median reported filters the outliers without averaging them into a fiction. Where runs fail entirely, the run count is recorded and the value is discarded rather than padded.
What do GEO-FENCED, NOT MEASURABLE and the other labels mean in your tables?
Each label states why a number does not exist, printed instead of guessed. GEO-FENCED: the operator blocks our test location outright. NOT MEASURABLE: it answered with a gate or skeleton page too thin to time. BROWSER-ONLY and NO APP: the brand ships no native app. STORE UNLISTED: our matching could not confidently identify the correct store listing. PLAY: VARIES: Google Play declares no fixed minimum OS. A label is a measurement boundary, never a zero.
Do operators know when you measure them?
No. Measurements hit the public lobby the way any phone visitor would, on no schedule an operator could prepare for. Nothing is announced, nothing is whitelisted, and operator-supplied performance claims are never used as a substitute for a run.
Can an operator pay to improve its position?
No. Affiliate commissions fund the site and are disclosed, but position is computed from the measured columns you can see in the table. An operator that wants to rank higher has exactly one route: ship a lighter, faster product and score better on the next run.
Why not test on real iPhones and Android handsets?
A physical device farm measures hardware rendering more precisely, and it is the direction we want to grow. Emulation measures the network and payload story identically across operators, which is where most of the mobile experience gap lives. We label the method honestly so you can weigh it yourself.