Are Access Memories of Computers Truly Random? - Algorithm Week


By Mark Falanga

Image by Rennett Stowe.

Computers are good at a lot of tasks, most tasks really. They can organize music, stream entire movies, complete the most intricate spreadsheets, sort data, and so much more. However, according to Steve Ward, Professor of Computer Science and Engineering at MIT’s Computer Science and Artificial Intelligence Laboratory, they’re bad at one thing… flipping a coin.

“They’re deterministic, which means that if you ask the same question you’ll get the same answer every time,” he said on MIT School of Engineering’s website. “In fact, such machines are specifically and carefully programmed to eliminate randomness in results. They do this by following rules and relying on algorithms when they compute.”

Ward goes on to explain that computers can be programmed to generate “random” numbers, emphasis on the quotation marks. “Typically, that means it starts with a common ‘seed’ number and then follows a pattern,” he said. This type of number is called a pseudorandom number. If this is how random numbers are generated then it would seem theoretically possible to predict what number they’ll choose.

This was the thought of Kevin French, a character in the short lived, but much beloved, adult cartoon, Mission Hill. In the episode “Kevin vs. the SAT”, he fears that a score of 1500 would not get him into his desired school, Yale University. He decides that he must get a perfect score of 1600, not by studying, but by finding a repeating pattern in the computer that chooses the answers for the SAT.

Image courtesy of College Guide.

Following the pattern of a pseudorandom number generator, he feeds old SATs into his computer to see if his computer can find a repeating pattern. Unfortunately, he needs more data than just a few tests and, in a satirical cartoon fashion, ends up going slightly insane in the process.

But pseudorandom numbers can be employed for purposes much more valuable than cheating. They also serve a vital function for states that have legalized gambling. Yes, it’s a tiny chip implanted in every slot machine and video poker machine that determines whether or not you hit the jackpot or have to try again.

BTR attempted to contact the New Jersey Board of Casino Control and the Pennsylvania Gaming Control Board to discuss the topic of these number generators, but both refused our request for an in-depth interview. However, both did acknowledge that the random number generators in their machines are set at the legal limit.

So how do they work? Well, each time the player presses the button for the reel to spin, the number generator produces a number anywhere from one to 1 billion and the slot machine compares it to the number based on the outcome of the spinning reels. If it’s not a match, the player loses and is awarded no coins or credits. If they match, then the appropriate amount of money is issued.

Unlike mechanical slot machines from the early part of the 20th century, the exact millisecond you pull the lever matters greatly, as the number generator comes up with hundreds of numbers every second, even when the game is not in use! So don’t feel cheated if the next person on your slot machine hits the jackpot right after you left.

Since these machines use pseudorandom number generators instead of true random number generators, could it be possible to cheat the system, find a pattern, and break the bank in Vegas? Don’t bet on it. You could always try to hack online poker though.

However, this still doesn’t answer our question: Can machines truly produce random numbers? As it turns out, yes they can… only if it’s not programmed. Sounds tricky, but the principle behind it is not. Sites like, use atmospheric noise in their calculation of a random number. The genuine random number generator works like a pair of dice attached to your computer, picking up noises created in everyday life, from background office noise to thunderstorms. Those waves then activate the true random number generator to create a completely random number.

So despite our technological advancements and complex algorithms, it seems that computers can only think of something purely random when they enter unpredictable reality. This idea sort of humanizes aspects of machinery, and for once gives us the satisfaction of knowing that we can do something better than a machine… even if it is completely random.