Skip to Content

The Digital Pit Boss

The Mohegan Sun casino goes “all in” on a sensor-riddled blackjack table.
August 1, 2005

The case: Connecticut’s Mohegan Sun casino is preparing to go “all in” on a sensor-riddled blackjack table that will give the house perfect knowledge of how its customers play. It found that a relatively low-tech system of cameras is more sensible than RFID – and that customers will tolerate electronic surveillance if they believe it offers them benefits.

For a casino, the expenses and profits at blackjack tables are difficult to precisely pin down. What’s clear is that the aggregate numbers are staggeringly high: a typical blackjack table in Reno, NV, can see more than $6 million wagered monthly, with about 1 percent of that sum staying behind as the casino’s winnings. But gauging the performances of individual players has long been an inexact science. Clay gaming chips slide back and forth between human hands. Paper playing cards are dealt, collected, and shuffled. Players signal their desire for additional cards by tapping their fingers on the table and refuse hits by waving their hands. Some expend all their chips; others walk away with more than they brought.

Casinos hunger for a better understanding of players. In particular, they want information that will help them refine how often, and to whom, they dole out “comps” – a sort of casino currency redeemable for treats like free hotel rooms, dinners, and drinks. This calculation requires two primary pieces of information: how much a given player is wagering, and – for blackjack and some other card games – how skilled that player is.

Knowing how much players are wagering requires watching their chips closely, and judging their skill levels accurately requires observing each decision they make. Right now, the onus for keeping track of these things falls to a manager known as a pit boss, who is, famously, backed up by surveillance staff eyeballing video monitors in a back room. Like other casinos, Mohegan Sun, in Uncasville, CT, thinks technology can help it track blackjack

players. “We have long been looking for a technology that would help us provide automated ratings of players at gaming tables,” says Dan Garrow, the chief information officer at Mohegan Sun. “If you spend $10,000, we will do something for you to keep coming back. It’s no different than any other business – how do you keep your customers coming back?”

But of course, gambling is different from any other business. While a casino does, as Garrow says, care about customer retention as much as any company, its relationship with its customers is adversarial: a casino wants its customers to lose. “Each player represents what we call a &lsquotheoretical win,’” says Garrow. “You would call that a loss.” That is what makes comping so important: it is the method by which casinos try to soften the edges of the hard reality of loss. And how comping is done matters greatly: the trick is to lavish the biggest gifts on the people who are most likely to not only place big bets but also make decisions that worsen their odds.

Casinos know that technology can help them identify those people. Garrow explored – but has rejected for now – prototype systems that use radio frequency identification (RFID) tags embedded in gaming chips. This technology gives each chip a unique identifying code; as a player buys chips (after first showing a player ID card to become eligible for comps), the chips are electronically associated with that player. At blackjack or other tables, a tag “reader” identifies each movement of each chip, registering how much has been bet, won, and lost. As a side benefit, such chips are nearly impossible for employees to steal or players to counterfeit.

While a few casinos are testing prototype RFID-chip systems, none has yet implemented them, says John Kendall, president of one RFID-gaming chip maker, Chipco International of Raymond, ME. When Garrow investigated RFID chips, he concluded they were too costly, though he acknowledges that prices have since come down. (The newest versions add about 50 cents to the 80-cent price of a traditional casino chip, Kendall says.) Moreover, while RFID technology provides detailed information about players’ betting patterns, it reveals nothing about the cards they base their bets on, and therefore nothing about their skill at blackjack.

While searching for alternatives, Garrow courted lone inventors proposing technology for blackjack tables. At one point he and his staff found themselves in the 13th-floor Manhattan apartment of an inventor who had rigged a blackjack table with computers and sensors to track all aspects of play. While the technology showed promise, the vendor was essentially looking for Mohegan Sun to provide his venture capital and expand his business – something Garrow was unwilling to do.

Then came MindPlay. Garrow was aware that a couple of casinos in Nevada had been trying out a system from MindPlay, a small, Bellevue, WA-based company that has since been bought by Bally Gaming and Systems. MindPlay builds blackjack tables with small cameras tucked into a slightly raised dealer platform facing the players. The accompanying gaming chips bear simple line patterns on their edges; these are read by the system’s pattern-recognition software. Special playing cards also bear line patterns that identify them; the patterns are on the faces of the cards and are read by a camera pointing up through the table at the mouth of the “shoe” – the box containing cards.

Players swipe their casino cards at the start of play – so they can earn comps – and thereafter, every detail of their play is tracked. The system knows which cards they receive, what decisions they make, and the amount of each bet they make.

A back-end system continually evaluates and reevaluates their skill, hand by hand, hit by hit. (The theoretical best player, who plays strategically perfect blackjack, will have a mathematical disadvantage of .45 percent; each mistake he or she makes drives that number upward.) The software’s skill assessments are fed into the casino’s management software. In a final, tangible touch, the playing surface of the MindPlay blackjack table has the feel of a mouse pad.

Paving the way for Mohegan Sun was the El Dorado casino in Lake Tahoe, NV, which did its beta testing four years ago. Now the casino has implemented 16 MindPlay tables. And what kind of data is it getting? Rob Mouchou, El Dorado’s vice president of operations, made a few mouse clicks during a phone interview and reported that in a recent 30-day period, 5,795 skilled players who swiped in using player ID cards wagered $16.6 million at the 16 MindPlay tables.

A few mouse clicks later, he saw the payoff to the casino on these players. Before MindPlay, Mouchou comped players at a flat 25 percent of their estimated losses – a figure the house fixed at 1 percent of the amount they were estimated to have wagered. But this was always very much a ballpark figure, he says. Now he comps at 25 percent of the amount their skill evaluations suggest they will lose, on average, multiplied by the exact amount they wagered. The 5,795 players cited by Mouchou were particularly skilled, so their projected average loss was just .63 percent. Previously, Mouchou would have comped them $41,500 – one-quarter of 1 percent of $16.6 million. Instead, he comped $26,145, one-quarter of .63 percent of $16.6 million. Thus he saved nearly $15,000 in comps. Since this was spread out among 5,795 players, each player’s comp reduction was tolerable: less than $3. (And as a side benefit, he can track his dealers and see which ones keep the momentum going, and which ones are sluggish.)

Now Mouchou is planning a marketing campaign based on El Dorado’s new technology. Most casinos won’t expend their pit-boss manpower on low-stakes tables and thus don’t issue comps to the players who frequent them. So the MindPlay tables give El Dorado a marketing edge. “We want to be able to comp $5 players, $10 players, that other properties don’t ever track,” Mouchou says.

These advantages were not lost on Mohegan Sun’s Garrow. But he faced one final hurdle: the gamblers themselves, who – just like anyone else – can be suspicious of electronic surveillance. Richard LeBaron, a product manager at MindPlay, says the company’s technology offers advantages to players, too. “Like any new technology, it takes time to be accepted with open arms,” he says. “It’s all in training dealers in handling questions that come from patrons. Patrons have felt their comps are never tracked properly. The patrons of a casino now have a better understanding that with the system, they are going to get comped accurately and fairly.”

Today, Mohegan Sun just has two MindPlay tables, which it keeps in its dealer-training facility – a steel warehouse a short drive from the casino itself. But it will install 10 of the new tables in the casino next month. And Garrow is planning to cash in all of Mohegan Sun’s chips – literally – in favor of a new batch that works with the MindPlay tables. The new chips won’t be as expensive as RFID chips, but they will be made of extruded nylon, not clay. The nylon gives more sharply defined edge patterns, allowing the camera’s pattern-recognition software to correctly identify them.

Mohegan Sun hasn’t given up on RFID entirely. It’s considering giving its customers special RFID tags they can put on their cars and installing tag readers on the road to the casino. When the high rollers with bad blackjack skills hit town, Mohegan Sun will know it before they even reach the valet parking. “We could have services available, credit-limit changes, or set up a gaming table in a particular area, or have a favorite drink or food ready,” Garrow says. “We might be able to make your experience here at Mohegan Sun that much more special.” As Mohegan Sun and other casinos – and indeed other businesses – identify cost-saving surveillance technologies that both work on a practical level and are accepted by consumers, you can bet they’ll be installing them.

Keep Reading

Most Popular

This new data poisoning tool lets artists fight back against generative AI

The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models. 

Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist

An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.

The Biggest Questions: What is death?

New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.

Data analytics reveal real business value

Sophisticated analytics tools mine insights from data, optimizing operational processes across the enterprise.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.