Once upon a time there was side betting, next came totals betting, followed by half-time betting, after that the layers gave the punters quarter lines to shoot at, and, when we follow this arc all the way out to its logical end, we find interactive wagering. Since every single play changes the relative fortunes of two teams playing, say, a baseball game, why not set a new line after every single play and let the bettors have it? If your product is chances to gamble, why not offer your customers two hundred lines a game instead of the dozen or two that even the most prop-happy shops manage to put up on a featured contest?
WSEX (World Sports Exchange) was founded on this idea of offering a huge number of lines by re-calculating them releasing them anew after every possession, play, or out. The company also offers all the standard bets, and takes large limits on those, but interactive gaming is the flagship product. This new form of sports betting Jay and Steve invented is structured roughly after specialist-centered trading, and as it works in practice a great part of the theoretical hold is conceded in order to gain multiplies of volume. If you’re familiar with the world of on-line gaming then you know about the success of their company. By running an interactive wagering market that offers bettors enormous and consistent depth at vig prices very near to those one pays eleven-to-ten shops for straight bets, they remain the leaders in the corner of the gaming industry they pioneered.
If you already understand the basic scheme of this kind of wagering, then skip the following paragraph. If you’re unfamiliar with it, though, then here is a short introduction to how WSEX two-team interactives work (there’s also multi-way interactives, as in golf tournaments, a topic I hope to talk Mach into doing a guest co-column on).
Let’s say the Atlanta Braves are -150 favorites over the Philadelphia Phillies. In the terminology of WSEX interactives the "ask price" (price for which the player can buy the Braves chances of winning) would probably be $60. The "bid price" (price for which the player can sell Atlanta’s chances of winning, or what comes to the same thing, bet on the Phillies) would be $56. The bid price, in baseball, is always $4 less than the ask price. At the end of the interactive event, shares of the winning team will be worth $100 and shares of the losing team will be worth $0. At any point while the outcome is in doubt, the shares are made available with prices in the range between $0 and $100, with, again, always that four dollar spread between the bid and asked prices. So, for example, if the Braves are trailing by two runs in the top of the fourth, their new price might be 30-34. A player who buys the Braves for $34 needs to think one of the following is true: (a) The Braves have a greater than 34% chance of winning the game. This would be akin to a straight bet placed for value. (b) The Braves are going to begin playing better at some point in the game (or the Phillies worse), causing the Braves ask price to rise above $34 allowing the player to sell out for a profit. In summary when you bet interactives you can try to capitalize on near term reversals in the teams’ fortunes, or you can take your position relative to an anticipation of the final result.
I have never done any interactive wagering, but when it first started I would sometimes log onto the WSEX website to watch their prices as an event unfolded and I’ve always been a little bit amazed at what their price-setter had to do. It seems to me that anyone who thinks about the art of odds-making has to admit that interactive gaming forces the handicapper and the bookmaker to face a more complex set of issues than any other type of wagering. Take a baseball game in which home team A takes the field as a -180 favorite and on the first pitch of the game the lead-off man grounds out to the second baseman. How do you adjust the chances to accommodate the fact that one team now has 1/27th of the outs they’ll need to win the game, and 27/26ths (or for totals futures perhaps 24/26ths) as many possible offensive opportunities as the opposing team? Is the value of getting the opposition’s first out adjusted more or less for, say, a +130 dog as opposed to the -180 favorite? How do you calculate exactly how much more or less? Would it matter to the chances if the lead-off man had hit a vicious line-drive that was speared by the third baseman instead of that scooter to the second baseman we supposed? And what if the pitcher has a track record as a fast starter? Does the value of an early out, in that case, need to be discounted in anticipation of the higher probability that middle-game outs are likely to be more difficult for him? And what if the grounder to the second baseman was hit on a 0-3 instead of a 0-0 count? Does this slight indication of possible wildness need to be factored into the price? How does the value of a count shift if there is a runner on second? One out? Two outs? How valuable is the first out compared to the second?
The interactive oddsmaker and his players have to have some feel for all these probabilities and they need to have them on the instant. If you’ve ever faced the pressures of making odds that real bettors fire into then you know it’s a weakest link kind of thing. In other words, you can be doing it just right almost all of the time, but a mistake can torch a lot of good work. To put up 180 lines a game, usually on more than one game a day, and to accept on almost all of them several thousands of dollars worth of bets from anybody who thinks they have an aim on them is an enormous achievement in the field of oddsmaking. So everyone with a clue tips their cap to the man who does it day in and day out. He posts on this board as WSEXSteve, and we not only tip our caps to him but we’re grateful for his contributions to our discussions.
Even though the baseball probability combinatorics are enough to make anyone’s head spin, one realizes that baseball is a game of discrete definite situations which could be coded into tables and assigned probabilities which could be multiplied together to come up with something like an objective answer to the question of the value of any possible game event at any given point of any particular contest. When I first began to think about what the WSEX price-maker had to do to set good prices I felt certain that he must have some kind of tables, perhaps established by guess and by gosh, to guide him through that staggering number of instant decisions he has to make. At the same time I wondered if there might exist somewhere some mathematically alert handicapper who could apply himself to the same set of problems and solve them better than had the powers that be at WSEX. Then on April 14, in the Mess Hall at MajorWager, WSEXSteve himself delivered the message that this Neo—whose powers over the Matrix had come to exceed those of its creators—was actually amongst us.
Sunday, April 14, 2002 12:28 PM
I think I know who you are. If I am right you are our best interactive player. Sometimes when I make what I consider to be a bad market, I go back and try to change it. Nine times out of ten you have already beaten me to it. If you ever want a job let me know.
This message sent me searching through Machiavelli’s posts, and after reading them I guessed that he was young as well as knowledgeable. I knew that talented gamblers, like the mathematicians whose talent theirs resembles, tend to make their best discoveries at the beginning of their careers. So I wanted to meet this Machiavelli and discuss with him our art. I met the twenty-five year old in the condominium that he paid for in full with money won mostly from WSEXSteve. He moved into it straight out of his old room in his parent’s house where he began his full-time gambling work and became a hundred-thousandaire. We talked for the better part of two days, and I watched him trade a golf tournament and a couple of baseball games. Mach’s development as a gambler began when he discovered the literature of blackjack at age twelve. Like many gamblers who were drawn to theories of gambling games before having had much experience playing them, Mach read his blackjack books like sacred texts. Such texts speak to one of another world where, in proportion to one’s devotion and mastery, causality enforces high scruples and absolute justice reigns. In Mach’s case the books sparked what we might call technical daydreams. For example, during history class when the teacher was going on about trench warfare in World War One, Mach might have been imagining a deck of cards without sixes in them. Now many of us, perhaps with some prompting from an old statistics textbook, would be able to take a paper and pencil and figure out what the consequences for the various situations of blackjack are when a deck doesn’t have any sixes in it. But mathematical educators tell us that when such a problem comes to occupy a precious child the fascination has a quality that exceeds the search for an answer and becomes the creation of a world. For if certain conclusions can be drawn about the chances of being dealt a twenty-one or the strategic implication for standing or hitting from a deck without sixes, the child hasn’t so much solved the problem as increased his own ability to take on more complex ones.
So it is that the mathematically inclined youngster, after solving to his own satisfaction the problem of a deck without sixes, instantly begins to wonder about a deck without, for example, eights and tens. In this way he increases his ability to inhabit worlds whose extent is limitless without being infinite. They’re kind of weird places to be, I would think. But—and let this be marked as pure speculation on my part, Mach and I didn’t talk about this exactly—the organizational capacity that enabled him to solve analytic problems specific to the WSEX interactives was born from the capacity he developed building counter-worlds from decks of cards during years of patient technical daydreaming.
After high school Mach attended and graduated from Lehigh College where he studied electrical engineering and often hung out at nearby Penn National Racecourse. Of course during this time he always looked forward to the day he could go to Atlantic City and legally apply himself to the game that he had formed so much of his mind around. After graduation he got a regular job, but left it after less than a year when a bunch of huge middles happened to fall for him one right after the other.
Always a big sports fan, Mach was destined to turn his attention from blackjack to the emerging opportunities to bet on the net. As an engineering student he learned to see the problem of handicapping as a problem of modeling, and he set about constructing mathematical ways of thinking about the value of points. So, for example, box scores told him that a turn-over in a basketball game is worth on average one point. Therefore, a basket, which causes a change in possession equivalent to a turnover, is really only worth one point. It follows, then, that a three-point basket is worth twice as much as a two-pointer. In college basketball, where the three-point shooting percentages are higher than in the pros owing to the closer arc, the insight formed the basis of a profitable rating procedure.
Other insights required less inferential power than the analysis that revealed to him the secret power of the three-pointer. For example an early advance in Mach’s bankroll came when he discovered several books that allowed bettors to parlay individual NHL totals to the Grand Salami (an over-under bet on the total number of goals scored in all the games of a given night). Mach didn’t know a lot about the NHL, but he knew a dependent prop when he saw one and he hammered those lines with all his bankroll every night by cross-parlaying books against each other. Mach had soon taught those books a little more respect for the nature of correlation, and they paid him well for the education.
About interactive wagering itself one of Mach’s early statistical discoveries was that once a game gets going people tend to forget the initial probabilities. In other words, if a heavy favorite and a dog are tied toward the end of a game people are too anchored to the drama of the game and seem to want to value it as closer to a pick-em than in fact long term results say they should. This was an interesting discovery because the interpretive stereotype has the square over-estimating the favorite’s chances, but in many interactive wagering situations Mach’s numbers told him that this wasn’t in fact the valuation pattern.
When trading a baseball game Mach has his computer set up to give him a new set of prices based on the score, original odds, and point in the game every half inning. So while the price adjustments he makes from half inning to half inning are based on his spreadsheet wizardry, he moves the prices from out to out within the inning by adjustments he more or less pulls from his own head. As I watched him trade the game he talked me through these adjustments. Below is a sample of the kinds of things he was saying to himself in the seconds before he put in his orders.
· Runs are worth twelve bucks at this juncture… · A double-play would be an eleven dollar game event, and Anderson is a slow runner… · This inning is worth fourteen dollars so that out is worth six dollars, at least… · This [early] inning is only worth four dollars to Chicago, that’s how confident the computer is that they’re going to get them out… · The Padilla situation [early in the game it looked like Philadelphia’s pitcher Vicente Padilla might be hurt] is keeping me from making trades the computer is telling me to make…
My sense was that there was a fair amount of intuition involved in these adjustments, though Mach seemed to think the adjustments were fairly standard and his tweaking came from a sense of what his normal distributions had revealed about the character of inner-inning events, and not from the kinds of things one might know by virtue of being a sharp baseball fan.
A couple of times Mach felt the WSEX oddsmaker was making a mistake, but the discrepancy in the price wasn’t quite enough to get him to pull the trigger. In one instance I recall the price was two dollars too high, and in that range Mach wanted three dollars to put in an order. He refreshed the screen continually waiting for extra dollar of value to show up. The next batter got a double, and when he next refreshed the screen he swore because he believed that the odds maker had misvalued the double by two dollars in the opposite direction of his original error. “Oh No!” he said, thumping his desk, now he’s exactly right for the wrong reasons!” The two mis-evaluations had canceled each other out. Such are the fine-grained frustrations of WSEX’s best interactive player.
As successful as he’s been gambling Mach doesn’t see it as his calling. For one thing he feels the stigma of what he’s doing, and hates explaining what he does for a living in social circles. When the term “gambling” comes up, people unfailingly react in uncomprehending or falsely comprehending ways. Mach is a candid and open guy who doesn’t want to do his work behind a cover story, and he’s young enough to think he has other worlds to conquer. He also wonders how much longer his matrix will be better than WSEXSteve’s. Surely readings are being taken on what he’s doing, and at some point the WSEX technologists will ferret out what his computer has ferreted out about their systematic biases in valuing game events.
Mach has looked at the other interactive companies, like Tradesports, but he finds that they rarely have a market maker who offers enough action to accommodate his orders. He has an account with Tradesports and uses it when he finds someone keeping a big inventory on hand, but where as WSEX’s markets are consistently deep at Tradesport it’s hit or miss. Also, unlike WSEX, Tradesports charges a small transaction fee for each trade and another fee for holding shares at the expiration of the event. He feels that the delicate incubation process of the member-to-member exchange has been stifled by the lack of guaranteed depth and the dual fees. The result has been, from his point of view, an illiquid market with few good opportunities to trade. Mach’s records tell him that since he joined tradesports in late April ninety one percent of his in-progress trading volume has taken place at WSEX, with the remaining nine percent occurring at Tradesports.
So what might he be doing in five years if he’s not gambling? He’s currently reading up on day trading, and thinking of turning his blackjack imagination to anticipating the behavior of bulls and bears perhaps in the currency market. I hope he’ll let me pay him another visit in five years so I can report back on his progress.