How we test games: we play the free demo, log the mechanics spin by spin, and rely on published JILI figures rather than marketing claims.
Every score starts in the demo
Before we write a word, we load the free Charge Buffalo demo and play it. We log what happens spin by spin: how often the Buffalo Coin scatter lands, what the multiplier wilds carry, how the free-spins counts (8, 15, 25, 100) actually feel over a long session. The score comes out of that session, not out of a marketing page.
We test on a virtual balance, never real money. That keeps us honest about one thing in particular: demo wins aren't real wins, and we say so everywhere it matters. But the demo runs the same maths engine the real game does, so it's the right place to watch mechanics without a bankroll on the line.
How a review gets built
Roughly in this order, though we loop back whenever a fact and an observation disagree.
1. Play
We load the demo and run enough spins to trigger the free-spins feature more than once. We watch the wilds land on reels 2 to 6 and note when the x2, x3 and x5 multipliers stack.
2. Log
We write down what we see, spin by spin. Counts, magnitudes, the shape of a dry spell, the one round that pays most of it back.
3. Cross-check
We line up our notes against published JILI specs and structured slot catalogs. When they agree, good. When they don't, the catalog conflict goes in the write-up.
4. Rate
We set a number, then argue against it internally before it ships. If we can't defend the score out loud, it changes.
Published specs beat marketing copy
Where a promotional page and a structured catalog disagree, we side with the catalog and flag the gap. Charge Buffalo is the perfect example. Some marketing pages call it "high" volatility. The structured ratings call it medium, so we lead with medium and note the disagreement once.
The same rule governs RTP. We quote it as around 97% by default, then add the caveat that operators can run lower-configured builds reported near 96%, so the real figure depends on the casino. We won't print a single flat RTP as if it were carved in stone, and we won't invent a "96.55%" that no source actually gives. False precision is worse than an honest range.
What we verify
The grid and ways (6x4, 4,096), the wild behavior, the scatter, the free-spins triggers and retriggers, the additive multiplier maths, the 4,000x max-win figure, and which studio actually made the game. We confirm each against at least one structured source and, where we can, against the demo itself.
What we flag as unpublished
Hit frequency isn't published, so we don't quote one. The exact per-symbol paytable is reported on two different bases, so we give rank order, not fake exact values. The certification lab for this specific title, JILI's founding year, and a single agreed release date all sit in "not confirmed," and we label them that way.
The name-collision check
One step exists only because this game demands it. Before publishing any figure, we confirm it belongs to JILI's Charge Buffalo and not to a lookalike.
TaDa Gaming, a separate and unrelated studio, publishes same-named buffalo slots with wildly different maths: Charge Buffalo Ascent (5,488 ways, 20,000x) and 3 Charge Buffalo (10,000x). Aggregators bleed those numbers into JILI listings constantly, and a few even mislabel Ascent as "JILI." We treat any 5,488-ways, 10,000x, 20,000x, or expanding-grid claim as a red flag that a TaDa figure has crept in, and we strip it out.
When we get it wrong
We will, eventually. Slots get reconfigured and catalogs update.
When a reader or a fresh source shows us an error, we fix the page and note what changed. Our corrections process is open to anyone. And if new data ever shifts the maths enough to move the verdict, the number moves with it. A score that never changes isn't a sign of confidence. It's a sign nobody's checking.