Sport & Performance

Skilled performance degrades under high stakes, in part because attention turns to the execution of an automated skill.

The explicit monitoring account holds that a practised skill runs better without supervision, and pressure invites the performer to supervise it. Home crowds can make it worse rather than better in decisive contests, which is one of the few conditions where home advantage reverses.

In practice

• A putter who has made the stroke ten thousand times thinks about the stroke and misses.
• A presenter who has given the talk twenty times loses the thread on the one occasion it counts.

Attributed to

Roy Baumeister and later work

Also called

Paralysis by analysis; performance under pressure

Page for this entry

Teams win more often at home than away, across essentially every sport and league measured.

One of the most robust findings in sport, and its cause is not settled: travel, familiarity, crowd, and referee bias have all been proposed. Matches played behind closed doors during 2020 offered a natural experiment, and home advantage shrank markedly, which points at the crowd and at what the crowd does to officials.

In practice

• A league that publishes home and away splits shows the gap in every season it has records for.
• A season played in empty stadiums narrows the gap, which is evidence about mechanism rather than existence.

Attributed to

Sports statistics; studied systematically from the 1970s

Also called

Home field advantage; home ground advantage

Further reading

Wikipedia

Page for this entry

Calling a timeout immediately before a kick is held to disrupt the kicker's success rate.

The coach who ices and loses has done the accepted thing; the coach who declines and loses has to explain. That asymmetry keeps the tactic in use whatever the evidence says, which is a general feature of low-cost superstitions inside professional practice.

In practice

• A manager makes the conventional substitution because the unconventional one that fails will be the story.
• A practice with no measured benefit and no cost survives indefinitely, since nothing selects against it.

Attributed to

American football coaching practice

Also called

Freezing the kicker

Possible explanations

One explanation is disruption of a practised routine and added time to dwell on the attempt. Another is that the tactic survives for reasons of blame rather than effect, since a coach who does the accepted thing and loses has less to explain.

In its favour

Some studies find a small effect on long or high-pressure attempts, which is where the tactic is used.

Against it

Others find no effect at all, the samples are small, and icing is chosen in exactly the situations where kicks are hardest, so the apparent effect may be selection.

Standing

Studies disagree: some find a small effect on long or high-pressure attempts, others none. The practice is near-universal regardless, which makes it a good example of a tactic sustained by the cost asymmetry of not trying it.

Further reading

Wikipedia

Page for this entry

Success within a contest is held to raise the probability of further success beyond what the underlying skill implies.

Commentators identify momentum from streaks, and streaks occur in random sequences at rates people badly underestimate. Where careful work has looked, the measured effects have been small. The concept survives because it is applied after the fact, when the sequence is known.

In practice

• A team that scores twice is said to have momentum, and the same two goals in a loss are called a consolation.
• A run of four wins in twenty is unremarkable arithmetic and is narrated as a turning point.

Attributed to

Sports commentary and psychology

Also called

The big mo; psychological momentum

Possible explanations

One explanation is psychological, that confidence and attention rise after success. Another is that streaks of the length people notice occur in random sequences, so the pattern is inferred from none.

In its favour

The psychological mechanism is plausible and there is some evidence for small effects in individual sports where opponents cannot adjust.

Against it

Studies find weak and inconsistent evidence, and much of what is called momentum is the clustering that random sequences produce, which commentators reliably underestimate.

Standing

Studies find weak and inconsistent evidence, and attribute much of what is called momentum to the clustering that random sequences produce. Related to the hot-hand debate, where the balance of evidence has shifted somewhat toward a small real effect.

Page for this entry

An exploitable mispricing disappears once the method for finding it becomes public.

Named for Lewis's 2003 book. Once on-base percentage was public, its price rose and the Athletics' advantage disappeared — a concrete case of the Grossman-Stiglitz paradox and a cousin of Goodhart's law.

In practice

• A recruitment firm's edge in sourcing from one platform disappears the year its competitors read the same case study.
• A trading strategy published in a well-read paper stops working within about eighteen months of publication.

Attributed to

Michael Lewis

Standing

A concrete case of the Grossman–Stiglitz paradox and a cousin of Goodhart's law.

Page for this entry

Performance rises to a peak in the mid to late twenties for most sports and declines thereafter, at a rate the sport determines.

Aggregate curves are built from players still playing, and the ones who fell away are absent from the later points. Corrections for this move the peak earlier and steepen the decline. Sprinting, endurance and skill-dominant sports peak at different ages, so a single curve describes none of them.

In practice

• A club paying for a thirty-year-old's last three seasons is buying the flattest part of a biased curve.
• A cohort measured only where it survives will show every trait declining more gently than it does.

Attributed to

Sports science and performance analysis

Also called

Ageing curve

Possible explanations

The explanation is the interaction of physical capacity, which declines from the late twenties, and skill, which accumulates, so the peak falls where the two cross, later in sports where judgement matters most.

In its favour

Peaks do differ systematically by sport in the direction the account predicts, and longitudinal data on individual athletes show the rise and decline clearly.

Against it

Aggregate curves are biased by survivorship, since players who decline early leave the sample, which flattens the measured decline, and corrections move the peak earlier and steepen the fall.

Standing

Estimated peaks are biased by survivorship: players who decline early leave the sample, which flattens the measured decline. Peak ages also differ substantially by sport, position and the demand being measured.

Page for this entry

Officials award marginal decisions to the home side, and do so more where the crowd is larger and closer.

Studies of added time, penalties and marginal fouls find home-favouring patterns that track crowd size and stadium layout, including natural experiments where matches were played without spectators or behind athletics tracks. It is the best-supported single mechanism behind home advantage, and it does not require any official to be aware of it.

In practice

• Added time is longer when the home side is behind and shorter when it leads, by a margin visible across a season.
• A panel that meets in one party's offices decides marginal points in that party's favour without anyone intending it.

Attributed to

Sports economics and officiating research

Also called

Social pressure on officials

Page for this entry

Children born just after a selection cut-off are heavily over-represented in elite squads, because early physical maturity is mistaken for talent.

The Canadian junior hockey study found birth months clustered immediately after the January cut-off. The mechanism is compounding rather than physical: the older child in the age band is selected, receives better coaching and more play, and the initial advantage of a few months' growth becomes a real one. The same skew appears in football, cricket and schooling.

In practice

• A squad list whose birthdays cluster in one quarter has told you about its selection process, not its talent pool.
• A school setting pupils by attainment at seven has partly sorted them by month of birth.

Attributed to

Barnsley, Thompson & Barnsley

Also called

Birth-month bias; the selection cut-off effect

Further reading

Wikipedia

Page for this entry

Results improve after a manager is dismissed, and improve about as much for comparable clubs that keep theirs.

Managers are dismissed after an unusually bad run, which is when results were furthest from the club's mean, so improvement was the expected next observation whoever was in charge. Studies matching sacking clubs against clubs in similar slumps that did not sack find little difference. It is regression to the mean with a press conference.

In practice

• A club wins three after a dismissal and the board learns the wrong lesson at some cost.
• A chief executive replaced after a poor quarter presides over a recovery that was already due.

Attributed to

Sports management research

Also called

The new manager bounce; managerial change effect

Page for this entry

A rating updated by the gap between actual and expected results, so a win against a stronger opponent is worth more than one against a weaker.

Devised by Elo for chess and adopted by the US federation around 1960 and internationally in 1970, the system converts a rating difference into an expected score and moves both players by the surprise. It has since been applied to football, esports and matchmaking, and its logic is Bayesian in effect if not in derivation.

In practice

• A player who beats a much stronger opponent gains far more than for a routine win, and the system needs no judgement to decide that.
• A matchmaking system that pairs equal ratings produces close games, which is the property it was not designed for.

Attributed to

Arpad Elo

Also called

Elo rating; chess rating system

Further reading

Wikipedia

Page for this entry

As skill rises and becomes more uniform among competitors, luck explains more of the difference between their results.

Mauboussin set it out in 2012, following Gould's account of the disappearance of the .400 hitter in baseball. Batting averages did not fall; the variation among players fell as training and selection improved, so extreme averages stopped occurring. When everyone in a field is excellent, what remains to separate them is what nobody controls.

In practice

• A fund manager beating the market by a wide margin is easier to find among amateurs than among professionals.
• An Olympic final is decided by hundredths of a second that the same two athletes would reverse the following week.

Attributed to

Michael Mauboussin; after Stephen Jay Gould

Standing

Follows from the arithmetic of variance rather than from a study of skill, which is what makes it general. Compare regression to the mean and survivorship bias.

Page for this entry

A team's expected win rate can be estimated from runs or points scored and conceded, better than from its actual record.

James's formula estimates wins from scoring ratios and predicts next season's record better than this season's record does, because the difference between the two is mostly luck in close games. It is a regression-to-the-mean instrument disguised as a scoring model, and it transferred from baseball to basketball and football with adjusted exponents.

In practice

• A team that outperformed its scoring difference is expected to decline, and usually does.
• A manager sacked after an unlucky season is replaced by one who inherits the correction.

Attributed to

Bill James

Also called

Pythagorean win percentage

Further reading

Wikipedia

Page for this entry

Where talent is imperfectly substitutable and reproduction is cheap, small differences in ability produce enormous differences in reward.

Rosen's 1981 argument: nobody wants two mediocre performances instead of one excellent one, and technology lets the best serve everyone at once. The result is a market where the hundredth-best earns a small fraction of the best, though the difference in ability is slight. It explains sport, music and increasingly any winner-take-most market.

In practice

• A player marginally better than the next is paid several times more, and the club is being rational.
• A field where recordings replaced live performance concentrated income within one generation.

Attributed to

Sherwin Rosen

Also called

The economics of superstars

Page for this entry

Talent ends up at the clubs that value it most, whichever way the rules assign the initial rights to a player.

Rottenberg argued in the Journal of Political Economy in 1956 that baseball's reserve clause distributed players among teams about as a free market would, because a club that values a player less will sell him to one that values him more. The general result is now known by Coase's name, and sports economists have objected to the demotion for fifty years.

In practice

• A draft intended to spread talent is followed by trades that undo it.
• A transfer restriction changes who is paid rather than who plays.

Attributed to

Simon Rottenberg

Also called

The invariance principle

Possible explanations

The explanation is that a club values a player by what he adds to its revenue, so whoever holds the initial right, the club or the player, the player ends up where that value is highest, with the rule determining who is paid rather than where talent goes.

In its favour

It anticipated the Coase theorem by four years, and the broad pattern after free agency, with talent distribution changing less than predicted, is consistent with it.

Against it

Transaction costs, payroll rules, and the fact that clubs differ in their ability to borrow all break the equivalence, and later work finds that the assignment of rights does affect where players end up.

Standing

Stated four years before Coase's 1960 paper and often absorbed into it, though the two differ: under Rottenberg the parties to the bargain change with the rule, and under Coase they do not. Empirical support is mixed.

Page for this entry

A draft that rewards the worst teams gives the worst teams a reason to lose.

The draft is meant to restore competitive balance by giving weak clubs first choice of new talent. The incentive it creates near the end of a season is to be as weak as possible, which is why leagues have layered lotteries and weighting on top, each of which creates a further incentive to manage. It is a clean case of a rule producing the behaviour it was designed to prevent.

In practice

• A club rests its best players in March and calls it squad management.
• A grant scheme that funds the worst-performing schools rewards not improving before the assessment.

Attributed to

Sports league design

Also called

Losing for draft position

Further reading

Wikipedia

Page for this entry

Spectators are held to value close contests, so a league's revenue depends on competitive balance.

Stated in Rottenberg's 1956 paper and developed by Neale in 1964, it underpins the standard defence of drafts, salary caps and revenue sharing. Testing it has proved difficult and the results are mixed, which matters, because the policies it justifies are substantial restraints on a labour market.

In practice

• A league imposes a salary cap to protect a demand relationship it has not measured.
• A dominant team fills stadiums for a decade, which the hypothesis does not predict.

Attributed to

Simon Rottenberg; developed by Walter Neale

Also called

Competitive balance hypothesis

Possible explanations

The explanation is demand: if spectators value contests whose result is in doubt, a league's revenue depends on no club being able to dominate, which is the standard justification for drafts, caps and revenue sharing.

In its favour

The reasoning is clear, it underpins the competitive design of most professional leagues, and some studies do find attendance responding to closeness of the contest.

Against it

The empirical support is weak and inconsistent, with many studies finding attendance and viewing driven more by star quality, stakes and habit than by closeness, and dominant teams have repeatedly coincided with record revenue.

Standing

The empirical support is weak and inconsistent: many studies find attendance and viewing driven more by star quality, stakes and habit than by closeness, and dominant teams often draw best.

Page for this entry

Many small improvements across every component are held to compound into a decisive advantage.

Brailsford's account attributed British Cycling's success to hundreds of one-percent improvements, from pillow choice to hand-washing. The programme also received a large increase in lottery funding and recruited exceptionally. The doctrine has been repeated in business almost entirely on the strength of the sporting result.

In practice

• A team that improves fourteen things by 1% has made a real difference or none, and the account cannot tell you which.
• A firm that copies the method without the budget that accompanied it has copied the visible half.

Attributed to

Associated with Dave Brailsford at British Cycling

Also called

Aggregation of marginal gains; the 1% rule

Possible explanations

The explanation offered is compounding across components: many one-percent improvements multiply into a decisive margin, provided each is real and retained.

In its favour

Attention to overlooked components is genuinely useful, and the approach produced measurable improvements in equipment, recovery and illness rates at British Cycling.

Against it

It is a framing rather than a measured finding: the same results are equally consistent with a large funding increase, talent and the marginal-gains programme together, and the arithmetic assumes gains are independent and additive, which they usually are not.

Standing

A framing rather than a measured finding. British Cycling's results are consistent with it and equally consistent with funding, talent and coaching, and the doctrine has since been applied far beyond any evidence for it.

Page for this entry

A team stays a contender until the local press believes it is one, at which point the season collapses.

The law is stated as though the press caused the fall. Regression to the mean produces the same pattern with no causal arrow at all: the media become convinced at the point of maximum unusual success, which is precisely the point from which the next stretch is most likely to be worse. Berg's version is memorable, and the mechanism it implies is the wrong one.

In practice

• A division promoted as the company's success story misses its next two quarters, and the promotion did not cause it.
• A fund profiled after its best year is profiled at the moment its returns are most likely to fall.

Attributed to

Mitch Berg

Also called

Berg's Fourth Law; the contender jinx

Possible explanations

One explanation is the one the law asserts, that coverage changes a team's behaviour. Another, which needs no causation, is regression to the mean: the press becomes convinced at the peak, and the peak is the point from which performance most often falls.

In its favour

The pattern is recognisable, and the timing it describes, acclaim arriving just before collapse, is real enough to have produced the same folklore in several sports.

Against it

It is a jinx claim with a mundane alternative available, and nobody has compared the fortunes of teams the press believed in against those it did not.

Standing

From Berg's numbered set, first codified on his blog in 2005. A jinx claim with a mundane explanation available: coverage peaks when performance peaks, and performance then regresses toward the mean, so the belief and the collapse are both effects of the same run rather than cause and effect.

Page for this entry