When the cash in the box doesn't match what the machine says it earned, the honest answer is: it's one of three things — a counting error, a hardware fault, or someone stealing — and you can usually tell which by comparing three independent numbers that should all agree, and seeing which one breaks. Don't jump to "I'm being robbed" until you've lined them up. Half the panicked messages I get about arcade machine accounts that don't add up turn out to be a meter the operator was reading wrong, or a free-play credit nobody remembered leaving on.
I'm Engineer Wang. I build and harden arcade machines at my own factory in Guangzhou, and I've spent years helping operators chase down exactly this gap. Here's how I work the problem.
Every machine gives you more than one way to count its earnings, and the whole trick is that honest numbers agree and a problem makes them disagree in a specific direction.
The hard coin/bill meter — the mechanical or electronic counter wired to the acceptor. It counts inputs and is hard to fake without opening the cabinet.
The soft meter on the game board — what the game software thinks it took in and paid out. You read it in the operator/bookkeeping menu.
The actual drop — the real cash you pull from the box, plus tickets or prizes dispensed valued out.
Pull all three for the same period. On a clean machine they reconcile within a tiny margin (jams, test coins, the occasional miscount). When they don't, the gap pattern tells you where to look:
Drop is short, but both meters look normal. Cash left the box without the machine knowing. That's straight cash handling — staff, route, or the box itself — not a game exploit.
Soft meter shows income the coin meter doesn't. Credits were added without coins going in. Classic score-theft territory: a device or trick that puts credits on the board directly.
Coin meter climbs but income is low and payouts are high. The machine is paying out more than it should per dollar in. Either the payout setting is wrong, the dispenser is faulting, or someone's forcing wins.
All meters agree with each other but disagree with last month at the same traffic. The leak is upstream of the counters, or your baseline assumption is off.
That last one is why a baseline matters more than any single reading. If you don't know what a machine normally clears on a normal week, you can't recognize "doesn't add up" in the first place.
I push every operator through this list before we talk theft, because it's free and it's usually the answer:
Free play / test mode left on. A tech sets a machine to free play to service it and forgets. Income craters, traffic looks fine. Check the operator menu first.
Payout / win-rate set wrong. A redemption or fish payout percentage bumped during setup quietly bleeds margin with no visible "event."
A failing coin or bill acceptor. A dirty or worn acceptor double-counts, under-counts, or rejects good money. The meter and the drop drift apart with nobody touching anything.
Ticket dispenser miscount. Dispensers that skip or run long throw the payout side off without any cheating.
Bookkeeping itself. Reading the meter on the wrong machine, mixing up reset vs. lifetime counters, counting a period that isn't aligned to the pull. More common than people admit.
If you want the wider version of this triage — slow bleeds, overnight drops, and how each boring cause looks different from an attack — I wrote it up in unexplained losses on arcade machines and what causes them.
Once the boring causes are out, certain gaps only really make sense as deliberate theft. The signature of credits appearing on the board without matching coin input is the big one — that's money created out of nothing, and a faulty acceptor undercounts, it doesn't invent income on the soft meter. When I see the soft meter running ahead of the coin meter on one cabinet and not the floor, I treat it as score theft until proven otherwise.
This is where understanding the actual move helps you read the gap. The physical methods that produce a "credits without coins" pattern — small cards on the acceptor, jumpers on the board, ribbon-cable tricks — are explained at an awareness level in how score theft on arcade machines works. You don't need to know how to do it; you need to know what it leaves behind so you can spot it. Pair that with a cabinet inspection from signs a fish game machine has been tampered with and you'll usually find your answer in the wiring or the acceptor.
The timing matters too. If the gap appears in clusters — fine on weekday mornings, off on Friday nights — you're probably looking at a person who shows up on a schedule, not a steady hardware fault. A fault bleeds constantly; a cheater bleeds when they're present.
Here's the limit of everything above: meters tell you that money went missing, never when or who. You reconcile on Monday and the move happened Saturday at 9 p.m. The evidence — the hand on the acceptor, the card on the board — is long gone. That's why a lot of operators reconcile a loss for months and never catch the cause.
A Score Theft Detection Unit sits at the machine and watches for the exact moves that create a credits-without-coins gap: physical tampering at the acceptor, the board, the ribbon, and abnormal credit-ups. The moment it sees one it alerts and logs a timestamped record, so instead of a mystery gap on Monday you get "here's the machine, here's the time, here's what happened." Gen 1 covers a single cabinet (about 1 to 1.5 m); Gen 2 reaches roughly 2.5 to 3 m for tighter clusters. It's plug-and-play, no wiring, and it doesn't touch normal gameplay. How that detection works in plain terms is in how our arcade anti-cheat detection devices work, and the full diagnostic map for symptoms like this is the warning signs hub.
If your accounts aren't adding up and you've already cleared the boring causes, send me the gap pattern — which of the three numbers is off and in which direction — along with the machine type. I'll tell you whether it reads like theft or a fault and what to check next. WhatsApp or WeChat +86 17620842078, or use the Contact Us page.