[dylan604]

Let's pick up an entire sunken submarine with a claw from the surface by attaching to the center of the boat instead of something applying more even pressure. It's like there was nobody in the room that had ever played the claw vending machine game and saw the futility of it. "The smartest guys in the room" seems apropos

[waste_monk]

>It's like there was nobody in the room that had ever played the claw vending machine game and saw the futility of it

The claw games are deliberately designed to hit a certain payout ratio (e.g. only 1:15 or so, configurable by the operator, attempts will have the claw clamp tightly enough to retrieve a prize).

I would assume the CIA did not include this feature on their claw.

[dylan604]

or did they?? they were able to bring the boat up off the floor tantalizingly to make them thing they were going to raise it at some ridiculous rate of ascent only for the boat to split. they too got fooled by the claw. their parents were out of quarters, so they couldn't try again

[NikkiA]

The machines can force a loss, but they can't force a win. If you're unlucky you're unlucky and the machine will make it up by not forcing losses later with someone with more luck and skill. Statistics give enough opportunity to not need to rig every game, and they only need to achieve a ratio over a long run, usually millions of plays ('we' always ran to 3 million plays as our statistical target, but that was fruit machines, aka slots).

[waste_monk]

I understand many (most?) claw machines these days have sensors in the prize chute and will go into a "strong claw" mode until such time as a prize is successfully dropped. So not quite forcing a win, but certainly making it more fair.

[vesrah]

I thought this happened because the claw was pretty heavily damaged during deployment, only a small bit of it was actually able to grip.

[wkat4242]

Well it actually did work, it was just that the sub broke apart and the pieces they were most interested in fell back to the ocean floor.

[dylan604]

This is like a project manager that agrees to ship code that is known to be buggy and doesn't actually do what was intended, but performs just enough to fool a casual user. No. No. NO. Please, update your definition of working