Глава 1
When a Broken Leg Leads to a Baseball Betting Revolution
On a cold January morning in 2011, I found myself in a wheelchair in a New York apartment, unemployed and placing bets on baseball games. Just months earlier, I had occupied a lucrative seat on a Wall Street trading desk with a family in San Francisco. Then an FDNY ambulance struck me while crossing West Broadway, leaving me with shattered bones in my right leg and a dozen screws and a plate after surgery. Six weeks after returning to work in a wheelchair, Nomura Securities fired me-they literally rolled me off the trading floor.
Jobless, immobile, and separated from my family, I inhabited a dark mental space. Unable to focus on books or movies, I suffered nightmares and obsessed over "what if" scenarios. What saved me was baseball-specifically, creating a model that combined my love of trading, statistical analysis, and sports betting to beat the Vegas baseball line.
This story of recovery through baseball analytics would become a sensation in both financial and sports circles. The book became required reading at several business schools, was featured in The Wall Street Journal, and even caught Warren Buffett's attention, who called it "a smart, funny exploration of the beautiful mind of a professional sports bettor." As Joe Peta transformed from Wall Street trader to sports betting innovator, he created what one venture capitalist called "the holy grail of uncorrelated returns"-a baseball betting system that would ultimately deliver a 41% return in its first year.
Глава 2
The Statistical Edge: Discovering Cluster Luck
My journey began with a simple question that would eventually reshape my approach to sports analytics: why had the 2010 Tampa Bay Rays scored so many runs despite mediocre offensive statistics? With a .247 batting average and merely average slugging percentage, they somehow managed to produce run totals comparable to offensive powerhouses like the Yankees and Red Sox - teams that consistently outperformed them in traditional hitting metrics.
Statistical analysis revealed something fascinating - Tampa scored 802 runs when they should have scored only 701 based on their hitting metrics. Their hit-to-run ratio of 1.675 was the best of any team in five years, far exceeding the league average of 1.532. Through detailed regression analysis examining over 2,000 team seasons, I confirmed they had scored an incredible 78 runs above statistical expectations - nearly a half-run per game more than their underlying statistics suggested possible.
This anomaly led me to identify what I call "cluster luck" - Tampa's hits happened to occur disproportionately when runners were on base, a phenomenon that defied probability. While all MLB hitters improved with runners on base (typically +12/+24/+15 improvements in BA/OBP/SLG), Tampa's improvement was extraordinary (+39/+45/+39). Carl Crawford exemplified this pattern with staggering improvement splits (+77/+78/+107) that far exceeded his career norms. Even more telling, their team batting average jumped from .238 with bases empty to an astounding .286 with runners in scoring position.
This wasn't repeatable skill but random distribution - like flipping a coin and getting ten heads in a row. Historical data showed that teams experiencing extreme cluster luck invariably regressed toward league averages the following season. After analyzing roster changes and applying regression models, I concluded the 2011 Rays would score about 100 fewer runs due to cluster luck regression and personnel changes. The question became: how could I profit from this insight in a market that hadn't recognized this statistical anomaly?
This discovery parallels what happens in financial markets when stocks become overvalued based on unsustainable metrics. Just as savvy investors short overvalued companies expecting eventual price corrections, I could "short" overvalued baseball teams by betting against them when the market (Vegas) hadn't accounted for statistical regression. This realization opened up a whole new approach to sports betting - one based not on gut feelings or traditional statistics, but on identifying and exploiting mathematical inefficiencies in how runs are scored and games are won.
Глава 3
Building the Model: How to Quantify a Team's True Value
To create accurate predictions, I needed to quantify each player's contribution to their team's success. The 2011 MVP debate around Justin Verlander perfectly illustrates this concept. While some argued pitchers shouldn't win MVP because they only play one-fifth of games, sabermetrics shows a different reality.
A team's success comes from roughly 12,400 plate appearances per season. Cy Young winners face an average of 938 batters yearly, while MVP position players average only 674 plate appearances. When Verlander pitched in 2011, the Tigers went 25-9, versus 70-58 without him.
Using the Pythagorean theorem (a statistical formula that predicts wins based on runs scored and allowed), I calculated that if Verlander started all 162 games (an unrealistic scenario), the Tigers would win approximately 115-116 games-not the 120+ most fans estimated. This exercise showed Verlander was worth about five more wins than another Tigers starter-valuable, but not enough to transform a last-place team into a contender.
This critical reasoning approach allowed me to quantify roster changes in terms of wins. The Phillies provided a fascinating case study after their 2010 season. Despite winning 97 games, my analysis showed they were lucky-their true talent level was closer to 90 wins.
The Pythagorean theorem reveals that for every 10 runs a team improves its differential, they can expect one more win. This elegant formula provides tremendous value to front offices-want five more wins? Find ways to score 50 more runs, allow 50 fewer, or some combination totaling 50.
When the Phillies signed Cliff Lee, I calculated precisely how this would impact their win total. In 2010, Jamie Moyer and Kyle Kendrick started 50 games, allowing 207 runs over 420 innings. Lee and Roy Oswalt projected to allow just 160 runs over 400 innings-a 47-run improvement worth nearly five wins.
By combining player projections with adjusted previous-year results, I created complete projected standings for the 2011 season. I then compared these projections against Vegas over-under lines, identifying nine teams with differences of five games or more-these became my futures positions.
Глава 4
Why Baseball Betting Offers Superior Advantages
Baseball betting provides fundamental advantages over other sports that make it particularly attractive for analytical approaches. Unlike point-spread sports where players' incentives often conflict with bettors' interests (like a football team running out the clock rather than scoring), baseball uses a money line system where both teams are simply trying to win the game. This alignment creates a purer betting environment where game strategy and betting strategy naturally coincide - there's no concern about teams playing conservatively to protect a lead at the expense of covering a spread.
The house edge in baseball betting is remarkably lower than other sports. While football and basketball carry a consistent 4.55% house edge through the standard -110 point spread, baseball's "dime line" system creates a house advantage that rarely exceeds 2.44% and often decreases as the favorite's price rises. In football, each team has implied odds of 52.38%, creating 4.76% of "juice" (excess over 100%), while baseball's standard -105 line for evenly matched teams creates just 2.44% juice. This lower house edge means bettors need a smaller edge to become profitable - a crucial advantage for long-term success.
Baseball also offers superior modeling potential due to its discrete nature. Individual performance can be isolated with remarkable precision - like Randy Johnson's consistent 34% strikeout rate across eight years regardless of team, league, ballpark or opposition. Similar patterns emerge with batting averages, on-base percentages, and ERA statistics that remain relatively stable across different contexts. This statistical isolation is nearly impossible in football or basketball, where performance depends heavily on teammates and specific matchups. A quarterback's completion percentage, for instance, can vary dramatically based on offensive line protection, receiver quality, and opposing defensive schemes.
The volume advantage in baseball betting cannot be overstated - 2,430 MLB games versus just 256 NFL games per season provides nearly ten times more opportunities to apply analytical edges. This increased volume allows for more rapid validation of betting systems, faster accumulation of meaningful sample sizes, and greater ability to overcome short-term variance. Additionally, the daily nature of baseball games enables more frequent deployment of betting strategies and quicker adjustment to market inefficiencies.
Weather factors also play a more predictable role in baseball. Wind direction and speed, temperature, and humidity all have measurable effects on ball flight and scoring potential. Sophisticated bettors can incorporate these meteorological factors into their models with greater precision than in other sports.
As one veteran sports bettor told me, "Baseball is the last frontier where a smart bettor can still consistently beat the house. The edges are smaller than they used to be, but they're still there if you know where to look." This observation is particularly relevant given the wealth of publicly available statistical data in baseball, from pitch-by-pitch information to detailed spray charts and defensive positioning data - all of which can be leveraged for betting advantages.
The combination of aligned interests between bettor and team, lower house advantage, behavior that can be modeled precisely, and abundant opportunities makes baseball betting uniquely attractive for systematic approaches. The sport's rich statistical tradition and growing availability of advanced metrics continue to provide new angles for analytical bettors to explore.
Глава 5
Risk Management: Lessons from Lehman Brothers' Collapse
The fundamental principle of business valuation - that a company's worth lies primarily in its future earnings - forms the foundation of my risk management approach. This concept reveals that no short-term profit justification can ever outweigh risking an enterprise's future earnings stream. The present value of all future cash flows represents not just theoretical value but the very essence of what makes a business worth owning.
Dick Fuld, Lehman Brothers' CEO during its 2008 collapse, failed to grasp this essential truth. While investing in real estate wasn't inherently wrong, Fuld's unforgivable error was placing a bet so massive it jeopardized the entire company's existence. The firm's real estate positions grew to more than $50 billion - nearly double its total equity. When questioned about this concentration, Fuld consistently dismissed concerns, believing the firm's diversification across commercial and residential properties provided adequate protection. Despite his continued blame of external factors like market manipulation and regulatory failure, the simple truth remains that Lehman went bankrupt because it risked everything on the American real estate market.
I witnessed this deterioration firsthand during my thirteen years at Lehman. By early 2008, warning signs were everywhere: rapidly declining real estate values, widening credit spreads, and mounting losses in our mortgage portfolio. By June 2008, just months before its collapse, a colleague with connections to the executive committee confirmed my fears: "The firm is insolvent and everyone on the 31st floor knows it." The trading floor buzzed with rumors of desperate attempts to raise capital and potential buyers, but the underlying problem remained - we had bet too much on a single thesis. Three months later, Lehman Brothers filed for bankruptcy, its stock went to zero, and shareholders' claims on future earnings vanished.
The Lehman Brothers collapse offers a crucial lesson for my baseball betting model: never make a bet that imperils your ability to exist the next day. With a model that has positive expected value, the goal should be measuring appreciation in basis points, not risking everything on a single wager - even one with a calculated 14% edge. This principle applies whether you're managing billions in institutional money or making individual sports bets.
I established a non-linear betting system where only games with exceptional edges (over 15% advantage) would receive substantial 2% allocations, while most wagers would risk mere basis points of capital. Even these larger bets were carefully structured with stop-loss provisions and position limits. This approach ensured that if the fund failed, it would be because the model didn't work, not because of poor risk management. We implemented daily value-at-risk calculations and stress testing scenarios to monitor our exposure.
As Warren Buffett famously said, "In order to succeed, you must first survive." This principle guided every aspect of my betting strategy. I created multiple layers of risk controls: position limits, correlation analysis, and careful liquidity management. The system was designed to weather not just normal market conditions but extreme scenarios - including the possibility that our edge calculations could be wrong by several percentage points.
Глава 6
The Human Element: Baseball's Emotional Connection
While statistics and algorithms drive my predictive model, baseball's inherent beauty and grace fuel my enduring passion. After watching Justin Verlander throw a devastating curveball that literally buckled Michael Young's entire body during a critical ALCS moment - the ball seeming to defy physics as it dropped from shoulder to knees - I realized my complex spreadsheets were merely tools to engage more deeply with what I truly loved: the game's ability to create moments of pure athletic poetry.
Baseball uniquely creates connections across generations in ways other sports rarely achieve. With my older daughter Lily, I've transformed mundane math lessons into baseball stories, linking multiplication tables to DiMaggio's legendary 56-game hitting streak and teaching percentages through batting averages. For my mother, baseball has become a bittersweet measure of my father's gradual decline into Alzheimer's - each forgotten player name and misremembered game score marking the progression of his illness. My father, an Italian immigrant's son born in 1929 amid the Great Depression, embraced baseball alongside the English language, eventually becoming both my dedicated Little League coach and, remarkably, an English professor who could quote both Shakespeare and baseball statistics with equal facility.
In May 2011, with my leg sufficiently healed from surgery to handle travel, I returned to San Francisco's AT&T Park. During interleague play weekend, I temporarily suspended my betting model operations and took my five-year-old daughter Calista to her first baseball game. When she burst out with "Daddy! I see grass!" upon glimpsing the perfectly manicured emerald field, I felt tears welling up, recognizing in that moment the same intergenerational connection I'd experienced with my father decades earlier at Yankee Stadium.
These ballpark moments create indelible memories that transcend the game itself. Baseball connects us across time in profound ways - just as parents never receive warning bells for the last time they'll read Goodnight Moon or help with ballerina tights, baseball provides these poignant markers of childhood and connection. The crack of the bat, the smell of fresh-cut grass, the vendor's cry of "Hot dogs!" - these sensory experiences become time capsules of shared moments between parents and children.
As screenwriter Aaron Sorkin eloquently noted in Moneyball, "How can you not be romantic about baseball?" Though football and basketball may dominate American sports culture now, commanding bigger audiences and advertising dollars, when I hear the phrase "Hey, Dad, wanna have a catch?" I still picture a well-worn baseball glove, not a football. The game maintains its unique ability to pause time, to create spaces for conversation and connection between pitches. I remain romantic not just about baseball's pastoral rhythms and strategic complexities, but even about betting on it - because at its heart, baseball remains a game of hope, of possibilities, and of bonds that span generations.
Глава 7
The Model in Action: Results and Analysis
After months of meticulous preparation and system refinement, I approached MLB's Opening Day 2011 with the calculated anticipation of a trader before a pivotal earnings announcement. My model processed vast amounts of historical data, player statistics, and team dynamics to convert comprehensive team assessments into precise single-game projections, each generating specific win probabilities and betting recommendations.
The Milwaukee-Cincinnati opener provided an early test case for the model's capabilities. I projected the Reds with a 57.62% chance of winning compared to the oddsmakers' implied 51.22% - a significant 6.4% edge that exemplified the type of market inefficiency I sought to exploit. Scanning across the full slate of opening games, the model identified fifteen plays with positive expected value. The projected edges varied dramatically, from a minimal 0.13% advantage in the Phillies-Astros matchup to an exceptional 14.13% edge in the Cardinals-Padres game - highlighting the model's ability to detect both subtle and substantial market mispricings.
April's performance validated the model's methodology in unexpected ways. Despite a losing record of 166-174 in selected games, the portfolio generated an impressive 8.79% return. This seemingly paradoxical outcome stemmed from two key factors: underdogs significantly outperformed expectations, posting a 79-73 record with a +15.15% return, while favorites struggled with a 78-89 record and -6.25% loss. Most notably, the model's highest-conviction bets delivered exceptional results - 2% allocation bets went a perfect 2-0, 1.5% bets achieved a strong 7-3 record, and 1% bets impressed with an 11-4 performance.
May's results further confirmed the model's effectiveness, generating a 12.03% return and pushing the year-to-date gains to 21.87%. The system demonstrated remarkable consistency, selecting 368 of 420 games played and achieving a winning record of 199-169. Favorites particularly excelled during this period, posting a dominant 124-83 record with a +6.63% return - suggesting the market had overcorrected from April's underdog success.
The middle months presented challenges, particularly when the Minnesota Twins defied statistical projections throughout June and July, impacting overall performance. However, the model showed resilience by delivering strong results in August and September. The regular season concluded with a portfolio return of 32.83%, combining the robust 30.18% daily return with a supplementary 2.65% from strategic futures positions.
The post-season trading proved equally successful, adding another 6.17% to bring the final year-to-date return to an impressive 41.03%. This comprehensive performance - spanning over 2,400 games and thousands of individual trades - prompted one industry observer to make an apt analogy: "You built an airplane from scratch, and it flew." This success validated not just the model's predictive capabilities, but the entire systematic approach to sports betting.
Глава 8
Beyond the Numbers: What Baseball Can Teach Wall Street
Major League Baseball generates $6.4 billion in annual revenue with operating profits around 7%. While impressive, Wall Street equity trading desks generate four times that revenue with better margins. Yet baseball teams have a much better grasp of personnel value than financial firms do.
Thanks to the sabermetric revolution, MLB teams collect, analyze and interpret data to determine player worth with remarkable precision. Front offices understand the marginal value of a win (about $5 million) and can project player contributions using metrics like WAR, allowing for sophisticated cost-benefit analyses when signing players.
While baseball front offices use advanced statistical analysis to identify skill sets that predict future performance, financial industry managers rely primarily on recent results. A baseball GM would identify Brandon Beachy as their best pitcher based on his repeatable skill set, regardless of wins or ERA. Meanwhile, financial firms simply point to traders with the highest P&L or portfolio managers with the best recent returns, without accounting for contextual factors.
Portfolio managers control two primary variables: market exposure (the percentage of assets invested long or short) and stock selection. Success depends on both market timing and stock selection abilities, which require completely different skill sets-as different as base-running speed and batting power in baseball. Yet the financial industry rarely acknowledges or attempts to quantify these distinct skills.
In my hedge fund trading desk experience, I maintained a database of eleven hundred days of trades that revealed fascinating insights. One PM was an exceptional stock picker but terrible at predicting market moves. His constant adjustments to market exposure consistently destroyed alpha he'd created through stock selection. When presented with this data, the PM ignored it, preferring to focus on better salesmanship rather than improving performance-like trying to boost attendance without improving the team on the field.
Глава 9
From Model to Fund: Creating a Baseball Hedge Fund
The parallels between managing a baseball betting fund and running a hedge fund are striking. Like hedge funds trading securities in secondary markets, baseball betting involves analyzing implied rates versus expected rates to identify value. While financial markets provide crucial economic benefits through capital formation, secondary trading is fundamentally about making money through analytical skill-exactly what model-based baseball betting represents.
After submitting my manuscript, my editor suggested I go to Las Vegas with the book's marketing budget to bet on baseball games-an offer as irresistible as asking a Kardashian to make a sex tape. With my wife's blessing and legal consultation, I headed to Vegas after the 2012 All-Star Game to run a baseball-centric hedge fund backed by friends and family.
Surprisingly, my investors asked sophisticated financial questions about NAV changes, turnover rates and Sharpe Ratios rather than baseball statistics, with one venture capitalist friend calling it potentially "the holy grail of uncorrelated returns."
After three months of daily betting totaling nearly $3 million across multiple casinos, I discovered bookmaking resembles NASDAQ trading in the early 1990s-clubby, fragmented, and suspicious of customers. Despite my significant business, no casino manager ever proactively asked about my intentions or how they could better serve me.
The model performed impressively in 2012, finishing 12th out of over two thousand contestants in WagerMinds.com's season-long competition. The final 2012 return reached 14.01%, combining regular season performance (8.12%), futures positions (3.60%), and postseason results (2.05%).
My journey from broken leg to baseball betting success represents more than just a personal comeback story-it demonstrates how analytical approaches can find edges in markets that conventional wisdom considers efficient. By combining financial principles with baseball's statistical revolution, I created something truly unique: a system that not only beat the Vegas line but provided healing through the game I've loved since childhood.
As I watched the Cardinals celebrate their 2011 World Series championship after my final two small bets on the Rangers lost, I reflected on how baseball had transformed my life once again. The broken leg that seemed like a catastrophe had become the catalyst for an entirely new career path-one where I could apply Wall Street rigor to America's pastime and find both profit and passion in the process.
