Game of Edges The Analytics Revolution and the Future of Professional Sports

1. 1List the convention you intend to test and state it as a falsifiable claim. Examples from the passages: 'a shortstop plays where he always plays against a given hitter,' 'a shot worth three points is no better than a closer one worth two,' 'a player who doesn't look like a standout isn't one.' If you cannot say what result would prove the convention wrong, you have not yet framed the test.
2. 2Find the gap between assumed and actual value. Ask where the market, the league, or your own staff is paying for appearance, tradition, or easy-to-count stats such as goals, assists, or a player's look. Beane's insight started here: baseball people overweighted appearance, so talent sat unrecognized in players who didn't look like standouts.
3. 3Identify your constraint. Beane's A's, Morey's Rockets, and Leicester all worked within limits: a smaller budget, or a payroll already committed to two players under a salary cap. Constraint makes the inefficiency worth chasing, because you cannot outbid rivals and must out-evaluate them.
4. 4Start with the cheapest-to-verify inefficiency. Morey's three-pointer was the 'low-hanging fruit': three points instead of two is simple arithmetic, so he pushed the shot. Take the obvious, checkable case first, before you build subtler models.
5. 5Then look for attributes that contribute to winning without showing up in conventional metrics. Morey did this for NBA talent, and Battier treated the court as a decision tree with one optimal action per moment. Define what the right answer is at each decision point, then measure how often a player or tactic produces it.
6. 6Improve the measurement itself, because the data is 'a foggy lens.' Graham's model could evaluate only passes, shots, and ball movement from the official play-by-play. Spearman's video-tracking model scored everything that happened to everyone, even off the ball: a fullback's run, a striker's positioning for a cross. Ask what your data omits before you trust a verdict.
7. 7Implement the change and commit through the lean period. Morey's Rockets missed the playoffs three straight seasons from 2009–10 before reaching the playoffs in 2013 and the conference final in 2015. Set the evaluation horizon before you start, since short samples mislead in both directions.
8. 8Build the organization to keep questioning. Lacob ran the Warriors on venture-capital precepts: open communication, outside advisers, and 'continuous reevaluation of what companies do and how they do it.' Make re-testing a standing process, not a one-time project.

Daryl Morey's Rockets show the full cycle. The convention was that teams should work the ball inside for an easy jumper, lay-up, or dunk. Morey tested it against arithmetic: a made three-pointer is worth 50 percent more than a two. In 2005–06 the Rockets attempted slightly more than 17 threes a game. He hired Sam Hinkie, urged coaches to design plays that got open shooters behind the line, and by 2019–20 the team launched more than 45 a game. In 2017–18 they became the first NBA team to attempt more threes than twos in a season. The cost was real: three straight playoff misses from 2009–10 and an SB Nation headline calling him 'The Genius Who Outsmarted Himself.' The edge also eroded, because Golden State, with Steph Curry, applied the same insight better and beat Houston in the playoffs in four of five seasons from 2014–15 through 2018–19. The lesson: the test was right, but the advantage lasts only while others haven't caught up.

• If a convention has an arithmetic or measurable basis you can check in an afternoon, then test it first and act on it before building complex models (Morey's three-pointer).
• If your budget or cap is smaller than your rivals', then look for value in attributes the market ignores instead of bidding for what it already prices.
• If the available data misses large parts of the game (off-ball movement, how hard a pass was hit), then treat any verdict as provisional and invest in better measurement before deciding.
• If the results are poor over a short sample, then ask whether the sample is too small to judge before abandoning the approach, and agree on the evaluation horizon beforehand.
• If you notice yourself dismissing a person or idea because of style (Keri disdaining Yost's brusque, traditional manner), then evaluate the substance on its results instead.

• Replacing one orthodoxy with another. Keri said 'We hate orthodoxy' yet judged Yost on persona, the same appearance-over-substance mistake the lifers made about Beane. Tell: you rate people by whether they sound analytical rather than by outcomes.
• Judging too early on a small sample. DePodesta's short-term assessments worked out poorly, and writers nicknamed him 'Google Boy' while the owner read the criticism daily; he was fired after one disastrous season. Tell: no pre-agreed horizon, and the owner or board is reading the press instead of the model.
• Trusting a model beyond what its data can see. Graham's model could only evaluate recorded passes, shots, and ball movements, and he could not say whether Keïta played well until he checked it. Tell: you defend a player the eye test contradicts without knowing what the data omits.
• Assuming the edge lasts. Houston's insight was matched and improved on by Golden State. Tell: your approach is now widely copied, and you are still citing its early success.
• Questioning without owner or organizational buy-in. Analytics work stalled or ended where leadership was skeptical. Tell: the analyst is isolated from the decision-makers, as when most Premier League managers and executives rarely understood the concept.

• ▸Convention: state the long-held assumption in one sentence.
• ▸Falsifiable claim: what measurable result would prove it wrong?
• ▸Assumed vs. actual value: where do you suspect the market or your peers misprice a player, tactic, or practice, and why?
• ▸Constraint: what budget, cap, or resource limit makes this gap worth pursuing?
• ▸Cheapest test: the simplest check available (arithmetic, existing data, a small trial).
• ▸Data blind spots: what does your data not capture, and how will you reduce the fog?
• ▸Evaluation horizon: how many games, seasons, or periods before you judge, and what sample size is enough?
• ▸Stakeholder buy-in: who must tolerate the lean period, and what have you told them?
• ▸Re-test date: when will you re-examine whether rivals have caught up or the inefficiency has closed?

1. 1Write the decision down on the day you make it, with the information available that day. DePodesta's rule: "You had to make the pick today, so was it a good decision?" Do not defer judgment to "we'll see in five or six years."
2. 2Define the gap you are betting on. Henry's and Beane's frame is that a gap between actual and perceived value is the opportunity. State what the market believes, what your data says the asset is worth, and why they differ.
3. 3Evaluate with information the consensus is not using. Beane saw that using the same information as the wealthier teams would make him covet the same players and lose the bidding. If your inputs match everyone else's, your conclusion will too.
4. 4State the probability, not a promise. Record that the decision gives the best possible chance of success, not a guarantee. DePodesta's trades included Brad Penny (worked) and Hee-seop Choi (didn't), and both were consistent with the same process.
5. 5Strip out sentiment by stating what you want separately from what the numbers say. Henry's version: whether he wanted to own a soccer team had little to do with whether his company should buy one. Write the emotional preference down and exclude it from the verdict.
6. 6Gather the evidence before concluding. Henry collected as much information as he could find, then analyzed it "until it revealed its secrets" (stocks, free agents, even restaurants in Italy). Do not stop at the first persuasive number.
7. 7Audit the process after the result, not the outcome. Ask whether the inputs were sound, whether the analysis was sound, and whether the sample was big enough to judge. If the answers are yes, move on. DePodesta called 2005 a "perilously small sample size."
8. 8Plan for the public and the principal. Decide in advance how you will keep the boss or owner on the process during a bad stretch. Unpopular, correct decisions fail when the person with authority reads only the critics.

Paul DePodesta, as Dodgers general manager, valued some players less than other GMs did and made a series of unpopular trades, sending off fan favorites for players whose value he believed was buried in their statistical profiles. Brad Penny worked out; Hee-seop Choi didn't. In 2004 the Dodgers won 93 games and the NL West. In 2005 they fell to 71–91, their worst season in more than a decade. DePodesta's position was that the perception of failure was incorrect: he had a better sense of player value than the Padres, Diamondbacks, and Giants, who finished ahead of him, because he had better information, not because he was more insightful. One season, he noted, is a perilously small sample. His test was "Was the process leading up to it sound? If it was, you move on." The limit is also in the passage: the writers nicknamed him "Google Boy" and said the season proved his method failed, and process discipline meant little unless the owner agreed.

• If you catch yourself judging a decision by one season or a few results, then ask whether the sample is large enough to say anything, and check the inputs instead.
• If the outcome was bad but the information, analysis, and probabilities were sound, then keep the process and move on. If the process was flawed, change it. Do not change it merely because it lost.
• If your evaluation matches the market's, then you have no edge. Look for information the consensus is not using before you bid.
• If a deal is attractive mainly because you would enjoy owning it, then set that aside and decide on value alone. Henry rejected NBA teams that looked ideal on first consideration.
• If the person with final authority reads the critics, then secure their agreement on the process and the timeline before executing an unpopular call.

• Results-based review: you judge the call only by whether it won. The tell is that the justification shifts with each result.
• Sentiment dressed as analysis: the preference came first and the data was gathered afterward. Sterling refused to sell the Clippers for more than they were worth because of what he would talk about at the country club. The tell is a reason that has nothing to do with value.
• Over-reading a small sample: one bad season is treated as proof the method fails, or one good season as proof it works. The tell is a verdict drawn from a single year.
• Using the same inputs as everyone else: you rely on the data the richer competitors use and end up coveting the same assets. The tell is that your shortlist matches the consensus.
• Analysis without a path to action: the work is produced but ignored, as at Wolves, where the analytics lead spent much of his time looking for someone to present findings as his own. The tell is that the manager never sees the model's output.

• ▸Decision and date: state what you are deciding and when, so it is judged on what you knew that day.
• ▸Market view vs. your view: what the consensus believes, what your analysis says, and the gap.
• ▸Information used: list the data sources, and note which ones the consensus does not use.
• ▸Probability of success: state the odds in plain terms; a guarantee is not an acceptable entry.
• ▸Sentiment check: write down what you want emotionally, then confirm it is excluded from the verdict.
• ▸Expected unpopularity: who will object, and what they will say if this loses.
• ▸Review horizon and sample size: how many instances you need before the result says anything.
• ▸Process verdict after the outcome: were the inputs and analysis sound? If yes, keep the process; if no, record the fix.
• ▸Principal sign-off: the owner or boss has agreed to the process and timeline in writing.

1. 1Define the market's consensus price signal first. Name the stat, reputation, or label that other buyers use to set value (in soccer, pass completion rate; in baseball, what counts as a successful plate appearance). The gap you exploit only exists relative to that signal.
2. 2Secure a data source others ignore or underuse. MLB transmits raw data to every team nightly; a few teams ignore it, others build a mechanism to download and convert it into insights. Your edge is the conversion step, not access to the feed.
3. 3Build your own valuation model rather than adopting the familiar metrics. Graham spent months building a model that calculated a team's chance of scoring before an action and immediately after it, then credited each player with the change.
4. 4Run the model against the market's view and sort players into three bins: overvalued, undervalued, and valued correctly. Players who rank well on familiar stats and also on your model are fairly priced. Players who rank at the bottom of your model despite good familiar stats are sell candidates.
5. 5Look for the behavior that explains the divergence. Keïta's low pass completion hurt his reputation, but his passes, if completed, put teammates in better-than-average scoring positions. Confirm you can state the mechanism, not just the number.
6. 6Act early and on a long horizon. Graham recommended Keïta in 2016; he arrived in 2018. Acquire before rivals reprice, and sell holdings the market still overvalues.
7. 7Check where you operate. The edge was largest in industries that disdain data (baseball for Beane, soccer for Graham). Prioritize such markets.
8. 8Plan for decay. Beane's A's finished first or second in their division for eight straight seasons only until other teams copied the strategy. Keep looking for the next inefficiency, such as new proprietary data from puck chips and uniform sensors.

Liverpool FC, owned by John Henry's Fenway Sports Group, hired Ian Graham, soccer's foremost analyst, and added a staff that included a particle physicist. Graham disdained raw statistics and even the more evolved metrics such as pass completion percentage. He built a model of each team's chance of scoring before and after every action (a pass, a missed shot, a slide tackle) and used it to quantify how much each player changed his team's chance of winning. Some players who looked good on familiar statistics also ranked high in his model, but others landed at the bottom, which showed him who was overvalued, undervalued, or correctly valued. Naby Keïta had a lower pass completion rate than some elite midfielders, and many observers at other European clubs marked him down for it. Graham's figures showed that Keïta often attempted passes that, if completed, put a teammate in a better-than-average position to score. Graham recommended pursuing him starting in 2016, and Keïta arrived in 2018. Liverpool won the Champions League in 2019 and its first Premier League title the next season, with the club's tactical insights and player advice credited as helping.

• If a player's market value rests on a familiar stat that your model does not support, then treat him as a sell candidate. If he ranks poorly on that stat but well in your model, treat him as a buy.
• If a player is only valued highly by teams that measure the relevant contribution (an average hitter who is a terrific fielder), then do not expect to buy him cheaply from them; look for sellers who are not measuring it.
• If rival clubs are using the same data and models, then the arbitrage is closing; move to a newer data source or a different inefficiency, as Beane's edge lasted only until others copied it.
• If the market in your sport traditionally resists quantification, then invest more in building the analytic capability, because the mispricing is likely larger.
• If your model cannot explain why the market misprices a player, then do not act on it; require a stated mechanism, as with Keïta's pass selection.

• Trusting the familiar stat. You buy the player with the best pass completion rate or batting line and pay a premium. The tell: your shortlist matches every other club's shortlist.
• Access without conversion. You receive the nightly data but never turn it into insight, which is what the teams that ignore it do. The tell: no model, no analyst, and no changed decisions.
• Assuming the edge is permanent. You keep running the same approach after rivals copy it. The tell: you no longer find bargains and your rivals' prices have caught up.
• Acting on a number without a mechanism. You buy a player who ranks high in your model without understanding why. The tell: you cannot explain in one sentence what he does that the market misses.
• Applying the strategy where the market already prices the trait. You pursue defensive value in a market that already measures it. The tell: the targets are priced above their familiar stats.

• ▸Asset: name the player, tactic, or position under review.
• ▸Market signal: the stat or reputation other buyers use to price it.
• ▸Market view: how others currently value it (high, low, fair) and why.
• ▸Your model's metric: the proprietary measure you use and what it captures (e.g., change in scoring chance before and after each action).
• ▸Model verdict: overvalued, undervalued, or correctly valued.
• ▸Mechanism: one sentence on the behavior that explains the gap (e.g., attempts passes that put teammates in better scoring positions).
• ▸Who else measures this: list rivals using similar data, and whether they price it in.
• ▸Action and horizon: buy, sell, or hold, with the timing of the move.
• ▸Decay check: the date to re-test whether the gap has closed.

1. 1Name the bond before you measure anything. Write one sentence that separates a member from a customer, using Robert Elms's test: his loyalty at Queens Park Rangers 'wasn't affected by whether his team happened to be up or down in any given season, or even if it won or lost.' Any decision you review gets checked against that sentence.
2. 2Audit your fixed-format 'best practices.' List every element of the game-day experience (announcer cadence, video presentation, music, anything the league office curates) and mark which ones came from a template. The source says the quirkiness separating one team from another 'had been deemed an inefficiency and removed,' and that this may have eroded 'some of the emotional attachment that drew us to sports in the first place.' Restore or protect the local elements that only your club has.
3. 3Price for attendance as well as yield. Before you set any price, estimate two values for each seat: the ticket revenue and the ancillary income (hot dogs, beer, parking). Then add the atmosphere value of a nearly full park, which is harder to quantify. The source gives the principle: 'There was value in having fans sitting in seats... no matter how much their tickets had cost them.'
4. 4Build a cheap on-ramp for the fan who wants nothing but the game. Keep at least one tier of good seats at a price a person with a school-night schedule can afford. The reason is stated: once someone attends a game, he or she is 'far more likely to develop an attachment to the team and return for more.'
5. 5Segment fans as individuals, not as a crowd. Sort your fan base by zip code, age, frequency of attendance, and spend. Use the segments to find who follows the team and to talk to them, instead of putting up a billboard on a freeway and hoping the right people glance up.
6. 6Design communal touchpoints for newcomers. Give casual visitors a reason to gather in person: a shared big screen, food and drink, promotions available only to those present. Aim the visit at a later purchase of merchandise or a ticket. The Cubs/DraftKings plan bets on this, hoping a casual visit becomes 'the catalyst for a lifelong relationship.'
7. 7Keep the local, tribal identity when you scale. Before expanding or adding properties, ask whether the move treats the club as a branch office. Peter Moore's warning: 'football is tribal and local,' and a conglomerate approach is 'ripping the soul out of what we're all about, a little bit.'
8. 8Treat analytics as one input to the decision. After the model gives its answer, ask what it can't see about the bond. Zue's rule: 'The analytics are certainly helpful, but what they're telling you isn't always the answer.'

The Red Sox ticketing team in 2014 faced a pricing choice. StubHub data showed the same infield seat that cost $52 in 2013 sold for an average of $96 across the 16 games designated Tier 1. Pure revenue logic pointed to charging near $96 for those games and maximizing the top seats. Bumgarner and Zue's team charged $70 instead. They offset most of the increase by cutting the same seat to $40 for the 16 Tier 5 games. A fan who wanted only to see the Red Sox could now sit between first and third base for $13 less than the year before, if willing to go on a weeknight when school was in session. Their reasoning was that empty seats at low-demand games, especially in years without a pennant race, would cost ancillary income and the atmosphere of a full ballpark. It would also cost the chance that a first-time attendee develops an attachment and returns. Zue summarized it: 'That's the ecosystem.'

• If the data says to raise price to the market ceiling, then check what happens to attendance at your lowest-demand games before you act. Keep a lower-priced tier so the seats are filled and the atmosphere survives.
• If a 'best practice' makes your game-day experience look like every other team's, then treat it as a possible cost to attachment and not as a pure efficiency gain. Keep the quirk unless you can show it doesn't matter.
• If a revenue initiative brings in casual or first-time visitors (betting parlors, events), then design the visit to end with a path to merchandise or a future ticket. Revenue from the visit alone does not build a bond.
• If a growth plan requires running multiple clubs or branches, then ask whether each club's local, tribal identity survives. If it doesn't, expect the brand to weaken even when the finances look good.
• If a metric conflicts with your read of what fans feel, then do not let the metric win by default. Ask what it leaves out, as Zue does.

• Treating fans as customers: you manage by transaction metrics only. The tell is Elms's line, 'the minute you start treating them like customers, they become customers,' and they leave after a bad experience as they would a restaurant with perfunctory service.
• Template-izing the experience: every park feels the same. The tell is that announcer cadence, video, and music are set by the league office and local charm has been labeled an inefficiency.
• Over-optimizing price: you charge the maximum for premium games and ignore the weak ones. The tell is a thin crowd at Tier 5 games and no new fans developing an attachment, even though the top-seat revenue looks better.
• Mistaking attention for attachment: you bring in casual bettors or visitors and count their spend as loyalty. The tell is that only the sportsbook or one-off purchases show up, with no merchandise or future ticket sales.
• Scaling like a conglomerate: you set up branch offices across clubs. The tell is language like Moore's, that the approach is 'ripping the soul out of what we're all about,' and fans describing it as detached from the local club.

• ▸Decision under review: describe the pricing, experience, or expansion change in one line.
• ▸Member test: would a lifelong fan stay loyal through a losing season if this goes ahead? Yes/No and why.
• ▸Standardization check: which local or quirky element does this remove or replace with a league-wide template?
• ▸Attendance effect: estimate change in filled seats at the lowest-demand games.
• ▸Ancillary and atmosphere value: list concessions and parking income plus the effect on a full-park feel.
• ▸First-visit path: how does a new attendee get from one game to a return visit?
• ▸Segment targeted: name the fan group by zip code, age, attendance frequency, and spend.
• ▸Local identity risk: does this treat the club as a branch of a larger operation? Note the evidence.
• ▸What the analytics miss: write the one thing the data does not capture about this decision.
• ▸Final call: proceed, modify, or reject, with one-line reason.

1. 1Treat competitive success as a zero-sum metric and measure it against opponents, not against your own history. Every win you gain is a loss for someone else, so record wins, division finishes, playoff appearances, and titles relative to the league, not to a prior-season baseline.
2. 2Separate the two scoreboards before any decision. Track profit and on-field results as distinct metrics: they are not completely aligned. A team can lose for years and still turn a profit through league revenue sharing, and a team can win a title and lose money, as the Florida Marlins did.
3. 3Set an explicit competitive target and write it down before spending. Define the number you will accept (for example, playoff games played or a division finish), because owners with no stated standard drift to hobby-style or overconfident expectations. Snyder said 8–8 would make him want to kill himself; his team went 7–9.
4. 4Find the market inefficiency that your budget can actually exploit. If you cannot outspend rivals, look for attributes that contribute to winning without showing up in conventional metrics. Beane's A's did this with undervalued players. Morey did it in the NBA because Houston's payroll was already committed to two long-term contracts.
5. 5Convert the inefficiency into a repeatable tactic. Morey's math was simple: a three-pointer is worth 50 percent more than a two. The Rockets attempted slightly more than 17 threes a game in 2005–06 and more than 45 by 2019–20, after coaches designed plays to get open shooters beyond the arc.
6. 6Pair analytics with resources and the right hire. Epstein's Red Sox were described as Moneyball with money: owner John Henry used data in options trading, and analysts cost less than one Jason Giambi contract. Check that you have the owner, the capital, and the executive aligned.
7. 7Set a decay clock on every edge. Beane's A's finished first or second for eight straight seasons only until other teams copied the strategy. Plan the next edge, such as proprietary data (Liverpool's data work preceded a 2019 Champions League title and a 2020 Premier League title), before the current one is copied.
8. 8Audit what the optimization costs. Check whether the tactic is hurting the product: baseball has more strikeouts and home runs but fewer extra-base hits and fielding plays, and NBA ratings are down. Keep winning, but watch the fan-facing effects.

Daryl Morey's Houston Rockets show the full loop. When he became general manager in May 2007, the Rockets had less money than the teams they competed against, and much of their permitted payroll was already tied up in two long-term contracts (Tracy McGrady and a seven-foot-six center). Money could not close the gap, so he looked for value elsewhere. The first step was shot selection: a three-pointer is worth 50 percent more than a two, so he wanted players to take more of those harder shots instead of working the ball inside. With Sam Hinkie as his analyst, he urged coaches to design plays that got the ball to open shooters beyond the line. The Rockets went from slightly more than 17 attempts per game in 2005–06 to more than 45 by 2019–20, a new architecture for an NBA offense. Morey called that the low-hanging fruit. The deeper work was using data to find player attributes that help win games without appearing in conventional metrics. The cost showed up elsewhere: jump shots from the perimeter look indistinguishable, and the NBA's ratings are down.

• If your payroll is constrained or smaller than rivals', then hunt for undervalued attributes that conventional metrics miss, as Beane and Morey did, rather than bidding for the same players.
• If an owner's confidence comes from success in a non-zero-sum business, then reset expectations to the zero-sum reality, in which every win is someone else's loss, before approving the plan.
• If your edge has become public or copied, then treat it as spent: Beane's eight straight first- or second-place finishes ended once other teams copied the strategy. Move to the next data source.
• If a title or winning record will not cover costs, then decide in advance whether you are paying for competitive success or for profit; do not assume they move together. Huizenga sold off his best players after a Marlins title.
• If an analytic tactic raises wins but visibly reduces the on-field spectacle (fewer extra-base hits, uniform-looking jump shots), then monitor fan interest and ratings alongside the standings.

• Assuming business skill transfers to winning. The tell: an owner predicts easy dominance after success in a non-competitive industry. Snyder, a direct-mail founder, saw 7–9 in 1999 and only five playoff games in 21 seasons.
• Assuming profit tracks wins, or the reverse. The tell: a championship season followed by a fire sale, as with the Marlins, or a losing team described as fine because it is profitable.
• Copying an edge after it is already priced in. The tell: you adopt a tactic that every rival already uses, and your record reverts to the league average once others copy it, as with the A's.
• Buying analysts without the surrounding conditions. The tell: you hire a data-minded general manager but lack money or owner alignment. Epstein was seen as the success and DePodesta, hired by the big-budget Dodgers, as the failure.
• Optimizing the metric and ignoring the product. The tell: more strikeouts and home runs, fewer extra-base hits and acrobatic plays, falling ratings, and fans who say they miss the intimacy and informality of earlier eras.

• ▸Target outcome: write the specific wins, division finish, playoff games, or title you are aiming for, and the season by which you expect them.
• ▸Profit versus wins: state the budget you will accept losing in pursuit of the target, and whether the revenue sharing covers it.
• ▸Resource gap: list how your payroll or spending compares with the top rivals, and what is already committed long-term.
• ▸Inefficiency identified: name the attribute or tactic the market undervalues (for example, three-point attempts at 50 percent more value per make).
• ▸Baseline metric: record the current number (for example, 17 attempts per game) and the goal number (for example, 45).
• ▸Owner and executive fit: note who in the organization understands the data and who has the capital to act on it.
• ▸Copy date estimate: write when rivals will likely replicate the edge and what the next edge is.
• ▸Product check: list the fan-facing indicators you will monitor (ratings, in-play action, attendance) as the tactic spreads.
• ▸Result versus league: after the season, record wins and losses relative to opponents, not only relative to last year.

1. 1Anchor on the scarcity logic before you price anything. Count the supply: fewer than 100 people held majority stakes in one of the four major US leagues while about 500 billionaires existed, and teams grew only from 43 (1960) to 120 (2001) before expansion stopped. Treat the franchise as oceanfront real estate: fixed supply, rising demand.
2. 2Trace the value to its media driver. Live sports kept strong ratings while scripted shows fragmented, which pushed rights fees up. Check whether your league's rights deals are growing: NFL network payments went from $900 million a year in 1990 to $14 billion in long-term deals by 2002. Model franchise value as a share of that rights stream.
3. 3Build a price ladder for your asset. Record the purchase prices over time as the Pacers, Spurs, Suns and Clippers examples do ($4.5M in 1983, $75M in 1993, $401M a decade later, $2B in 2014). Compare your entry price with the last comparable sale to see how much appreciation is already priced in.
4. 4Separate the two scoreboards: winning and earning. The metrics are related but not aligned. A losing team can still turn a profit through league revenue sharing, and a title team can lose money (the Florida Marlins). Set a revenue target and a competitive target as separate line items.
5. 5Use competitive success as the lever on sponsorship and media rights. In most markets winning is the surest way to generate fan enthusiasm and the sponsorship and media money that follows. Budget on-field spending as a driver of off-field revenue, not as a cost centre.
6. 6Apply data analysis to the business side, not only the roster. Run the same dispassionate analysis you use on players against real estate, ticketing, and brand reach. John Henry's group made small investments in undervalued lots around Fenway to grow company value or set up later growth.
7. 7Open new revenue streams adjacent to the franchise. Ted Leonsis spread bets across nearly a dozen gambling-related plays: a bookmaking website, a game-data company, a gambling-ready sports bar in his arena, and a proposed interactive gambling channel on his regional sports network. List every stream (media, real estate, retail, gambling, hospitality) and assign an owner and a test for each.
8. 8Screen every acquisition by business merit, not fandom. Henry's group rejected NBA teams that looked ideal on first pass, and wanting a soccer team was irrelevant to whether the company should buy one. Write the investment case first, then check whether you like the team.

Take the Boston Red Sox under John Henry. In 2003 Henry and his partners paid $700 million for the team and some real estate, more than double the price ever paid for an MLB franchise. The Sox hadn't won a title since 1918, so the owners set out to build both a World Series winner and a business that would validate the price. Between 2004 and 2017 the team won four World Series. By 2010 relentless data analysis had helped win two titles and moved the club from a regional following to one of the most popular franchises in America. Off the field, New England Sports Ventures bought undervalued lots around Fenway, and it turned down NBA teams that looked ideal but didn't pass a business test. The same logic explains the silent-partner investors: Randy Vataha pointed out that once-blue-chip Boston companies such as Wang Laboratories vanished, while the Red Sox, Patriots, Celtics and Bruins were the four equities that gained the most value over 25 years.

• If the league's media rights keep growing while scripted TV fragments, then treat franchise appreciation as structurally supported and focus on capturing your share of it.
• If the league has frozen its team count (supply fixed, demand from billionaires rising), then accept a premium price for access, as Steve Ballmer did with the $2 billion Clippers, but only if your holding period can absorb that premium.
• If a team is attractive mainly because you want to own it, then stop and rebuild the case on cash flow, adjacent assets and portfolio fit before bidding, as Henry did with the soccer team.
• If a title cannot be shown to raise sponsorship and media revenue in your market, then don't pay for winning at the expense of profit; the Marlins won and still lost money.
• If analytics-driven play lowers the product's watchability (more strikeouts, indistinguishable three-pointers, falling ratings), then discount the growth forecast, because ratings feed rights fees and franchise value.

• Running the team as a hobby. Bob Tisch said he didn't care if he never made a cent on his $75 million; the tell is no revenue plan beyond ticket sales and no one accountable for growth.
• Assuming skill in another industry transfers. Daniel Snyder made his fortune in direct mail, where growth doesn't erode competitors, but football is zero-sum; the tell is a confident on-field promise (never going 8-8) followed by a 7-9 season and five playoff games in 21 seasons.
• Treating winning and profit as the same goal. The Marlins won a title and lost money, and Huizenga sold off his best players; the tell is a payroll built for a title with no matching revenue source.
• Overweighting on-field efficiency and hurting the product. Optimized baseball and three-point-heavy basketball coincided with falling ratings; the tell is declining TV numbers despite better analytics.
• Buying on emotion or prestige alone. Fans and owners want to talk about teams and watch practice, but the passage shows Henry's group declining attractive-looking NBA teams; the tell is a bid justified by the team's glamour, not by valuation or adjacent assets.

• ▸Entry price and year: record what you paid or the last comparable sale, and the date.
• ▸Scarcity check: note the league's team count, whether expansion is frozen, and how many people can realistically buy in.
• ▸Media rights trajectory: list current and prior rights-fee totals and the next renewal date.
• ▸Revenue stream inventory: for media, real estate, retail, gambling and hospitality, write the current status and one next test for each.
• ▸Adjacent assets: list undervalued land, venues or businesses near the franchise that could appreciate or support growth.
• ▸On-field vs. off-field link: state how a competitive result changes sponsorship and media revenue in your market.
• ▸Product-health signal: record current ratings and attendance trends and any style-of-play changes that may depress them.
• ▸Business case without fandom: write the investment rationale in two sentences with no reference to liking the team.
• ▸Go / no-go and review date: record the decision and the date you will re-check the numbers.

1. 1Name the entertainment unit you are protecting before you optimize anything. The source's candidates are balls in play, extra-base hits, acrobatic fielding plays, varied shot types, and a compelling narrative. Write them down. Research cited in the book says extra-base hits and acrobatic plays are what fans most want to see.
2. 2Run an optimization audit on every strategy you adopt. Ask what it raises (wins, efficiency) and what it removes from the watchable product. Baseball's efficient style raises strikeouts and home runs and lowers balls in play, extra-base hits, and acrobatic fielding. Basketball's perimeter emphasis produces jump shots that 'tend to look indistinguishable' where hooks, double-pump lay-ups, and dunks were singular.
3. 3Track the aesthetic side with the same discipline as the win side. Log how often the ball is in play, the mix of play types, and the viewing trend (the book ties falling ratings to baseball and to the NBA, while calling the NBA link possibly 'no accident'). Treat a ratings decline alongside an efficiency gain as a warning, not proof.
4. 4Check whether your metric actually predicts winning before you trust it. Soccer's corners and possession time look good in theory, but they help some teams more than others: teams with attackers skilled at redirecting centering passes work to create corners, while teams with finishers who elude defenders prefer open play and are not especially pleased when corners happen. Some teams do not even want possession.
5. 5Separate who bears the cost from who captures the gain. A team incentivized by winning (the Rays under Cash) is not at fault for an unwatchable style. The fix belongs to the league, so push for rule changes: in Morey's words, set up the field 'such that, when you do optimize for wins, it's also entertaining.'
6. 6Use rule tweaks as the precedent. The NBA added the three-point line to unpack the mass of bodies beneath the basket, then allowed zone defenses once perimeter shooting had grown good enough to counteract it. Propose a rule only when you can name the behavior it corrects.
7. 7Apply the same test off the field. Cleaner arenas, better food, sharper replays, and less-crowded corridors are real gains, yet fans report missing the intimacy and informality of corner-grocery-style teams. Before adding a business-driven change, ask what it costs in the fan-team relationship.
8. 8Discount the vested-interest voice. Morey called himself the wrong person to ask about entertainment because his stake in winning made aesthetics 'irrelevant.' Get the entertainment read from fans and viewership data, not from the optimizers.

Take baseball as the book frames it. The analytically optimal style raises strikeouts and home runs and cuts extra-base hits and acrobatic fielding plays, so the ball is in play less often. A front office that follows this is rational: Tampa's Cash and his superiors were 'incentivized by winning, not creating a compelling narrative.' The cost shows up in the audit. Research says fans most want extra-base hits and acrobatic plays, and television ratings have plummeted. Morey's remedy is that the league should change the rules so winning and entertainment line up again, and he cites the NBA, which added the three-point line and later allowed zone defenses as player ability evolved. The takeaway is to measure the watchable product and push for rule-level correction when the two diverge, since no single team can afford to unilaterally play 'less efficiently.'

• If a strategy raises win probability but lowers balls in play or play variety, then treat it as a league-level rules problem and flag it, rather than blaming the team for adopting it.
• If your team's results ride on a style fans find monotonous, then expect fan patience to last only as long as you win (Morey on Utah: fans want to watch it 'if Utah won'), and plan for the losing seasons.
• If a proxy statistic (corners, possession) is cited as proof of quality, then test it against how your specific team scores before treating it as good.
• If ratings fall while efficiency rises, then investigate the style change as a candidate cause, but do not claim causation from the correlation alone.
• If a rule change is proposed, then tie it to a specific distortion, as the NBA did with the three-point line (crowding under the basket) and zone defense (perimeter shooting).

• Treating winning as the only objective and assuming entertainment follows. The tell: the team wins while the play looks repetitive and ratings slide.
• Reading a stat as a goal in itself, such as chasing corners or possession. The tell: the number improves while results do not, or while the team's actual scorers prefer open play.
• Letting the vested party judge aesthetics. The tell: the person who profits from the strategy says entertainment value is irrelevant or that fans will accept it if you win.
• Upgrading the stadium experience while ignoring intimacy. The tell: the facility is cleaner and the food and replays are better, yet fans say the place feels corporate rather than like a corner grocery.
• Assuming rules are fixed. The tell: you accept a distortion as inevitable even though the NBA changed its rules as players' capabilities evolved.

• ▸Strategy or change being evaluated: describe the optimization in one line.
• ▸Expected win or revenue gain: state the efficiency benefit.
• ▸Watchable elements affected: list balls in play, extra-base hits, acrobatic plays, shot variety, narrative.
• ▸Direction of change: mark each element up, down, or unchanged.
• ▸Fan evidence: record ratings, attendance, and fan complaints, noting the trend direction.
• ▸Proxy-stat check: name any stat used as a goal and whether it fits your team's actual scoring style.
• ▸Off-field cost: note any loss of intimacy or informality alongside the comfort gains.
• ▸Rule-level remedy: propose a league rule change, naming the distortion it corrects.
• ▸Decision and review date: adopt, modify, or lobby for a rule change, and set when to re-audit.

1. 1Name the decision before you collect data. Liverpool did not hire Will Spearman to 'do analytics'. The question was whether the club could play 'in a slightly different way' using the trajectories of all players. Write one decision per project, in a sentence a manager could say yes or no to.
2. 2Inventory the data you can actually get, and map its gaps. Tracking data gives player trajectories at 25 frames per second, but not every European league has the cameras. Most of the young talent Liverpool scouts plays in leagues that have no tracking and may never get it. List which leagues, teams, or units have the data, and decide what you can infer from those that do.
3. 3Hire for modeling skill, not sport résumé. Spearman is a physicist. He recast player speed as 'charge' and, in one weekend, built a model that could show where each player controlled space at any moment. Theo Epstein had not played past high school and carried no preconceptions of how the game should look. Screen candidates on whether they can reshape raw data into a model, and treat playing background as secondary.
4. 4Price talent against what it replaces. One line from the source: you could hire a stadium full of young analysts with laptops for what the Yankees were paying Jason Giambi. Compare analyst salaries with one marginal player contract to see the budget logic.
5. 5Move quickly on rare talent. After Spearman's 2016 OptaPro presentation, Waskett called Graham and said, 'We have to hire this guy. We have to take him off the market.' Keep a short path from spotting a strong analyst, to the decision-maker, to an offer.
6. 6Apply the same machinery off the field. Januszewski's case for Liverpool was an asset analysis: a property whose potential value was higher than its price. Henry then spent a day reading and sifting on the club's structure, history, and city. The business levers were Anfield's ancillary revenue (a café, rentable conference rooms), nearly every ticket sold at the same price, and online content. Run the player-side process (question, data, model, comparison to a benchmark) on pricing, venues, and media.
7. 7Test the model against the sport's evidence and its limits. The Burnley match shows the risk. Liverpool had 80.6 percent possession and lost 2–0. Compare what the metric rewards with what wins, and write down which signals would make you distrust the output.
8. 8Build the reverse check into the process. In the 2020 World Series, the Rays pulled Blake Snell, who was pitching the best game of his life, because the data said to. The Dodgers then scored and won. Define in advance when an on-the-spot override of the model is allowed, and review those calls afterward.

Liverpool and Will Spearman. The club's analysts had tracking data: the trajectory of every player at 25 frames per second. Spearman, a physicist, treated each player's speed as a 'charge' that determines how much influence he exerts over space. He spent a weekend shaping the data into a more accurate model. By Monday he could calculate where on the field each player exerted control at any moment. He presented it at London's OptaPro conference in early 2016. Waskett watched, phoned Graham, and said they had to take him off the market. Graham explained the rationale: with the data available, the club could ask why it should not play football in a slightly different way. The limits were stated plainly. Not every league has the cameras, and most of the young talent Liverpool scouts is in leagues that never will, so the club relied on patterns inherent to the sport that show up in leagues that do track. The edge came from a specific question, a hire from outside the sport, a model built in days, and a fast offer.

• If you can afford one star or a whole analytics group, then price them against each other. Epstein's Red Sox implied a stadium full of analysts for roughly one Giambi contract.
• If a candidate shows a model that answers your question in new terms, as Spearman's did, then move before a competitor does; delay costs you the hire.
• If your data covers only some of the leagues, teams, or situations you care about, then build on the patterns that hold in the covered sample and label every conclusion about uncovered cases as inference.
• If a model's recommendation conflicts with what you can see in front of you, as with Snell in 2020, then do not override ad hoc. Follow the pre-agreed rule, and log the result either way. The Rays followed the data and lost; Friedman himself privately hoped the doctrine was wrong that night.
• If the optimal strategy lowers the quality of the product, then add fan appeal to the objective. Baseball's strikeouts and home runs and basketball's perimeter jumpers are efficient, yet ratings are down.

• Hiring analytics as a label rather than a capability. DePodesta, a Moneyball hire, became the symbol of analytics failing at a big-budget club, while Epstein with the same ideas and money won. The tell: the analyst is hired but has no budget, authority, or decision to inform.
• Optimizing a single metric until the game stops being watchable. The tell: efficiency rises while extra-base hits and acrobatic plays disappear and television ratings fall.
• Treating possession-style statistics as proof of performance. Liverpool held 80.6 percent and still lost 2–0. The tell: a headline metric is high and results do not follow.
• Applying a model rigidly at a moment where the real-time evidence is strongest. The tell: a decision-maker removes a pitcher having the best game of his life because the algorithm says so, and the team loses the lead.
• Confining analytics to the field. Henry was held back by MLB commercial restrictions, while at Anfield there were flat ticket prices and no café or rentable rooms. The tell: the analytics group never touches pricing, venue, or media revenue.

• ▸Decision to inform: one yes/no or choose-between sentence a manager owns.
• ▸Data available: list sources, resolution (e.g., frames per second), and which leagues/teams/units they cover.
• ▸Data gaps: name where the data does not exist and what you will infer from covered cases.
• ▸Model owner: name the analyst and the skill (e.g., physics, modeling) that makes the model possible.
• ▸Benchmark cost: price of the analyst/group versus the marginal player or asset it informs.
• ▸Output metric: the single number or map the model produces, e.g., control of space or ticket price tiers.
• ▸Reality check: the on-field or in-market result that would make you distrust the metric.
• ▸Override rule: the pre-agreed condition for departing from the model, and who decides.
• ▸Product-quality cost: how the optimization could hurt fan experience, ratings, or atmosphere, and the signal that flags it.

1. 1Name the objective and the loss you will tolerate before you recruit anyone. Grousbeck's stated goal was 'to raise more championship banners,' and he was 'willing to lose money in the service of that next championship.' Write down the goal and the maximum annual loss the group can absorb. Every later choice gets checked against these two lines.
2. 2Spread the risk across a partner roster instead of carrying it alone. Grousbeck brought in his father Irv and Bain Capital's Steve Pagliuca, then 'corralled more than a dozen additional investors.' His logic was 'the goal was to spread the losses around, so nobody had to lose too much.' Size the group so no single partner's loss is painful, and so a bad year does not force a sale.
3. 3Choose partners for what they add besides cash. Pagliuca came from a private-equity firm, and the source treats his background as part of the team's strength. For each partner, write one line stating the skill or network they bring that the others lack. Drop any partner whose only contribution is money, unless you need the money.
4. 4Hire outside advisors early and listen to the junior ones. Grousbeck and Pagliuca hired the strategy firm EY-Parthenon to help close the Celtics purchase. One analyst on that project was Daryl Morey, 29, who called himself 'the low, low-level guy' and later became the Rockets' general manager. Give outside analysts direct access to decision-makers, and judge the analysis on its content, not the analyst's rank.
5. 5Import a method from another industry and apply it to both the field and the business side. Henry decided by gathering as much information as he could and analyzing it 'until it revealed its secrets.' He applied this to the Red Sox on the field and then to the business, including Fenway-area real estate lots. Cuban had collected data on his players since buying the Mavericks and used it to examine lineup combinations and coaching influence. Pick one such method and name the first decision it will govern.
6. 6Audit the revenue side as an operating business would. For Liverpool, Henry's list was a multi-tiered ticket structure instead of one price for every seat, a café for pre- and postgame dining, a rentable conference room overlooking the field, and unrestricted use of the club's brand (a pub in Boston's Back Bay or in Shanghai). Walk your own venue and brand and list the same kinds of gaps. Then check each against league rules, since baseball's rules blocked Henry's Red Sox pubs.
7. 7Let the analysis veto an attractive deal. Henry's group rejected some investments, 'such as the NBA teams, that on first consideration had seemed ideal.' Set an explicit rule that a purchase must pass the evidence test and not merely appeal to the partners. Record who voted and why, so the group's judgment can be reviewed later.
8. 8Set the exit path. Grousbeck knew that 'as long as the team stayed solvent,' investors could 'cash out, probably at a healthy profit.' Tell every partner in advance what solvency means for the club and when they can sell, so a group of many investors does not turn into a group of people who feel trapped.

The Celtics purchase in 2002 is the clearest documented case in the passages. (They do not describe Lacob's venture-capital deal directly, so the Celtics and Henry's group stand in as the documented evidence for the model.) Grousbeck's group paid $350 million for a team whose 'prime asset was history,' which was 16 championships. The going rate for an NBA team was $280 million, the amount Cuban paid for the Mavericks in 2000. Grousbeck admits 'we were overpaying.' He solved the risk problem with structure: his father Irv and Bain Capital's Steve Pagliuca anchored the group, and more than a dozen additional investors joined so 'nobody had to lose too much.' He and Pagliuca also hired EY-Parthenon, where the 29-year-old analyst Daryl Morey worked on the deal. Morey was a computer-science and MIT Sloan graduate, junior in the room, and within five years was running the Rockets. The result is a team bought at a premium, with the downside shared and outside analytical talent in the process. Compare Bob Tisch, who treated his $75 million as money he would 'pretend I never had,' or Donald Sterling, who 'didn't run his team like a real business' because he 'didn't need to.'

• If a single owner would carry a loss that hurts, then add partners until no one partner's share of the downside is painful. Grousbeck overpaid and made it viable by spreading the losses.
• If a candidate partner brings only capital, then ask whether you need the capital badly enough to give up a seat. A slot is better spent on someone like Pagliuca, who adds a skill the others lack.
• If an outside advisor's finding contradicts the group's instinct, then test it with data before overruling it. Cuban's lineup and coaching analysis and Morey's shift toward three-point shooting are examples of outsider methods that paid off.
• If a deal looks ideal on first consideration, then run it through the same evidence test as every other deal. Henry's group rejected the NBA teams after analysis.
• If a revenue idea is common practice in other businesses (tiered pricing, hospitality, brand extension), then check league rules before costing it. Selig's veto of Red Sox pubs shows where the ceiling can sit.

• Running the team as a hobby with a pile of money. Tell: the owner says he doesn't care if he makes a cent, as Tisch did, or sees ownership as buying a boat, as Vataha put it.
• Collecting investors without defining roles. Tell: the partner list is long but nobody can say what each person contributes beyond the check.
• Hiring advisors and then ignoring them, or filtering them by rank. Tell: the outside analysis appears in a deck but no decision changes. Morey was the 'low, low-level guy,' and the model works only if that kind of person is heard.
• Applying data to the field but not to the business. Tell: the club has an analytics group yet sells every ticket at one price and has no hospitality or brand plan. Henry noted that 'such dispassionate assessments were rarely applied to the business side of sports.'
• Spreading the losses but not the information. Tell: partners learn about solvency problems after the fact, so the cash-out path in the plan stops being credible.

• ▸Objective: state the one outcome the group is buying (e.g., more championship banners) in one sentence.
• ▸Maximum tolerable loss: the annual loss the whole group can absorb and the share each partner carries.
• ▸Partner roster: each partner, their stake, and the one skill or network they add besides money.
• ▸Outside advisors: firms or individuals hired, the question each answers, and who receives their findings.
• ▸Imported method: the practice from another industry (e.g., data analysis, tiered pricing) and the first decision it governs.
• ▸Revenue audit: three untapped revenue sources at your venue or brand, each with the league rule that may restrict it.
• ▸Veto rule: the evidence a proposed investment must meet, and what happens when it fails despite partner enthusiasm.
• ▸Exit path: when and how a partner can cash out, and the solvency condition that makes that possible.
• ▸Review date: when the group rechecks all of the above against actual results.

1. 1Set the ownership frame before you look at a price. Write one sentence stating whether this is an investment or a hobby. Bob Tisch's line, "I don't care if I never make a cent on the investment... I'm just going to pretend I never had that $75 million," is the frame you reject. If the sentence can't name a return, you are not yet doing investment logic.
2. 2Establish the asset's appreciation case from long-run comparables, not from fandom. Randy Vataha's test: of Boston's blue-chip companies, Wang Laboratories included, nearly all disappeared, while the Red Sox, Patriots, Celtics and Bruins gained the most value over a quarter century. Ask what comparable franchises did to their owners' capital, and what the alternatives did.
3. 3Price the scarcity. Count the supply: expansion took the four major leagues from 43 teams in 1960 to 120 in 2001 and then stopped, and fewer than 100 people held majority stakes among roughly 500 US billionaires. Treat a franchise as oceanfront real estate: fixed supply, so entry price reflects access as well as cash flow. State how much of the price you are paying for access (Ballmer paid $2 billion for the second-tier Clippers) and decide whether that premium is justified.
4. 4Find the gap between perceived and actual value, which is the arbitrage logic the source attributes to Moneyball. Ask what the market, the fans and rival executives believe, then what the data says. Gelman's Canadian fans are the model: thousands who traveled hundreds of miles for home games and whom nobody had contacted. List the unrealized revenue or performance sources the current owner ignores.
5. 5Build the buying group around the specific expertise the thesis needs, and size the group to the price. Franchise prices make solo purchases rare, so owners combine in groups of three to five. Henry's group included Tom Werner (television producer) and two former Fidelity fund managers. Lacob, who could have bought alone, partnered with a film producer and YouTube and Zappos-era founders, each expert in one field. Assign each partner to a part of the thesis.
6. 6Define the business plan with two goals: win on the field and build a business that validates the price. The 2003 Red Sox owners set out to do both. Write a measurable target for each, and commit to a data-driven method for pursuing them.
7. 7Install a decision pipeline that turns analysis into action. The Callaway quote applies: good information is worthless if the people executing it don't buy in. Hire people (managers, coaches, executives) who will implement the analytics, and replace those who won't, as five clubs did with their managers.
8. 8Audit for second-order damage. Optimized baseball produced more strikeouts and home runs, fewer extra-base hits and acrobatic plays, and falling television ratings. Before scaling any optimization, check what it does to the product fans pay for, and write down which asset (the game's appeal) you may be depleting.

In 2003, John Henry, a commodities trader, and his partners paid $700 million for the Red Sox and some real estate. That was more than double the price ever paid for an MLB franchise, for a club that hadn't won a championship since 1918. A hobby owner would call the price absurd or romantic. An investor reads it differently. The group was built like a portfolio team, with Tom Werner plus Jeff Vinik and Mike Gordon, former Fidelity fund managers. Their stated aim was a World Series winner and a business that would validate the investment. The scarcity case supported the premium: a closed club of fewer than 100 majority owners, and Boston's four sports franchises as the area's only sure appreciating equities. The result the source reports is four World Series wins between 2004 and 2017 and a data-driven methodology that grew into a sports empire. The passages do not show Henry's actual valuation model, so the replicable lesson is the structure: an evidence-based case for the premium, a team of financial experts, and two measurable goals.

• If you can only justify the price by what owning the team feels like (practice on a Tuesday afternoon, prestige, access), then call it consumption, price it as such, and don't call it an investment.
• If the asking price is far above prior records, as the $700 million Red Sox price was, then require a written, data-based case for why the asset's value is understated before you bid.
• If your thesis depends on skills you lack (video, content, finance, data), then add partners who hold them. Franchise prices are already pushing buyers toward groups.
• If your people won't act on the analytics, then replace them rather than diluting the model. Clubs fired managers criticized for not implementing analytics recommendations.
• If an optimization raises efficiency but lowers what fans want to watch, then treat it as a cost against the asset and test it before extending it.

• Treating the team as a trophy: the tell is a return expectation you can't state, or language like Tisch's "pretend I never had that money."
• Paying only for scarcity: the tell is a bid justified entirely by 'there are only so many of them' with no operating plan to make the asset earn.
• Analysis without execution: the tell is an analytics department producing recommendations that managers and coaches routinely override by instinct.
• Copying the method without studying it: the tell is dismissing a rival's working approach unread, as rival GMs did with Moneyball, instead of studying and imitating what works as a real business would.
• Optimizing the metric and eroding the product: the tell is rising efficiency alongside falling ratings, fewer exciting plays, and fans noticing.

• ▸Ownership frame: one sentence stating the expected financial return, or admit it is consumption.
• ▸Price and benchmark: the price versus the highest prior comparable (the Red Sox deal was more than double the previous record).
• ▸Scarcity premium: what share of the price buys access or limited supply rather than cash flow, and why it is justified.
• ▸Value gap: what the market misperceives, and the data that shows it (the Canadian-fans type of unreached opportunity).
• ▸Partner map: each partner and the specific expertise they bring to the thesis.
• ▸Dual goals: a measurable on-field target and a measurable business target that validates the price.
• ▸Execution check: who turns analytics into decisions, and what happens if they won't act on them.
• ▸Product-risk audit: how optimization could reduce what fans want to see, and the metric that would show it (ratings, attendance).
• ▸Exit or review trigger: the evidence that would make you reconsider the investment.

1. 1Trace the causal chain before you act: fragmented non-sports ratings, live sports holding its audience because outcomes are unscripted, rights fees rising, franchise values rising. Write it as four links and confirm your league or club sits on it.
2. 2Put the price history in one table for the franchise or league you are studying, with years, buyers, and prices. The source's NBA chain runs $4.5M (Pacers, 1983), $75M (Spurs, 1993), $401M (Suns, ~2003), $2B (Clippers, 2014). Mark each buyer's prior business, since the buyer profile shifts alongside price.
3. 3Normalize old figures to current dollars before comparing. NFL network fees of $900M in 1990 are about $2B in 2022 dollars, and the 2002 deals of $14B are about $23B. The real growth is the gap between the nominal and normalized figures, not the nominal rise alone.
4. 4Test for scarcity separately from media money. The source counts teams growing from 43 in 1960 to 120 in 2001, then stopping, with fewer than 100 majority owners against about 500 US billionaires. Check whether supply is fixed, and whether each sale reprices every other franchise.
5. 5Re-derive the entry ticket in percentage terms. Beane's figure: $60M bought about 5% in 2016 and about 2% by 2022. Compute what a given check buys today, and the minimum price of a meaningful stake, such as at least $250M for 5% of the Warriors.
6. 6Redefine your audience. At $1B-plus valuations, the pool of hobby-driven buyers narrows to the richest tech billionaires and sovereign or oligarch money. Everyone else is a rational investor who compares the team to biotech or software and looks at it 'with a squinty eye.'
7. 7Convert expectations into operating requirements: a growth narrative, a voice for minority investors, data-driven decisions, and off-field business discipline. Write each as a measurable commitment you can show an investor.
8. 8Pressure-test the hobby exception. Ask whether the owner is insulated by shared revenue, as Sterling was, or exposed. If the owner is insulated, expect weaker discipline despite the macro pressure.

Use the Boston Red Sox sale in 2003. John Henry and his partners paid $700M for the team plus some real estate, more than double any price previously paid for an MLB franchise. Compare that to Bob Tisch, who paid $75M for half the Giants and said, 'I don't care if I never make a cent on the investment.' Tisch treated the purchase as a hobby because the price was small against his wealth. At $700M, the buyers had to build a World Series winner and 'a business that would validate their huge investment.' The Red Sox then won four World Series between 2004 and 2017. A second case is the Warriors. Joe Lacob had sold investors on growth, gave all of them a voice, and Stevens, who put in about $50M, said the team 'is being run like a real business.' The pattern is that price level sets the owner type, and owner type sets management standards.

• If a franchise price is large relative to the buyer's portfolio, then evaluate it with investment rigor and expect the owner to demand business discipline, not hobby logic.
• If a minority investor writes a $50M–$60M check, then give them a growth case and a channel to be heard, because they treat it as an equity.
• If supply is fixed and each sale resets valuations league-wide, then treat appreciation as structural and cheap entry as gone, but do not assume any single team's operations are healthy.
• If valuations exceed $1B, then expect the buyer pool to narrow to sovereign-scale, oligarch, and top-tier tech wealth plus consortiums formed to limit risk.
• If revenue is pooled and the owner is insulated, as with Sterling, then do not read price appreciation as evidence of good management.

• Comparing nominal prices across decades. The tell: you call the NFL's $900M to $14B jump larger than it was, ignoring the 2022-dollar figures of $2B and $23B.
• Crediting rising value to management skill. The tell: you cannot separate the scarcity and media-rights lift from the operator's contribution, as with the Clippers, which were terrible for decades while the franchise's value rose.
• Using the hobby-owner playbook with investor-owners. The tell: minority owners ask whether they are heard, as Lacob worried about at the Celtics, and you have no answer.
• Assuming every buyer shares a passion for the sport. The tell: you build pitches around fandom, while Stevens saw the Warriors as a sound place to park money.
• Treating the expansion-era population as current. The tell: you assume local car dealers or attorneys are plausible buyers when entry costs $250M for 5% of a franchise.

• ▸Franchise and league: name the team and league being analyzed.
• ▸Price history: list year, buyer, and price for at least three sales.
• ▸Real-dollar adjustment: restate each price and any rights-fee figures in one common year's dollars.
• ▸Rights-fee driver: note the media-rights revenue trend and whether live-event demand held.
• ▸Scarcity test: count teams, note when expansion stopped, and say whether sales reprice other franchises.
• ▸Buyer profile: describe who bought at each price point and their prior business.
• ▸Entry ticket: state what a given investment buys in percentage terms today versus earlier.
• ▸Owner insulation: say whether revenue sharing protects the owner from poor performance.
• ▸Required operating standard: list the growth, voice, and data commitments investors now expect.

1. 1Sort every data source you could use into three tiers: public (box scores, the on-base percentage that Beane's analysts crunched), league-owned and distributed raw to every team (Statcast, delivered nightly), and not yet available (NHL puck chips, NFL uniform sensors). Your edge differs by tier. Public data gives an edge only to whoever moves first, because rivals copy it once it works. Raw league feeds give an edge to whoever builds the pipeline.
2. 2For any league-owned feed, build the ingestion mechanism first. The passage describes teams that download the raw server data each night and convert it into insights, and others that simply ignore it. Assign an owner, a nightly schedule, and a defined output, such as a metric the front office can act on.
3. 3Point the new data at a question the old data could not answer. Statcast made defensive skill measurable for the first time: balls reached versus the average shortstop, speed versus jump versus positioning. Write the question down before you touch the data, and note which old stat was blind to it.
4. 4Convert the measurement gap into a valuation gap. A team measuring defense can value an average hitter who fields superbly above a team that does not. Compare your internal valuation with the market's on the same player and list where the two diverge.
5. 5Audit coverage before you generalize. Not every European league has the tracking cameras, and most of the young talent Liverpool scouts plays where tracking doesn't exist and may never exist. Check that the leagues feeding your model have the capability, and use them to find patterns inherent to the sport rather than to rate individual players.
6. 6Staff for the problem's structure, not for sport experience. Liverpool hired a particle physicist who knew nothing about soccer. He modeled player control of space by treating speed as charge, and he built a working model in a weekend. Pair a domain expert with someone who can reshape raw tracking data into a new model.
7. 7Record what the data still can't see. Graham called the play-by-play data 'a very foggy lens'. Passes, shots, and ball movements were all his model could evaluate, so off-ball actions were invisible until video tracking scored them. Keep a list of unmeasured actions and rank it by how likely each is to matter.
8. 8Watch for second-order effects once the whole league optimizes. Analytically tuned baseball has more strikeouts and home runs and fewer extra-base hits and acrobatic plays, which research suggests are the plays fans like most. Check your edge against the product, not just the win column.

Liverpool's analytics group, led by Ian Graham, had tracking data on the trajectory of every player at 25 frames per second, but the zone-control models it could build were simplistic. In 2016 Will Spearman, a particle physicist who knew nothing about soccer, presented at London's OptaPro conference. He treated each player's speed as a 'charge' and spent a weekend reshaping the data into a model that located where each player controlled space at any moment. Michael Waskett called Graham and said, 'We have to hire this guy… take him off the market.' Graham's reasoning was that the data now made it possible to ask whether the team should play 'in a slightly different way.' The limits were explicit: only some leagues have the cameras, and the play-by-play remained a foggy lens. Spearman's video-tracking model went after what that lens missed, scoring off-ball moves such as a fullback forcing a defender to choose or a striker positioning for a cross. The passages credit these tactical insights and player recommendations with helping the club to the 2019 Champions League title and, the next season, its first Premier League title.

• If the data is public and cheap, expect any edge to vanish once a winner emerges (the A's, then nearly every team). Move fast, or look for proprietary data instead.
• If a league distributes raw data to every team, the advantage lies in your pipeline and interpretation, not access. Build the nightly conversion before debating what to buy.
• If the data covers only some leagues or venues, use it to find sport-wide patterns, not to grade players from uncovered leagues.
• If a metric newly makes a skill measurable (defense via Statcast), price that skill against rivals who don't measure it and target the gap.
• If the model only sees events the official feed records, treat its conclusions as provisional and fund work on off-ball or unrecorded actions.

• Ignoring the feed: your team receives the raw data nightly but nobody ingests it. The tell is that no one can name the person or process that handles it.
• Treating a public statistic as a lasting edge: you build strategy on box-score data. The tell is that rivals now run the same analysis and your advantage has shrunk.
• Over-generalizing from partial coverage: you apply tracking-based conclusions to players from leagues with no cameras. The tell is that your sample comes mostly from a few well-equipped leagues.
• Mistaking the foggy lens for the whole game: you judge only passes, shots, and ball movements. The tell is that off-ball contributions like positioning and forcing decisions never appear in your ratings.
• Optimizing the metric and degrading the product: everyone chases the analytically best play. The tell is rising strikeouts and home runs, fewer extra-base hits and fielding plays, and a ball rarely in play.

• ▸Data source: name the feed or technology (e.g., Statcast, soccer tracking cameras, puck chips).
• ▸Tier: public, league-owned and distributed to all teams, or not yet available.
• ▸Coverage: which leagues, venues, or players it actually captures, and which it misses.
• ▸Ingestion owner and schedule: who pulls and converts it, and how often.
• ▸Question it answers that old data couldn't: one sentence, naming the blind stat it replaces.
• ▸Valuation gap: where your estimate of a player or tactic differs from the market's.
• ▸Unseen actions: what the data still cannot record (the 'fog').
• ▸Rival status: which competitors ignore it, use it, or are building it.
• ▸Product side effect: how acting on this data could change the game's appeal.