Most operators don't catch a cheater in the act. They catch a pattern: a machine that used to clear a steady number every week and quietly stopped, a ticket count that doesn't match the prize cabinet, one regular who wins on a game nobody else can beat. The earliest warning signs your game machine is being cheated almost always show up in the numbers and in small physical changes, not in some dramatic moment on the floor. This page is the map. It walks through the signals worth watching, points you to the deep dive on each one, and helps you tell a real attack apart from bad luck or a setup mistake.
I'm Engineer Wang. I've spent fourteen years building and hardening arcade game-machine hardware out of my own factory in Panyu, Guangzhou, and I run a small attack-and-defense research group that breaks these machines so we can teach operators how to defend them. Almost every "my machine is acting strange" message I get traces back to one of the signs below. Read the cluster that matches what you're seeing, then follow the link into the full write-up.
This is the cluster that catches the most operators, because the symptom is the same whether the cause is theft, a hardware fault, or sloppy counting: the cash and the meters disagree.
The first thing to rule out is whether your books are wrong before you assume someone's stealing. Coin meters, the soft meters on the board, and the actual drop should tell the same story across a week. When they drift apart consistently on one machine, that's your flag. I go through how to read those three numbers against each other, plus the common bookkeeping traps that look exactly like theft, in why my arcade machine accounts do not add up.
A slower, scarier version of the same problem is the gradual fade: same foot traffic, same hours, but the take on one or two machines keeps sliding week over week. That usually means a recurring exploit, not a one-time grab. The overnight version, where a machine clears normally on Friday and barely registers Saturday, points at something that happened in between. Those two are covered in game machine revenue dropped overnight and what to check and in the broader unexplained losses on arcade machines and what causes them, which separates cheating from the boring causes — a failing acceptor, a wrong payout setting, free-play left on — before you go hunting for a culprit.
The second cluster is physical and behavioral: signs that someone has been inside the cabinet, attached something to it, or learned to drive it in a way the game was never meant to allow.
Fish game tables and other high-payout cabinets draw the most attention from cheaters, so they're the ones to inspect first. Fresh tool marks on the lock or door edge, a coin-door screw that won't sit flush, a ribbon cable reseated at an odd angle, a board that runs warmer than its neighbors — small tells add up. I list what to actually look for, in inspection order, in signs a fish game machine has been tampered with.
Sometimes the tell isn't damage but an addition: a small board spliced into the wiring, a sensor stuck to the cabinet, a device that wasn't there at the last service. Knowing where these get hidden and how to spot one on a walk-through is its own skill, and I cover it in how to tell if someone installed a cheat device on my machine.
If you want to understand why these physical changes work — what an attacker is actually after when they open a coin door or splice a cable — the awareness-level breakdown lives in our overview of how arcade machine cheating actually works, so you can recognize the attempt instead of just reacting to the damage.
The third cluster is the human one. A cheater doesn't always touch the machine. Sometimes they exploit a result that's been leaked to them, or they feed the game a sequence it was never built to handle.
The classic question here is the regular who keeps winning. Skill is real, and so is variance, so you can't accuse someone over a hot streak. What you can do is watch for results that beat the math the game is supposed to enforce: wins on outcomes that shouldn't be predictable, timing that's too clean, a pattern that survives across sessions. I lay out how to separate genuine skill from a fed result in how to tell if one player who keeps winning is actually cheating.
Redemption and ticket games show their own version: a payout rate that runs hotter than the machine's setting allows. Before you blame a cheater, the dispenser and the payout percentage need ruling out, but a machine that pays heavy only for certain players or at certain times is a different signal. That's the focus of what to do when my redemption machine pays out too many tickets.
Claw and crane machines deserve their own note because operators panic about them the most and are wrong the most often. Most "my claw is hacked" cases are just grip settings, win-rate configuration, or a genuinely skilled player — but not all of them. The line between a tuning issue and an actual attack is drawn in is my claw machine being hacked or just unlucky.
Reading the signs is step one. The reason an experienced cheater can run for months is that nobody's watching the machine when they're not standing in front of it, and by the time the numbers look wrong the evidence is long gone. That's the gap our detection gear is built to close.
The Score Theft Detection Unit watches a machine for the physical score-theft moves — the small cards, jumper tricks, ribbon and coin-acceptor manipulation, abnormal credit-ups — and alerts the moment it sees one while logging a record you can go back to. For the cabinets where someone feeds the board a cheat sequence through the joystick and buttons, the AI Cheat Code Interceptor recognizes that illegitimate input at the machine's input layer and stops it from driving a payout, and it does that without transmitting any radio signal. The V5 / K8 Result Integrity Monitors watch for the suspicious wireless and data activity around a cabinet that points to leaked results, alerting and saving evidence when they see it; they detect and record, they don't transmit interference or block anyone's signal. If you want to understand what these actually catch before deciding which fits your floor, start with how our arcade anti-cheat detection devices work, and weigh it against the real numbers in stopping revenue loss from arcade cheating.
If you're seeing one of the patterns on this page right now, the fastest way through it is to tell me what's happening. Send me your machine type and a short description of the symptom — the numbers, the timing, what looks off — and I'll tell you which sign you're actually looking at and what to check next. Reach me on WhatsApp or WeChat at +86 17620842078, or through the Contact Us page. I'd rather talk you out of buying something you don't need than sell you a box that won't fix your problem.