Claw machines are often less games of skill as chance

September 5, 2015 2:53 AM

Here's some information on claw, also called crane, machines. Did you ever play one of those games to try to win a plush toy, only for the prize to tumble out of the claw at first grasp, or even outright drop it halfway up? Think you just didn't play well enough? Maybe the machine just decided to make you lose by randomly reducing claw power. Here's more information. Via Dubious Quality, although I did page through a few manuals on my own....

Some months ago (before the linked article and video were published) I binged on reading claw and crane game manuals online, in order to write a post on claw games that cheat. I didn't make the post back then because with everything I did, I felt it fell too far under the category of my own work. I have been waiting for someone else to write an article on this so I could present my own findings, some of which I present below.

If you want to see the adjustments for cheating in claw game manuals with your own eyes, follow the links below and be prepared to search, as they're not always easy or obvious to find. When looking through crane game manuals and sifting through material like maintenance and pricing information, it helps to look for the Adjustments sections. Be on the lookout especially for settings that say "bonus power," "win percentage," or "games to win," it is generally difficult to implement win reflexing functions with confidence without sometimes monkeying around with claw power during play.

Panda Vending Treasure Center (PDF) Description of Main Functions: Weak Claw Control: "If the Rate of Weak Claw Happening = Controlled Times/Controlled Base = 1/5, there is one chance of weak claw happening in every five times game. When it's controlled by weak claw power, the claw has not enough power to grab toy, and the players can't catch it. In the other four times, the claw will grab up the toy firstly by strong power, then change to weak power according to the set time and the toy will fall unconsciously, thus to tempt the players to try more times."

ICE "Winner Every Time" Crane Game Manual (PDF) Contains an "auto-percentaging" mode that monkeys with the crane power to manipulate how often the game is won. The operator specifies a percentage from 20 to 50 percent; if the game hasn't been won enough to match that percent, then the claw will be at full strength more often, and if it's been won more than the target percentage, it'll be at full strength less often. See page 13 of manual. This kind of feature is common in these machines; I've even seen one manual that lets operators randomly drop prizes -- sorry, activate low power -- while the claw is on its way up, but lost the link to it. (Aside: In pinball, these kinds of techniques are used to dynamically adjust extra ball and replay difficulty over many games. But of course, pinball is a much more dynamic game that can change in complex ways over time, and there are usually no prizes.)

LAI Games' "Stack N Grab" machine (PDF)
This option is for adjusting the game payout and sets the Average Games played per Prize Won. The game sets the game difficulty to achieve a player win ratio that will average very close to the (AGW). Example an AGW setting 20 means on average a prize will be won for every 20 games played. The setting is adjustable from 1 to 200. With adjustment from 1-30 in steps of 1 and from 30-200 in steps
of 5. The default setting is 10, or one win every 10 Games which suitable for $4 Plush on $0.50 per play." Page 25, Setting P09, "Average Games Per Win (AGW)"

Coastal Double Play Crane Operation Manual (PDF) "This Feature will allow you the operator to choose the amount of game plays until you want the claw to receive full strength. When set to (0), full power will only be received every 256 times. Factory default is (10)" Page 3, Inner Value, Item 9, "Number of times for claws’ strong power given as bonus." Note, the "bonus power" is typically necessary in order to win. No information is given in the manual for the effects of the default setting of 10.

LAI Games Toyzone machine manual (PDF) This manual doesn't describe the effects of many settings, and so doesn't have such an overt smoking gun in the text. It does, however, list as one of its adjustments "Games To Win," and does carefully explain how to set two power levels for the crane: "STD Power," which should be just enough to pick up a "Plush Prize," and "Minimum Power," which should cause it to drop. Claw setting instructions are on page 17.

ELAUT E-Claw machine manual (PDF) This manual is full of engrish. The adjustments on page 9, which talk about how to set up the claw so that it automatically figures out how much reduced claw power causes it to drop a prize, is kind of interesting.

posted by JHarris (31 comments total)

One of the first manuals I read, which I've been unable to find since I first saw it some months back, outright talked about player "win fate" and "lose fate." "Lose fate" meant the claw would reduce power midway up. Win fate meant, not a guaranteed win, but that the machine wouldn't reduce power, and it would be merely possible to win.

I am not sure all crane games do this. Sega's UFO Catcher manual seemed lacking in these kinds of settings. But I'm convinced some crane games do it even if the manual doesn't refer to it. Some machines that support reducing claw power to try to trigger losses allow you to disable that feature. And I ran into one manual that outright stated a few of its features were illegal under certain jurisdictions, and to disable them if that was the case at the place of operation!
posted by JHarris at 2:58 AM on September 5, 2015

Also, some manuals allow setting two power levels, "bonus power" and "normal power." Bonus power would be enabled on some small fraction of games (usually one over a power of two, like 1/16 or 1/32). But the power levels for each state were technically adjustable, so nothing stopped an operator making the normal power as strong as bonus. But it doesn't take a lot of thought to figure out how the feature would typically be used.
posted by JHarris at 3:09 AM on September 5, 2015

Warning: flipping this switch concealed within an inspection panel will make you more money, but may be illegal! Not that anyone's going to check! So don't do it!
posted by Joe in Australia at 3:40 AM on September 5, 2015 [2 favorites]

...this isn't how everyone has always assumed it's worked? If anything surprises me, it's that some crane games sometimes don't cheat.
posted by Dysk at 4:53 AM on September 5, 2015 [7 favorites]

States regulate slot machines to make sure they're not rigged too unfairly against players. But they rarely do the same thing for claw machines.

A focus of our society is the indoctrination of children into normalizing and accepting economic exploitation. They ought to give out introductory credit card offers with every prize.
posted by XMLicious at 4:54 AM on September 5, 2015 [7 favorites]

...this isn't how everyone has always assumed it's worked?

No. Probably not all or even most adults, but I can assure you that kids don't think so - and they're the target audience.
posted by Joe in Australia at 4:56 AM on September 5, 2015 [5 favorites]

This pisses me off. As a six year old I fanatically attempted to win prizes on one of these machines in Asbury Park N.J.

Now I am finding out someone was ripping me off? I hate people who screw with little kids.
posted by notreally at 5:10 AM on September 5, 2015 [3 favorites]

There is one way to win!
posted by sammyo at 5:11 AM on September 5, 2015 [8 favorites]

Related. (I could swear I saw this posted here but couldn't find it with a cursory search.)
posted by TedW at 5:19 AM on September 5, 2015

I figure after one play got me five Mega Man dolls (one of 'em is still on my desk at work) in a crane game on the boardwalk in Point Pleasant, I was done with my crane game luck for life.
posted by graymouser at 5:20 AM on September 5, 2015 [4 favorites]

No. Probably not all or even most adults, but I can assure you that kids don't think so - and they're the target audience.

It was common knowledge for me and the other kids I grew up with in Hong Kong (where crane games were also very commonplace).
posted by Dysk at 5:23 AM on September 5, 2015 [1 favorite]

I think it was always a sort of urban myth that there was a way to consistently win at crane games. But it's pretty clear just watching one of the machines that the claw is designed to go weak and drop the toy almost all of the time.

I'm much more of a fan of the 'coin pusher' type games, where it often feels like you're getting your money's worth in entertainment-value. Games that give back a reasonable percentage always seem a lot more fun to me (and my kids).
posted by pipeski at 5:44 AM on September 5, 2015 [1 favorite]

RNGesus gave, and RNGesus hath taken away...
posted by ethansr at 7:58 AM on September 5, 2015 [1 favorite]

My brother actually consistently won at these as boardwalk games in the 80s, getting a whole bunch of decent prizes. Is there anything about the history of the games - for instance, has technology made them more adjustable today? It seemed in the past they worked a much larger percentage of the time.
posted by Miko at 8:04 AM on September 5, 2015

My best friend in high school claimed to know some secret that let him win repeatedly at a claw machine at Dennys. The rear windshield of his car was STACKED from bottom to top.... He'd win over and over. The rest of us? Never won anything at all - until he pulled the "Let me try for you" and would win within 2-3 tries. Every. Freaking. Time.

I had the one he won for me on my desk for years as a reminder of this, as it stuck with me so well. The lesson I got from it was that unless I was him, I shouldn't play games of chance.
posted by MysticMCJ at 8:52 AM on September 5, 2015 [1 favorite]

This claw machine is perfectly rigged.
posted by elgilito at 9:25 AM on September 5, 2015 [4 favorites]

The claw a metaphor for God, for Authority, for Mom and Dad.

Why, Mom and Dad, why?
posted by LastOfHisKind at 9:38 AM on September 5, 2015 [1 favorite]

sammyo, flagged as fantastic; I have found my new hero and role-model. It is only through direct action that the rigged and phoney invisible claw of the market can be overcome.
posted by howfar at 10:33 AM on September 5, 2015 [1 favorite]

On winning streaks, here are some possible explanations:

1. As said before, it's possible for claw machines to be set up fairly.
2. There is always confirmation bias: you notice winning streaks a lot more readily than losing streaks. People tend not to tell stories of the time they lost several in a row.
3. Miko: This kind of rigging depends on a number of things that microprocessors enabled, so back in the 80s it would seem that claw machines would be a lot less likely to be rigged.
4. MysticMCJ: As I mentioned before, some machines "reflex," adjust their winning chances automatically based on how often they've been won. One effect of this system is, when people lose at such machines, they subtly increase the chances of winning for subsequent plays. If people lose at such a machine a lot, the win percentage will drop. Even rigged crane games don't aim to produce no winners, because people won't play them; they seek to condition players, to create an experience that encourages play (footnote: see below). So in effect, if a lot of people lose at these games, subsequent skilled players can reap the rewards. In other words, MysticMCJ's friend was unknowingly benefiting from his friends losing. Since lack of skill can cause losing streaks, skilled players can have winning streaks, if a lot of other people have lost lately/recently.

Footnote-- Some personal design philosophy: this "creating an experience" when behavior is supposed/expected/promised to be random is fundamentally dishonest, but it's something that computer game design has been toying with a lot more often recently. Like games that adjust events to make themselves easier when you die several times, or intentionally make you fail if you've won several times in a row. Mark my words: this is going to become a big problem in coming years, and some ways it is a problem now.
posted by JHarris at 11:21 AM on September 5, 2015 [3 favorites]

Here's a link to Matt Magnone's claw machine videos, which was mentioned in the video in the OP.
posted by filthy light thief at 12:47 PM on September 5, 2015 [1 favorite]

One effect of this system is, when people lose at such machines, they subtly increase the chances of winning for subsequent plays. If people lose at such a machine a lot, the win percentage will drop.

Sometimes it's not so subtle. I worked on one redemption game (one of those "stop-the-light-with-the-button" games) and the algorithm for stopping the light looked like this:

Check the stopped lamp position:
If the Light is the Jackpot light:
   Does the payout cause the average payout % to exceed the operator setting?
    Yes: Move the lamp one spot over and stiff the kid.
    No: Pay the Jackpot.

posted by JoeZydeco at 2:47 PM on September 5, 2015 [3 favorites]

Yeah, I've heard rumors about the way those jackpot light-stopping machines work in the past, although I haven't looked for hard data. Really, at this point I'm unwilling to trust any electronic prize-giving amusement machine that doesn't have completely open source workings.
posted by JHarris at 3:01 PM on September 5, 2015

I watched a video where this was demonstrated. You had to stop a square of light in the right position, over the tower you had built earlier. The guy slowed it down to the point where you could literally see the square of light skip over the necessary point rather than give you a chance to win.
posted by Joe in Australia at 4:30 PM on September 5, 2015 [1 favorite]

Really, at this point I'm unwilling to trust any electronic prize-giving amusement machine that doesn't have completely open source workings.

There are a few states that actually wrote in laws requiring redemption games to pass third-party lab validation just like casino slot machines do. Iowa and New Jersey come to mind but that may have changed since my days in coin-op. I recall having to change the stop-the-light game for NJ to get rid of the gaff and give the jackpot light a "fair" chance each time.

The games I worked on would be settable in the 20% to 30% range. I recall Chuck E Cheese coming along at some point in time and mandating a higher payout (40%?), probably because enough parents complained about getting virtually no tickets for their cash. AFAIK it's still 1 ticket = $0.01 when you look at the merchandise counter. Do the math on that one next time you're trapped there for a birthday party.
posted by JoeZydeco at 5:13 PM on September 5, 2015 [1 favorite]

At least give them their correct name - Teddy Cranes.
posted by essexjan at 1:23 AM on September 6, 2015

At least give them their correct name - Teddy Cranes.

Huh, never heard this name before, even when living in the UK for several years. They seem far more prevalent in the U.S., though.
posted by grouse at 3:13 PM on September 7, 2015

JHarris: "On winning streaks, here are some possible explanations: "

There's another possibility: The games are rigged to reduce claw power. If you watch skilled players, you will see that 1) they don't pick toys that require much claw power, and 2) they hook them in ways that even if the claw reduces power, the toys are still held. (i.e they hook on the edge of the claw, not between the two claws, etc.) For example, look at the different ways the UFO catcher expert wins in this clip (skip the first 30 seconds, as that's just video of people losing, a la "Don't you hate when this happens?" infomercials)
posted by Bugbread at 11:36 PM on September 7, 2015

Even skilled players cannot overcome a too-weak claw; you can't always hook a prize on a prong of the claw, and even if weak claw is barely strong enough to pick up the lightest toys, those aren't always on top or available, and you're talking about settling for the cheapest prizes in the machine. It's still a random aspect that the machine hides behind skill, basically telling players they're bad when a lot of it is random chance. How is a player supposed to get better if he doesn't have consistent information on his performance?
posted by JHarris at 10:24 AM on September 8, 2015

Here's the video of someone playing Stacker and being cheated by the machine. He's also got a video of him cleaning out a claw machine. An almost empty claw machine, but still.
posted by Joe in Australia at 4:03 PM on September 8, 2015

Oh, and here's he's got a shot of the owners' manual that says how to set the payout ratio. Done and dusted.
posted by Joe in Australia at 4:10 PM on September 8, 2015

I wasn't defending the claw weakening, JHarris, just pointing out that there can be skill-based winning streaks that aren't dependent on others losing or confirmation bias.

On a tangent, friends of mine have reported that recently, in arcades here in Japan, arcade staff members will often set up prizes in optimal positions to be easily won (as in, they will open up the actual machine right in front of you and put something in a can't-lose position, like where just hitting it with the claw will cause it to topple over into the win bin). So it's apparently now extremely unlikely that you will go home without winning a prize. I don't know if it's because they're trying to recapture lost business, or using crane games as loss leaders for other games, or what.
posted by Bugbread at 4:39 PM on September 8, 2015