Computers — Threat or menace?

Sunday, May 29th, 2016

“Ready or not, computers are coming to the people,” Stewart Brand wrote in 1972, in Rolling Stone — which explains the context for what comes next:

That’s good news, maybe the best since psychedelics. It’s way off the track of the “Computers — Threat or menace?” school of liberal criticism but surprisingly in line with the romantic fantasies of the forefathers of the science such as Norbert Wiener, Warren McCulloch, J.C.R. Licklider, John von Neumann and Vannevar Bush.

The trend owes its health to an odd array of influences: The youthful fervor and firm dis-Establishmentarianism of the freaks who design computer science; an astonishingly enlightened research program from the very top of the Defense Department; an unexpected market-Banking movement by the manufacturers of small calculating machines, and an irrepressible midnight phenomenon known as Spacewar.

Reliably, at any nighttime moment (i.e. non-business hours) in North America hundreds of computer technicians are effectively out of their bodies, locked in life-or-death space combat computer-projected onto cathode ray tube display screens, for hours at a time, ruining their eyes, numbing their fingers in frenzied mashing of control buttons, joyously slaying their friend and wasting their employers’ valuable computer time. Something basic is going on.

Rudimentary Spacewar consists of two humans, two sets of control buttons or joysticks, one TV-like display and one computer. Two spaceships are displayed in motion on the screen, controllable for thrust, yaw, pitch and the firing of torpedoes. Whenever a spaceship and torpedo meet, they disappear in an attractive explosion. That’s the original version invented in 1962 at MIT by Steve Russell. (More on him in a moment.)

[...]

The setting and decor at AI is Modern Mad Scientist – long hallways and cubicles and large windowless rooms, brutal fluoresccnt light, enormous machines humming and clattering, robots on wheels, scurrying arcane technicians. And, also, posters and announcements against the Vietnam War and Richard Nixon, computer print-out photos of girlfriends, a hallway-long banner SOLVING TODAY’S PROBLEMS TOMORROW and signs on every door in Tolkien’s elvish Fëanorian script – the director’s office is Imladris, the coffee room The Prancing Pony, the computer room Mordor. There’s a lot of hair on those technicians, and nobody seems to be telling them where to scurry.

[...]

I’m guessing that Alan Kay at Xerox Research Center (more on them shortly) has a line on it, defining the standard Computer Bum: “About as straight as you’d expect hotrodders to look. It’s that kind of fanaticism. A true hacker is not a group person. He’s a person who loves to stay up all night, he and the machine in a love-hate relationship… They’re kids who tended to be brilliant but not very interested in conventional goals. And computing is just a fabulous place for that, because it’s a place where you don’t have to be a Ph.D. or anything else. It’s a place where you can still be an artisan. People are willing to pay you if you’re any good at all, and you have plenty of time for screwing around.”

The hackers are the technicians of this science — “It’s a term of derision and also the ultimate compliment.” They are the ones who translate human demands into code that the machines can understand and act on. They are legion. Fanatics with a potent new toy. A mobile new-found elite, with its own apparat, language and character, its own legends and humor. Those magnificent men with their flying machines, scouting a leading edge of technology which has an odd softness to it; outlaw country, where rules are not decree or routine so much as the starker demands of what’s possible.

A young science travels where the young take it. The wiser computer research directors have learned that not trusting their young programmers with major responsibility can lead immediately to no research. AI is one of perhaps several dozen computer research centers that are flourishing with their young, some of them with no more formal education than they got at the local Free School. I’m talking to Les Earnest, the gent who went for beer. He’s tall, swarthy, has a black and white striped beard, looks like a Sufi athlete. He’s telling me about what else people build here besides refinements of Spacewar. There’s a speech recognition project. There’s the hand-eye project, in which the computer is learning to see and visually correct its robot functions. There’s work on symbolic computation and grammatical inference. Work with autistic children, ‘trying to get them to relate to computers first, and then later to people. This seems to be successful in part because many of these children think of themselves as machines. You can encourage them to interact in a game with the machine.”

I’ve written about that piece before, but it came up again because Rolling Stone itself is revisiting it:

Brand had something of a knack for staging epochal cultural happenings. In 1966, he co-produced the infamous Trips Festival with Ken Kesey. Thousands of hippies attended this three day event in San Francisco to listen to psychedelic rock and drink punch dosed with LSD.

[...]

Brand was equally tuned in to the technological revolution that was rocking the Bay Area at the time. “I discovered that drugs were less interesting than computers as a way to expand your consciousness,” he says.

In 1968, Brand helped the inventor Doug Engelbart orchestrate a presentation at a computer conference that has come to be known as The Mother of All Demos. Engelbart demonstrated video conferencing, the computer mouse, email, hypertext, word processing and a windows-based organizational structure. Basically, he predicted most of the elements of the modern personal computer and the modern workplace.

Ecco the Dolphin

Saturday, May 21st, 2016

I never played Sega’s Ecco the Dolphin, but I’m not surprised that it would be linked to John C Lilly:

Lilly was once a renowned and respected American scientist, with a particular interest in marine biology and interspecies communication. In the early 1960s he was given funding by NASA to research whether it was possible to teach dolphins to speak. NASA’s logic was that if we could learn to communicate with dolphins, we would have a better understanding of how to converse with extra-terrestrials if they were to ever pop down for a visit.

Lilly flooded a house in the Caribbean so that dolphins could live as closely as possible with him and his team, amongst them Margaret Howe Lovatt, who apparently had sex with one of the animals. The experiment fizzled out as, unsurprisingly, nobody was able to get any of them to talk – although check out YouTube for one of his subjects attempting a pretty close “Hello Margaret”. Useful, if all aliens were called Margaret. Lilly lost funding for the project, moved away from traditional science and threw himself further and further into 1960s pseudo-mysticism and chemical experimentation.

Around 1971 Lilly was looking for a cure for his chronic migraines, and a friend suggested that ketamine could help get rid of them. Back then ketamine wasn’t a widely used drug, probably only used recreationally by a small group of dedicated trippers, quite unlike its status today as a popular party drug. When he was under the influence of a small dose of K, Lilly said that he felt the migraine being pushed out of his body and, miraculously, he never had one again. Encouraged by this, he developed a longstanding affection for the substance he dubbed “Vitamin K”, and started taking it regularly, gradually injecting it in higher doses.

Just shooting up ketamine on its own wasn’t enough for Lilly, though, and soon he was IV-ing it inside a sensory deprivation tank with the help of his friend, Dr Craig Enright. They thought that by using the tank external stimulation would be significantly reduced, giving a psychedelic or, in this case, a dissociative experience at a higher level of intensity. Neither appreciated that what they were doing was incredibly $#@!ing dangerous – tranquilising drugs and floating on water aren’t to be mixed under most circumstances, and sure enough Lilly’s wife, Antonietta, had to resuscitate him on one occasion where he nearly drowned. These experiments would form the foundation for Paddy Chayefsky’s 1978 novel Altered States, later adapted into a movie by director Ken Russell.

During his sessions, Lilly came to believe that he was being contacted by an organic extra-terrestrial entity called the Earth Coincidence Control Office – ECCO. This alien group was benevolent, omniscient and in control of all earthly matters. Except for when they weren’t quite so friendly, as at one point Lilly thought they’d made off with his penis.

The similarities between Sega’s Ecco the Dolphin and Lilly’s ketamine fantasies are undeniable. It’s almost like the game’s story is an amalgamation of his interest in dolphins and the wacky philosophy he spouted when returning to reality from his phenomenal K-hole trips.

Alongside ECCO, Lilly encountered another alien life force, which he called the Solid State Intelligence. Unlike the entities from ECCO, the SSI were spawned by a mechanical solar system, and their main aim was to ravage the earth and destroy mankind. It’s not unlike the much-documented cinematic battles between us fleshy creatures and advanced AI turned malevolent, and it’s no stretch to compare the SSI with Ecco’s Vortex enemies, those evil, dolphin-kidnapping, interstellar villains.

(Hat tip to Scott Alexander.)

Can Boys Beat Girls in Reading?

Saturday, May 7th, 2016

Boys outscored girls on reading tests — when they were told the tests were a game:

The latest study, in France, involved 80 children, 48 boys and 30 girls age 9 years old on average, from four third-grade classes at three schools. All classes received a silent reading test that required students to underline as many animal names as possible in three minutes from a list of 486 words (animal names comprised half the list). Two classes were told the test was an evaluation of their reading abilities, and two were told it was a new animal fishing game designed for a fun magazine.

In classes given reading evaluations, boys made an average of 33.3 correct answers compared with 43.3 by the girls. But when the tests were framed as animal games, boys’ average scores were significantly higher: 44.7 compared with 38.3 for the girls.

It looks like the boys’ performance improved and the girls’ performance declined when they said it was a game?

Wargaming in the Classroom: An Odyssey

Monday, April 25th, 2016

James Lacey designed the perfect course on the Peloponnesian War for his students at the Marine Corps War College — only to realize that no one learned or remembered much of anything. Then he completely redesigned the course to use wargaming in the classroom, and the results were amazing:

I selected Fran Diaz’s Polis: Fight for the Hegemony, because, unlike many games, it has a heavy economic and diplomatic element. After dividing the seminars into teams, I was able to run five simultaneous games.

The results were amazing.

As every team plotted their strategic “ends,” students soon realized that neither side had the resources — “means” — to do everything they wanted. Strategic decisions quickly became a matter of tradeoffs, as the competitors struggled to find the “ways” to secure sufficient “means” to achieve their objectives (“ends”). For the first time, students were able to examine the strategic options of the Peloponnesian War within the strictures that limited the actual participants in that struggle.

Remarkably, four of the five Athenian teams actually attacked Syracuse on Sicily’s east coast! As they were all aware that such a course had led to an Athenian disaster 2,500 years before, I queried them about their decision. Their replies were the same: Each had noted that the Persians were stirring, which meant there was a growing threat to Athens’ supply of wheat from the Black Sea. As there was an abundance of wheat near Syracuse, each Athenian team decided to secure it as a second food source (and simultaneously deny it to Sparta and its allies) in the event the wheat from the Black Sea was lost to them. Along the way, two of the teams secured Pylos so as to raise helot revolts that would damage the Spartan breadbasket. Two of the teams also ended revolts in Corcyra, which secured that island’s fleet for Athenian purposes, and had the practical effect of blockading Corinth. So, it turns out there were a number of good strategic reasons for Athens to attack Syracuse. Who knew? Certainly not any War College graduate over the past few decades.

All of these courses of action were thoroughly discussed by each team, as were Spartan counter moves. For the first time in my six years at the Marine Corps War College, I was convinced that the students actually understood the range of strategies and options Thucydides wrote about. In the following days, I was stopped dozens of times by students who wanted discuss other options they might have employed, and, even better, to compare their decisions to what actually happened. A number of students told me they were still thinking about various options and decisions weeks later. I assure you that no one even spent even a car ride home thinking about my Thucydides lectures.

[...]

For six or more hours at a sitting, classes remain focused on the strategic choices before them, as they try to best an enemy as quick-thinking and adaptive as they are. Every turn presents strategic options and dilemmas that have to be rapidly discussed and decided on. As there are never enough resources, time and again hard choices have to be made. Every war college administrator can wax eloquently about their school’s mission to enhance their students’ critical thinking skills. But they then subject those same students to a year of mind-numbing classroom seminars that rarely, if ever, allow them to practice those skills that each college claims as its raison d’etre. Well, wargaming, in addition to helping students comprehend the subject material, also allows them an unparalleled opportunity to repeatedly practice decisive critical thinking. Moreover, it does so in a way where the effects of both good and bad decisions are almost immediately apparent.

At the end of each wargame, students walked away with a new appreciation of the historical circumstances of the period and the events they had read about and discussed in class. And even though all wargames are an abstract of actual events, I am sure that no student exposed to historical gaming will ever again read about the Peloponnesian War without thinking about Sicily’s wheat, the crucial importance of holding the Isthmus of Corinth, or what could have been done with a bit more Persian silver in the coffers of one side or the other’s treasury.

Revitalizing Wargaming

Sunday, April 17th, 2016

We must revitalize wargaming to prepare for future wars, Deputy Secretary of Defense Bob Work and Gen. Paul Selva argue:

Few historical periods match the dynamic technological disruption of the inter-war years of the 1920s and 1930s. During these decades, militaries the world over struggled to adapt to new inventions such as radar and sonar, as well as rapid improvements in wireless communications, mechanization, aviation, aircraft carriers, submarines, and a host of other militarily relevant technologies. Military planners and theorists intuitively understood that all these new technologies, systems, and advances would drive new ways of fighting, but they were forced to envision what future battlefields would look like with few clues to go by.

To help navigate through this period of disruptive change, the United States military made extensive use of analytical wargaming. Wargames were an inexpensive tool during a period of suppressed defense spending to help planners cope with the high degree of contemporary technological and operational uncertainty. They were used to explore a range of possible warfighting futures, generate innovative ideas, and consider how to integrate new technologies into doctrine, operations, and force structure. For example, faculty and students at the Naval War College integrated wargaming into their entire course of study, analyzing the then-novel concept of carrier task force operations, the role of submarines in scouting and raiding, and how to provide logistics support to fleet operations spread over the vast Pacific Ocean. Wargames in classrooms at Quantico helped the Marine Corps develop new concepts for amphibious warfare and conceive of new techniques for capturing advanced naval bases. Wargamers at the Army War College explored how to employ tanks and artillery on infantry-dominated battlefields and examined the logistical challenges of fighting a war far from American shores.

As valuable as they were, wargames were not in and of themselves sufficient to prompt organizational and operational change. As such, all of the services worked hard to test wargame results in fleet and field exercises. Exercises were used to verify game insights using systems at hand or with surrogates that represented desired advanced capabilities identified during game play. The observations and lessons learned in exercises were in turn fed back into new wargames, thus creating a cycle of creative ideas and innovation that generated requirements for new systems, suggested new operation concepts, and influenced force design.

Once the Second World War began, those warfighting communities that had pursued wargaming and exercises with vigor proved far better prepared for modern combat than those that did not. For example, of the three major warfighting communities in the Navy — naval aviation, surface warfare, and submarine — the naval aviation community carried out the most innovative pre-war experimentation and exercises. Although aircraft carriers were originally envisioned as operating in support of the fleet battle line, carrier aviators explored a wide range of futures, including independent carrier operations. As a result, the U.S. carrier force was ready on day one of the Pacific war, and within six months had inflicted a major, lasting defeat on the superior Japanese carrier force at the Battle of Midway. By contrast, pre-war wargames and exercises in the submarine community had emphasized rote doctrine using the submarine fleet as a scouting force for the main battle line, and policy strictures dampened any exploration of independent submarine operations. Unsurprisingly, then, the submarine community proved unprepared for the tactics, techniques, and procedures needed to execute unrestricted warfare on Japanese merchant shipping. Similarly, surface warfare wargames failed to anticipate long-range torpedoes or account for the Japanese emphasis on night surface action. As a consequence, they suffered badly in early clashes against a highly trained Japanese cruiser and destroyer force that excelled at night fighting and was armed with the deadly, long-range Long Lance torpedo.

Today, we are living in a time of rapid technological change and constrained defense spending, not unlike that of the inter-war years.

The story of how the Navy learned to learn to fight is fascinating, by the way.

Blood, Dice, and Darkness

Monday, April 4th, 2016

Games Workshop started humbly, as three game-loving friends selling Go boards and backgammon sets from their London flat. Then their gaming fanzine Owl and Weasel founds its way across the Atlantic to Gary Gygax, who sent them a copy of Dungeons & Dragons. They became the European distributor — still from their London flat:

“We were desperate not to let Gygax know that we were running the company from our flat,” Livingstone said.

“But what we didn’t know at the time was that he was publishing Dungeons & Dragons from his flat as well. Both parties were assuming that the other was some big-time operation, but it was very much a fledgling industry at the time.”

Eventually their side-project, a miniature-making company called Citadel, became the main project, as the games they devised to sell those miniatures took off:

“Bryan told us: ‘We need a game to sell more toy soldiers, get on with it.’ He’s like that,” Priestley said.

“His role was very top-down. He only laid down a couple of paramaters. He told us the game had to have rules for every model the company made at the time, and that it could only use standard six-sided dice, because every kid had them in their Monopoly sets.

“Richard wrote the initial manuscript, and then I did a lot of the production and development work, so a lot of the game mechanics were down to Richard and I did a lot of the refinement and detail.”

Warhammer First Edition

The result of their efforts was Warhammer’s first edition. Published in 1983 as a set of nondescript black and white booklets, it included rules for manoeuvring and fighting with a variety of fantastical creatures and soldiers. But it came with none of the fictional background that’s now synonymous with the Warhammer brand. Instead it was marketed as a “dual system” allowing roleplay gaming groups to fight large battles as part of their ongoing campaigns.

After Warhammer, they created its science-fiction cousin, Warhammer 40,000 — which has a deeper origin story than I realized:

“We just plundered everything. Obviously Tolkien was a big influence, and in terms of 40K there’s a lot of Frank Herbert’s Dune in there. If you’ve read Dune, every chapter starts with a bit of an excerpt, and I rather enjoyed that, so I just copied the idea by putting little bits of pseudo fiction in.”

Other influences included the works of Robert Heinlein and H.P. Lovecraft, but it was a much older source – the 17th century poet John Milton – who would provide the inspiration for the game’s greatest conflict.

In a reimagining of the epic poem Paradise Lost, which deals with an attempt to overthrow God by a faction of rebel angels, Warhammer 40,000 featured a cataclysmic schism within the forces of the Empire of Mankind. In an event known as the Horus Heresy, chapters of Space Marines – genetically engineered, fanatically religious super-soldiers – turned against their Emperor after falling prey to the influence of the Chaos Gods, the supreme antagonists of this dark future setting.

“The original idea for Chaos was Bryan Ansell’s,” Priestley said.

“He wrote a Warhammer supplement called Realms of Chaos where he came up with the gods and the demons. He produced this huge hand-written manuscript where he defined all of that, and I took what he’d written and developed it as a book.”

But Priestley’s idea of Chaos differed from Ansell’s, and in 40K he sought to expand on the concept.

“Bryan’s idea of Chaos was very much derived from [science fiction and fantasy author] Michael Moorcock,” he said. “I always thought it was a little too close for comfort, it looked like we were just copying.

“But I’d always had this sense of Chaos existing as described in Paradise Lost. I’d tried to bring elements of that into the background and gradually change it from a description of demons into a kind of force out of which came realities, a kind of literal primal chaos.

“Unless you’ve read Paradise Lost you don’t get it. The whole Horus Heresy is just a parody of the fall of Lucifer as described by Milton.”

Is Shooting a Martial Art?

Wednesday, March 16th, 2016

“Do you consider shooting a martial art?” I asked Jocko (@jockowillink). “How has your firearms training been like or unlike martial arts training?” He answered, and it went something like this:

Shooting is absolutely a martial art, maybe not the way people picture martial arts nowadays, because we picture a guy in a gi doing karate, that’s the generic picture. But for me that’s not martial arts actually. For me martial art is the art of war, the individual warrior skills which it takes, and firearms are absolutely a martial art. Because it’s something that you train, something that you get good at, something you need to maintain your skill at. To me it’s another piece of the puzzle, another thing that you need to know how to do, just like tactics that go along with shooting are an important part of being a warrior, you need to know how to shoot.

The training is very similar in my mind to martial arts training, in that it takes repetition, you have to know what the basics are, you have to repeat those basics, then you get more advanced.

It’s about movement and getting efficient with your movement, you want to train very similar to the way you train mixed martial arts. And once you get all those mechanical skills down, then you want to train your mind around this skill, so that your mind knows how to utilize it when things are unexpected and when there’s chaos and mayhem going on.

(Thanks to our Slovenian guest for transcribing that.)

From there, Jocko tells a fun story about Nerf “lazer” tag and the importance of good tactical training.

League of Legends Prodigy

Wednesday, February 3rd, 2016

ESPN looks at League of Legends prodigy Faker — a young pseudonymous Korean eSports “athlete” who has mastered the popular PC game:

When I ask the group why Faker is regarded as the best player in the world, MonteCristo, who goes by Monte, jumps in: “How would you grade a professional athlete? Like, what makes LeBron great?”

I rattle off a few words: athleticism, skill, decision-making.

“It’s the same. It’s exactly the same,” Susie says.

The League equivalent of athleticism is called mechanics, which refers to a player’s ability to use his mouse and keyboard to make swift movements, like dodging shots. In this respect, Monte says, Faker is peerless. He points me to a video of what is widely seen as the greatest play in League history, clipped from a 2013 game between SK Telecom and the KT Bullets. Faker is dueling another player, Ryu, and they’re both playing the same champion, a ninja named Zed. After a brief skirmish, Faker’s Zed appears about to die, so he darts away. Then, just when Ryu thinks he has the fight sewn up, Faker unleashes a startling set of moves, cutting down his opponent in a blinding flash. The audience goes nuts. “He used six different abilities in the span of two seconds,” Monte says.

Even more impressive, DoA adds, is the breadth of Faker’s champion pool, which makes it easier for him to adapt to new patches to the game — the “meta,” in eSports parlance. Because Riot upgrades League every few weeks, players live in perpetual fear of having their favorite champions’ skills diminished. Imagine if the NFL suddenly announced next year that rushing touchdowns were worth only five points, or if MLB expanded the strike zone for left-handed pitchers. Although the constantly changing meta keeps the game fresh, it can be agonizing for professionals. Some players never recover from an ill-timed patch.

That’s one of the reasons the average eSports career is so short. Professional players typically retire before their mid-20s; like figure skaters, they peak long before then. Older gamers must battle slowing reflexes and fatigue, as well as injuries to their necks and wrists. “As a male teenager, it’s easy to play video games for 16 hours,” Monte says.

Because many Korean players skip college, their career options after retiring are limited. “A lot of pro gamers don’t come from wealthy backgrounds,” Susie says. “A good number of them are doing this because they’re supporting their families.” Increasingly, she says, players realize they have limited time to capitalize on their skills, which is driving some of them to leave the country. Although most professional gamers in Korea earn five-digit salaries and a few elite players make over $100,000 (Monte says Faker probably makes more than twice that; SK Telecom declined to comment on his salary), Chinese teams boast massive war chests. One squad, Invictus Gaming, is owned by the son of Wang Jianlin, the richest man in mainland China. This winter, Invictus added four Korean players to its roster.

Pro players also make money by streaming, allowing fans to watch them practice while advertisements pop up. One retired player in China, Wei “Caomei” Han-Dong, has said he makes more than $800,000 a year streaming. Korean teams have begun to stream a little, but in general, “they think it’s inefficient,” says Lee “CloudTemplar” Hyun-woo, a retired-gamer-turned-caster. “In Korea, to make money you have to put up results.” Demand is out there, though. This February, a minor scandal flared up when a Twitch user started streaming Faker’s practice games without permission.

AlphaGo

Friday, January 29th, 2016

Researchers at DeepMind staged a machine-versus-man Go contest in October, at the company’s offices in London:

The DeepMind system, dubbed AlphaGo, matched its artificial wits against Fan Hui, Europe’s reigning Go champion, and the AI system went undefeated in five games witnessed by an editor from the journal Nature and an arbiter representing the British Go Federation. “It was one of the most exciting moments in my career, both as a researcher and as an editor,” the Nature editor, Dr. Tanguy Chouard, said during a conference call with reporters on Tuesday.

This morning, Nature published a paper describing DeepMind’s system, which makes clever use of, among other techniques, an increasingly important AI technology called deep learning. Using a vast collection of Go moves from expert players — about 30 million moves in total — DeepMind researchers trained their system to play Go on its own. But this was merely a first step. In theory, such training only produces a system as good as the best humans. To beat the best, the researchers then matched their system against itself. This allowed them to generate a new collection of moves they could then use to train a new AI player that could top a grandmaster.

“The most significant aspect of all this…is that AlphaGo isn’t just an expert system, built with handcrafted rules,” says Demis Hassabis, who oversees DeepMind. “Instead, it uses general machine-learning techniques how to win at Go.”

[...]

“Go is implicit. It’s all pattern matching,” says Hassabis. “But that’s what deep learning does very well.”

[...]

At DeepMind and Edinburgh and Facebook, researchers hoped neural networks could master Go by “looking” at board positions, much like a human plays. As Facebook showed in a recent research paper, the technique works quite well. By pairing deep learning and the Monte Carlo Tree method, Facebook beat some human players — though not Crazystone and other top creations.

But DeepMind pushes this idea much further. After training on 30 million human moves, a DeepMind neural net could predict the next human move about 57 percent of the time — an impressive number (the previous record was 44 percent). Then Hassabis and team matched this neural net against slightly different versions of itself through what’s called reinforcement learning. Essentially, as the neural nets play each other, the system tracks which move brings the most reward — the most territory on the board. Over time, it gets better and better at recognizing which moves will work and which won’t.

“AlphaGo learned to discover new strategies for itself, by playing millions of games between its neural networks, against themselves, and gradually improving,” says DeepMind researcher David Silver.

According to Silver, this allowed AlphaGo to top other Go-playing AI systems, including Crazystone. Then the researchers fed the results into a second neural network. Grabbing the moves suggested by the first, it uses many of the same techniques to look ahead to the result of each move. This is similar to what older systems like Deep Blue would do with chess, except that the system is learning as it goes along, as it analyzes more data — not exploring every possible outcome through brute force. In this way, AlphaGo learned to beat not only existing AI programs but a top human as well.

Can Videogames Make You a Better Race-Car Driver?

Sunday, November 15th, 2015

A few years ago Top Gear put an iRacing champion in a real race car and found that he was virtually prepared — but not at all physically prepared.

The Wall Street Journal now reports that videogames can make you a better race-car driver:

The first time that Brendon Blake, a 41-year-old physical therapist from Flowery Branch, Ga., careened around the nearby Road Atlanta racetrack, his instructor was taken aback. Mr. Blake, despite being a total beginner, was fast. That’s because, long before he’d enrolled in the one-day racing class, he’d “driven” the same course hundreds of times. It didn’t matter that he had done so virtually, in the Xbox car-racing game “Forza Motorsport.”

“The instructor sitting in the passenger seat said he was surprised I knew where to place the car on the track,” said Mr. Blake of that maiden drive, about four years ago. “I recognized every single corner and knew where the race line was and where all the apexes were — all from the game.” Since then, Mr. Blake — who plays using a force-feedback steering wheel and mock pedals like those shown below — has taken his 291-horsepower Mitsubishi Lancer Evolution to courses all over the country, from Tennessee’s Nashville Speedway to Talladega in Alabama. He pays about $250 a day for further driving instruction on “track days,” when average Joes can rent time on a course that’s not being used for a race.

The Tangled Cultural Roots of Dungeons & Dragons

Wednesday, November 4th, 2015

Jon Michaud laments how Michael Witwer’s Empire of Imagination fails to untangle the roots of Dungeons & Dragons:

We get just a paragraph on Gygax’s unhappy stint in the Marines, for instance; the idea that boot camp might have had some bearing on Gygax’s lifelong effort to re-create combat conditions in tabletop games never seems to cross Witwer’s mind. Likewise, there are only passing mentions of Gygax’s years of work as an insurance underwriter. But one needs only to browse the Advanced D. & D. “Player’s Handbook” or the “Dungeon Master’s Guide” to see how similar the books’ numerous charts are to actuarial tables.

They really do resemble actuarial tables, don’t they?

How a Video Game Helped People Make Better Decisions

Tuesday, October 20th, 2015

Carey K. Morewedge and his colleagues developed a couple “serious” computer games to help people make better decisions:

Participants who played one of our games, each of which took about 60 minutes to complete, showed a large immediate reduction in their commission of the biases (by more than 31%), and showed a large reduction (by more than 23%) at least two months later.

The games target six well-known cognitive biases. Though these biases were chosen for their relevance to intelligence analysis, they affect all kinds of decisions made by professionals in business, policy, medicine, and education as well. They include:

  • Bias blind spot — seeing yourself as less susceptible to biases than other people
  • Confirmation bias — collecting and evaluating evidence that confirms the theory you are testing
  • Fundamental attribution error — unduly attributing someone’s behavior to enduring aspects of that person’s disposition rather than to the circumstance in which the person was placed
  • Anchoring — relying too heavily on the first piece of information considered when making a judgment
  • Projection — assuming that other people think the same way we do
  • Representativeness — relying on some simple and often misleading rules when estimating the probability of uncertain events

We ran two experiments. In the first experiment, involving 243 adult participants, one group watched a 30-minute video, “Unbiasing Your Biases,” commissioned by the program sponsor, the Intelligence Advanced Research Projects Activity (IARPA), a U.S. research agency under the Director of National Intelligence. The video first defined heuristics — information-processing shortcuts that produce fast and efficient, though not necessarily accurate, decisions. The video then explained how heuristics can sometimes lead to incorrect inferences. Then, bias blind spot, confirmation bias, and fundamental attribution error were described and strategies to mitigate them were presented.

Another group played a computer game, “Missing: The Pursuit of Terry Hughes,” designed by our research team to elicit and mitigate the same three cognitive biases. Game players make decisions and judgments throughout the game as they search for Terry Hughes — their missing neighbor. At the end of each level of the game, participants received personalized feedback about how biased they were during game play. They were given a chance to practice and they were taught strategies to reduce their propensity to commit each of the biases.

We measured how much each participant committed the three biases before and after the game or the video. In the first experiment, both the game and the video were effective, but the game was more effective than the video. Playing the game reduced the three biases by about 46% immediately and 35% over the long term. Watching the video reduced the three biases by about 19% immediately and 20% over the long term.

In a second experiment, involving 238 adult participants, one group watched the video “Unbiasing Your Biases 2” to address anchoring, projection, and representativeness. Another group played the computer detective game “Missing: The Final Secret,” in which they were to exonerate their employer of a criminal charge and uncover criminal activity of her accusers. Along the way, players made decisions that tested their propensity to commit anchoring, projection, and representativeness. After each level of the game, their commission of those biases was measured and players were provided with personalized feedback, practice, and mitigation strategies.

Again, the game was more effective than the video. Playing the game reduced the three biases by about 32% immediately and 24% over the long term. Watching the video reduced the three biases by about 25% immediately and 19% over the long term.

The games, which were specifically designed to debias intelligence analysts, are being deployed in training academies in the U.S. intelligence services. But because this approach affects the decision maker rather than specific decisions, such games can be effective in many contexts and decisions — and with lasting effect. (A commercial version of the games is in production.)

Confessions of an Anonymous Free-to-Play Producer

Thursday, September 24th, 2015

We own you, an anonymous free-to-play game producer explains:

One of our engineers came up with a rather simple solution that today would seem like a joke. We could have a JSON file online that contained all the level information. Then we could update the file to make a level easier (or harder). This way we could watch user reactions (mostly app store reviews, Twitter was still pretty basic at this time). This worked great, we were able to balance the game in the wild.

During a meeting about the game, the guy who ran our website brought up some interesting information. He started watching the web logs and seeing all the connections to the JSON file. Unbeknownst to him (or our team) he was getting us a DAU. For the engineering and production teams, this was just a neat thing to know, a feel good “look how many people love our game” statistic. The CEO saw something else. Pretty quickly we started getting more requests for what our users were doing. Upper management was disappointed by our lack of answers. I found a new service online called Pinch Media, they were an analytics tracker. I got the team to integrate Pinch into a few products and finally I had answers. Of course then more answers were asked. Around this time, free to play started happening. Suddenly, Marketing and our bosses demanded to know more than ever. In response to the pressure to explain our user base, I ended up building an event matrix. I had no schooling in this, so I was just making it up as I went along. My first matrix was awesome for a game developer. It was full of all those cool stats like “How far has the player run” or “How many bullets has he shot”. But this did not impress my bosses. They wanted to know how we could get the player to buy more stuff, tell his friends to play the game (and thus I learned about cohorts, all I wanted to do was make games).

Time passed, Free to Play became a thing. I went from company to company. Each time, every new project became less and less about how we can do cool things, and more about how we can track and target users to get the most whales possible, boost chart position and retain users to shove as many ads on them as possible.

All of this already seems bad. But along the way, a major thing happened. Facebook. I forget when I did my first Facebook required app, but it was a game changer. Facebook has changed how it has worked over the years. Today you can’t quite get as much information (easily) as you could with the first API, but you still get a lot. We collect as much information about a player as possible, thanks to Facebook we have a ton. Even users who don’t really use Facebook or fill it with “fake” data actually tell us a lot. You might not use Facebook, but your connections give you away. If you play with friends, or you have a significant other who plays, we can see the same IP address, and learn who you are playing with. When we don’t know information, we try to gather it in a game. Have you played a game with different country flags? We use those to not only appeal to your nationalistic pride, but to figure out where you are (or where you identify). Your IP address says you are in America, but you buy virtual items featuring the flag of another country, we can start to figure out if you are on vacation, or immigrated. Perhaps English is not your first language. We use all of this to send you personalized Push Notifications, and show you store specials and items we think you will want.

And if you are a whale, we take Facebook stalking to a whole new level. You spend enough money, we will friend you. Not officially, but with a fake account. Maybe it’s a hot girl who shows too much cleavage? That’s us. We learned as much before friending you, but once you let us in, we have the keys to the kingdom. We will use everything to figure out how to sell to you. I remember we had a whale in one game that loved American Football despite living in Saudi Arabia. We built several custom virtual items in both his favorite team colors and their opponents, just to sell to this one guy. You better believe he bought them. And these are just vanity items. We will flat out adjust a game to make it behave just like it did last time the person bought IAP. Was a level too hard? Well now they are all that same difficulty.

Scrabble Francophone

Monday, July 27th, 2015

The French-language Scrabble world championship just went to a New Zealander — who doesn’t speak French:

The BBC reported that Nigel Richards, originally from Christchurch, defeated a rival from French-speaking Gabon in the final in Louvain, Belgium, on Monday.

He had only started studying the French dictionary about eight weeks ago, said a close friend of Mr Richards, Liz Fagerlund.

“He doesn’t speak French at all, he just learnt the words. He won’t know what they mean, wouldn’t be able to carry out a conversation in French I wouldn’t think.”

Mr Richards, now in his late forties, is a previous English Scrabble champion. He is based in Malaysia.

He has won five US National titles and the World Scrabble Championship three times.

Why The Next Sports Empire Will Be Built On eSports

Tuesday, June 30th, 2015

So-called eSports tournaments are reaching audiences of tens of millions:

Last year’s League of Legends championship, for example, drew nearly 30 million viewers, putting it in line with the combined viewership of the 2014 MLB and NBA finals, or the series finales of Breaking Bad and Two and a Half Men, plus the Season 4 finale of Game of Thrones. As with most sports, competitive gaming is now firmly entrenched in the US college system. The country’s largest collegiate league counts more than 10,000 active players, some of whom are on full athletic scholarships. Eager to capitalize on growing interest in the sport, Major League Gaming (MLG) opened the first dedicated domestic eSports arena in October 2014, and major brands such as Ford, American Express and Coke have begun forming partnerships with game developers, teams, players, event organizers and video distributors. What’s more, the US Department of State has been issuing athlete visas to competitive gamers since 2013.

[...]

In March 2015, Twitch averaged more than 600,000 simultaneous viewers, reached an audience of 51M worldwide and delivered more than 26B minutes of video entertainment. On a domestic basis, 11B minutes were watched in March – representing roughly 14 hours for each of the 13M American viewers. This consumption is so great that Twitch is already larger than 70% of American television networks, as well as Amazon’s own OTT video service.

However, the value of this consumption isn’t just its magnitude. An estimated 70% of all viewers are under the age of 35, making Twitch’s audience both highly valuable to advertisers and hard to reach via traditional television. Moreover, eSports fans, unlike linear TV viewers, are highly engaged in the content. Major League Gaming, for instance, consistently beats the industry average on key digital ad metrics such as completion rates (90% vs. 72%), click-through rates (4% vs. 2%), and ad viewability (99% vs. 44%). What’s more, Twitch shows little sign of slowing down. Total minutes delivered (both domestic and abroad) have grown by an average of 7% each month for the past three years, while per viewer consumption has doubled over that same period.

[...]

Despite ever-growing consumer interest and potential, eSports are still far from becoming an industry. In 2014, eSports generated less than $200M in revenue worldwide, including sponsorship, advertising, licensing, ticket sales and game-publisher investment according to Newzoo. By comparison, the US-only NFL and MLB gross roughly $10 billion a year each, while the European professional soccer/football leagues generate close to $21 billion.

Even as eSports tournaments have proliferated and audiences have expanded into the millions, the value of these tournaments continues to languish. The average event offers only $18,000 in total prize winnings (a figure almost unchanged from 1998) and 2014’s 1,990 tournaments handed out a relatively unimpressive $35M collectively. The largest prize pool did surge in 2014, from $3M to $11M, but only five players made more than $1M during the year. The remaining 6,200 e-athletes took home an average of $7,000.