Comparability breaks

There is no chart here, and the reason is different from every other absence on this site. A weightlifting world record is a weight lifted in a bodyweight category. The weight is a real, comparable number — kilograms are kilograms. What has not held still is the category. The classes were completely redrawn in 1993 and again in 2018, so there is no bodyweight division that has existed continuously for the sport's modern history, and a record has nowhere to live across either line. Even the heaviest class, which sounds as though it should be continuous, is not: its lower bound has moved repeatedly, so the superheavyweight record of 1990 and of 2020 are records in different competitions that happen to share a name. And before both of those, the sport removed one of its three lifts. Every total recorded before 1973 includes a press. Every total after it does not. Athletics changed an implement and reset one event. Weightlifting changed the definition of who is competing against whom, twice, and reset everything.

Roller speed skating's marks were severed around 1990 by an equipment change the sport made deliberately and completely. Competitive speed skating moved from quad skates to inline within a few years. An inline skate rolls faster and holds a line better, and the change was not a marginal gain — it was a different way of moving. Times before and after describe two sports. Unlike the equipment cases this site records elsewhere, nobody had to decide what to do about it. Athletics redesigned the javelin and reset the book; swimming banned the supersuits and kept the records; finswimming split its books by fin type. Roller speed simply changed over, and the quad-era times stopped being relevant because nobody was skating on quads any more. The other obstacle is that the discipline races on banked tracks and on open roads over distances from 200 metres to a marathon, and a road time is a function of the circuit. Records are kept by surface and distance and are not pooled. Nothing is entered because no series carrying that context has been taken from a cited source.

A set to 15 under side-out scoring and a set to 25 under rally scoring are not the same object. Under the old system a rally won by the receiving team produced no number at all; under the new one every rally produces a point. Points per set, points per match, match duration and the scoreline itself all changed meaning in 1999. This is the third time this exact absence appears on the site, and by now the repetition is the argument. Volleyball in 1999, table tennis in 2001, badminton in 2006 — three federations, three continents' worth of member associations, one stated reason, and not one of them kept a series across its own change. The homepage claims a wave of broadcast-driven scoring rewrites; this empty chart, three times over, is what the wave left behind. A run of match durations from one tournament held continuously across 1999 would be a real series and would show the break plainly. Nobody publishes one.

Fistball severed its own match statistics in 2000 and did not publish either side of the break. Under the old scoring, a side could only score while serving, so a set was a sequence of service exchanges of unpredictable length. Under rally scoring every rally is a point and a set ends at eleven. Counts of points, of rallies, and of anything expressed per set describe two different games before and after. This is the same severance volleyball made in 1999, table tennis in 2001 and badminton in 2006 — four sports, one broadcast pressure, four severed record sets within seven years. Fistball is the smallest of them and the only one where the before-and-after figures were never compiled in the first place. Nothing is entered because nothing has been taken from a cited source.

A game to 21 and a game to 11 are not the same unit, and a match is now best of five or seven rather than best of three. Points per match, rally counts and match duration all changed meaning on 1 September 2001, and none of them can be rescaled across that date. This is the second time this exact absence appears on the site. Badminton has it too, for the same reason and five years later, and the repetition is the finding rather than an oversight: two racket federations shortened their scoring for broadcast within five years of each other and neither kept a series across the change. A run of match durations from one tournament held continuously through 2001 would be a real series and would show the break plainly. Nobody publishes one. Table tennis has a complication badminton does not. The ball changed size in 2000 and material in 2014, and both changes altered rally length on their own. Even a scoped series would be measuring three changes at once.

A game to nine under hand-in hand-out and a game to eleven with a point on every rally are not the same unit. Under the old system a rally won by the receiver produced no number at all; under the new one every rally produces a point. Points per game, rally counts, match duration and the scoreline all changed meaning in 2004. This is the fourth time this absence appears on this site, and by now the pattern is the finding. Volleyball rewrote its scoring in 1999, table tennis in 2001, squash in 2004, badminton in 2006. Four federations, four continents' worth of member associations, one stated reason — a match had to fit a broadcast slot — and not one of them kept a series across its own change. Squash adds a second variable the others do not have. The tin was lowered for professional men's play in 2001, enlarging the target area and shortening rallies, so a rally-length series spanning that date is measuring the court as well as the scoring. A run of match durations from one tournament held continuously across 2004 would be a real series and would show the break plainly. Nobody publishes one.

Rally scoring changed what a point is. A point under the fifteen-point side-out system could only be won by the serving side, so a rally won by the receiver produced no number at all; from 2006 every rally produces a point. Rally counts, points per match, match duration and the scoreline itself all mean different things on either side of May 2006, and none of them can be rescaled from one system to the other. This is a different kind of absence from gymnastics. A series here is not impossible in principle — a run of match durations from one tournament held continuously across 2006 would be a real series, and it would show the break clearly. It is absent because no governing body publishes one, and because assembling it from tournament records is work this site has not done. When it is done, the chart appears here, in two segments, with the 2006 rule change between them.

Kabaddi's scoring changed shape in 2014 and the sport has not published a series across it. Before the raid clock, a raid ended when the cant broke or the raider was caught, so raids varied in length and a cautious team could run a low-scoring match by declining to engage. The thirty-second clock and the do-or-die raid removed both options. Scoring rose, and points per match before and after describe different games. The change also arrived through an unusual route. It was written for a professional television league in India in 2014 and adopted into the international code afterwards — so the severance in kabaddi's statistics was caused by a domestic commercial competition rather than by its federation. That is a first on this site. The figures are not entered because they have not been taken from a cited source. The league publishes detailed statistics for its own matches; the international game's are thinner, and nobody has reconciled the two.

Bowling has the strangest comparability problem on this site, because nothing about its scoring system ever changed. A strike has scored ten plus the next two deliveries for more than a century. The maximum is 300 and always has been. The arithmetic is identical in 1960 and today, which is exactly why the numbers cannot be compared. What changed is underneath. Reactive resin coverstocks from the early 1990s grip the dry lane far harder than the polyester and urethane balls before them, hook into the pocket more aggressively, and carry pins that older equipment left standing. Averages rose and perfect games went from rare to routine — with no change whatsoever to the scale they were measured on. Then the sport responded by changing the surface rather than the equipment. Competition oil patterns are longer, flatter and far less forgiving than the recreational house conditions most bowling is done on, so the same bowler with the same ball scores dramatically differently depending on a coating they cannot see. So a score requires three pieces of context that are almost never recorded together: the era's equipment, the oil pattern, and how far into the block it was applied. Without them the number is arithmetic without a measurement. Nothing is entered here because no series carrying that context has been taken from a cited source.

Sailing does not produce a number that could be charted, and it has arranged itself that way deliberately. There are no times. A result is a finishing position, aggregated across a series of races, with the worst result discarded. That structure exists because the wind is the propulsion: it changes across a course, across a race and across a day, so a single race's order contains a large amount of luck and a time contains no information at all. Rowing has the same problem in weaker form and answered it by timing its races and refusing to call the times records. Sailing went further and stopped producing times. There is a second discontinuity on top of that one, and it is unique here. The Olympic classes — the actual boats — are reselected between cycles. A result in the Finn class and a result in the class that replaced it are not comparable in any sense at all, because the equipment is different and the event no longer exists. It is as though athletics changed which events it contested every four years.

There is no chart here because fencing changed what a hit is three times, one weapon at a time, over fifty-two years. Before electrification a hit existed when a president and four judges agreed they had seen one. Afterwards it exists when a circuit closes. Those are different quantities in exactly the way a taekwondo judge's button and a taekwondo sensor are different quantities — one is an agreement about perception, the other a physical event. What makes fencing unlike taekwondo is that it did this three separate times and produced three separate discontinuities in one sport. Épée in 1936, foil in 1957, sabre in 1989. For twenty-one years the sport ran one measured weapon and two judged ones; for another thirty-two years, two and one. So a 5–3 in épée in 1938 and a 5–3 in sabre in 1938 were not the same kind of result, and neither is comparable with the same score in its own weapon before its own electrification. There is no continuous quantity across any of it.

Slalom scores can be ranked and cannot be charted, because a score is three numbers rather than one. A result is quoted as buoys at a rope length at a speed — "three at 10.75 off at 58 kph". Ordering two such results requires knowing the protocol: more buoys at a shorter rope beats fewer, but a comparison across speeds or divisions needs the whole triple. There is no axis a line could be drawn on without collapsing information the sport considers essential. There is a second problem, and it is the more interesting one. The boat improved. Hull design, wake characteristics and electronic speed control advanced enormously over the decades, and every one of those improvements makes a pass easier. A progression of slalom scores would therefore chart boat development at least as much as skiing. That is a rare shape on this site. Elsewhere the equipment that improved was the athlete's — a javelin, a swimsuit, a bat, a bench shirt, a monofin. Here the decisive equipment is a vehicle the competitor does not own, does not control and does not drive, and its improvement is invisible in the score. Nothing is entered because no series carrying the rope length, the speed and the boat protocol has been taken from a cited source.

A canoe slalom result is a run time plus penalties, on a course that is different at every event and different again depending on where the gates were hung that morning. Nothing about it is comparable. The water flow varies, the gate positions are set for each competition, and artificial courses differ from each other and from natural rivers. A time from one venue means nothing at another, and the federation has never suggested otherwise. Sprint is the opposite case and belongs in the same entry for contrast. A 1000 m sprint on flat water is closer to rowing's problem: a standardised distance, in lanes, with wind and water that nobody controls. Rowing decided its times were world best times rather than records, and canoeing recognises best times in the same way and for the same reason. So this federation runs one discipline whose times cannot be compared because the course changes, and another whose times cannot be compared because the conditions do — and it has handled both by declining to claim they are a series.

Shooting measures more precisely than any sport on this site and has made its own measurements harder to compare than most. Three changes did it. In 1988 finals were added, so a result stopped being a qualification total and became a total plus a final. In 2013 finals were reset to start from zero, so the qualification score stopped counting towards the result at all — a competitor can lead qualification by a wide margin and finish eighth. And in 2013 decimal scoring came into qualification in the air events, so a qualification total is now measured in tenths of a ring where it used to be measured in whole rings. A 1984 result, a 2000 result and a 2020 result are therefore three different quantities: a qualification total, a qualification total plus a final, and a final score from zero after an elimination. Records in these events are kept separately for exactly this reason. What makes it striking is the direction. Archery moved to sets in 2011 and taekwondo to rounds in 2022, both for broadcast, and both accepted that a competitor could out-shoot or outscore an opponent and lose. Shooting did the same thing while simultaneously making its instrument ten times more precise. The measurement improved and its relevance to the result fell.

A surfing score is a judgement about a ride on a wave that will never occur again. Two surfers in the same heat do not get the same waves; two heats on the same day do not have the same sea; and two events in different places are not comparable in any respect at all. This is the most extreme version of the problem rowing and sailing named about themselves. Rowing has an uncontrolled variable and refuses to call its times records. Sailing abandoned times and scores positions. Surfing cannot even do that, because the competitors within a single heat are not attempting the same task. What makes this entry different from every other absence on this site is that surfing is about to be able to fix it. An artificial wave pool produces an identical wave on demand, and Olympic competition moves to one for 2028. Every surfer would ride the same wave, and the sport would become comparable for the first time in its history. A large part of the sport regards that as the end of surfing rather than the perfection of it. This site takes no position and notes only that no other federation in this collection has ever had the option of removing its central variable, and that surfing's is the only comparability problem here with a technological solution.

Amateur boxing has changed what a score is twice, in opposite directions, and neither change was a rescaling. Until 1992 judges scored a bout on their overall assessment. From 1992 a point existed only when at least three of five judges registered the same blow within about a second — an arithmetic, deliberately built to remove discretion after the 1988 judging scandal. From 2016 the sport reverted to the professional ten-point must system, in which judges score each round on their assessment of it. A bout won 12–7 in 2004 and one won 29–28 in 2018 are not two values of one quantity. The first counts registered blows; the second counts rounds awarded on judgement. The sport went from opinion to arithmetic and back to opinion inside twenty-four years. What makes this the strangest entry in this collection is the direction of travel. Every other sport here that automated a judgement kept it automated — taekwondo, archery, tennis, swimming. Boxing automated, disliked what automation did to the sport, and reversed. And then, three years later, an independent investigation found that bouts under the restored human system had been manipulated.

A triathlon time is a sum of three legs plus two transitions, and in 1995 the middle leg stopped being an individual effort. Before that year, drafting was illegal: a rider maintained a gap and the bike split measured what they could do alone. After it, riders sit in packs where the aerodynamic saving is very large, and a bike split measures which pack you were in. Those are different quantities, and no adjustment recovers the difference. Two further things make even the post-1995 numbers uncomparable. Courses are not standardised — every triathlon is a different swim, a different bike course with different hills, and a different run — so a finishing time is a fact about a venue. And the swim can be shortened or cancelled for water quality, which means the same event's results sometimes describe a duathlon. Triathlon is therefore closer to rowing and sailing than to athletics: the conditions cannot be standardised and the numbers are not a series. Unlike rowing, it has never said so.

Chess has the best measurement instrument on this site and it still cannot be charted honestly across a century. The rating is a continuous number, updated by results, calibrated so that a difference predicts an expected score. It orders every rated player in the world without their having to meet. Almost nothing else covered here has an equivalent — most sports can rank the people who competed and no more. The problem is that a rating is relative, and the population it is relative to has changed enormously. Ratings are computed against other rated players, so the scale drifts as the pool grows, as the entry threshold moves, and as improvements in training raise the whole distribution. Whether a rating of 2700 in 1980 describes the same strength as 2700 today is a genuine and unresolved argument inside the sport, and rating inflation has been studied and adjusted for more than once. Engines complicate it further, in both directions. They raised the standard of preparation for everybody, which lifts the pool. And they provide the only absolute yardstick anyone has — comparing historical games against engine evaluation — which is a different measurement from rating and does not reconcile with it. So the number is excellent within its era and drifting between them, and this site does not draw a line through it.

Taolu scores cannot form a series across the introduction of the difficulty component, and the reason is the same one this site records for gymnastics. Before *nandu*, a routine was marked holistically for quality and completeness against a maximum. After it, an athlete declares prescribed difficulty elements — jumps with rotation counts, and their landings — and is scored on execution against that declaration plus an overall performance mark. The number means a different thing on each side of the change, and the maximum available is constructed differently. Gymnastics made the equivalent move in 2006 with the open-ended Code of Points and severed its own history. Wushu made it in 2003, and the effect on the sport was the same in kind: competitive emphasis moved sharply towards the elements that carried declared value. There is a second obstacle that gymnastics does not have. Wushu's two disciplines are not comparable with each other at all — a taolu score and a sanda result are a judged performance mark and a fight outcome — so there is no sport-level number to chart even within one era.

An eventing score is a penalty total across the phases of the competition. In 2005 three of those phases were removed. Before the change a horse arrived at the cross-country course having already covered roads and tracks twice and a steeplechase in between; the score recorded performance under accumulated fatigue. Afterwards it records performance on a fresh horse over one course. The dressage and jumping phases did not change, but the quantity they are added to did. So a penalty total from 2003 and one from 2007 are not two values of one thing. The sport did what modern pentathlon did seventeen years later — removed a component and kept the name — and the score stopped meaning what it had meant. This is worth stating alongside pentathlon's entry, because the two are the only cases here of a sport changing what it is composed of. Eventing did it for welfare reasons in 2005 and pentathlon did it for welfare reasons in 2022, and both removed the phase involving a horse.

Wrestling scores are not comparable across 2013, and the reason is unusually stark: before that year, a bout's winner was not necessarily the wrestler who scored more points. From 2005 a bout was best of three two-minute periods. Each period was won separately, so a wrestler could lose 1–0, win 5–0 and lose 1–0, and be defeated by an opponent who had scored two points to their five. A scoreless period could be resolved by a clinch, or in freestyle by drawing a coloured ball from a bag to decide who started with the advantage. From 2013 a bout is two periods of three minutes with points accumulating across the whole match, and the higher total wins. So a scoreline before 2013 records a sequence of separate period results, and one after it records a total. Those are different quantities, and no mapping exists — the same wrestling could produce either outcome depending on how the points fell across the periods. It is the same structural break archery made in 2011 by moving to sets, and taekwondo made in 2022 by moving to best of three rounds. Wrestling is the only one of the three that moved in the opposite direction, from sets back to a total.

There is no chart here, and there is not going to be one. Until 2006 a routine was scored against a maximum of 10.0. From 2006 the score is a difficulty value with no upper bound plus an execution score out of 10, and the two numbers are added. A 9.85 and a 15.633 are not two values of one quantity. They are values of two different quantities that happen to share the word "score", and no conversion between them exists — not an approximate one, not a rough one, not one good enough for a dotted line. This is a stronger claim than the one made about the javelin. There, two incomparable series each measure a real distance in metres, and a chart of each is meaningful even though a chart across them is not. Here the quantity itself was replaced. Nadia Comăneci's 10.0 in 1976 does not sit lower than a modern 15.7; it sits outside it. The vertical axis a chart would need cannot be constructed. It is also worth saying what was lost. Between 1949 and 2005 gymnastics had a public, legible, unambiguous top of the scale, and the perfect 10 became one of the few numbers in sport that a general audience understood without explanation. The 2006 Code bought accountability and the ability to reward difficulty, and it paid for them with that.

Esports produces more data than any sport on this site and almost none of it survives a comparison across time. Every action in a competitive match is logged by the server, so the raw material is extraordinary — complete positional, economic and outcome data for every game ever played. It is also worthless for progression, because the game the data describes is replaced every few weeks. A publisher patch can rebalance a character, a weapon, an economy or a map. A statistic drawn from one version measures a game that no longer exists, and a professional season is played across several versions. Compare gymnastics, which rewrote its scoring scale once in 2006 and severed its history doing it: esports does the equivalent on a schedule, by a company, without consulting anybody. Then the discontinuities that are not patches. A title can be discontinued, and its competitive record ends. A title can be replaced by a sequel, and the numbers do not carry across. A title can be contested at a multi-sport games in a specially modified version, as several were at Hangzhou 2022 — so even a single tournament's data may describe a game built for that tournament. There is no line to draw here, and no federation with the standing to draw one, because none of them owns the game.

A modern pentathlon score is a sum of points from separate disciplines, converted onto a common scale by tables the federation writes. That means the score depends entirely on which disciplines are in the sport and how each is converted — and both have changed repeatedly. In 2009 five events became four, when shooting and running merged into a handicap race in which the finishing order is the result. In 2011 the pistol stopped firing a projectile. From 2028 riding is gone and obstacle racing replaces it. So a total from 2008 and a total from 2028 are not two values of one quantity. They are sums over different sets of things, on conversion tables the sport rewrote in between. This is the gymnastics case rather than the javelin case: the scale itself was replaced, more than once, and there is no mapping in either direction. It is also the only sport on this site where the *composition* of the competition changed. Athletics redesigned an implement; weightlifting redrew its categories; modern pentathlon removed a discipline and added a different one.

Twice over, taekwondo replaced the quantity rather than rescaling it. Before 2009 a point existed when a majority of four judges pressed a button within a fraction of a second of each other. After 2009 a point exists when a sensor registers force above a threshold. Those are not the same measurement: one is an agreement between people about what they saw, the other is a physical reading. A score of 8–5 in 2007 and a score of 8–5 in 2011 record different kinds of event. Then in 2022 the contest stopped being a cumulative score at all. A match is now best of three rounds, so the result is a count of rounds won, and a competitor can outscore an opponent across a contest and lose it — the same thing archery did in 2011. There is no chart because there is no continuous quantity to chart. This is the gymnastics case twice in thirteen years.

There is no chart here, and this is not the badminton case. Before 2011 a recurve match produced a number: the sum of the arrows, out of a known maximum, and the higher number won. After 2011 a match produces a scoreline in set points, six-four or six-two, which is not a measurement of anything the archer did with a bow. The quantity was not rescaled. It was replaced by a different kind of object, in the way gymnastics replaced its score in 2006. The severity is easiest to see in the fact that makes archers angriest: an archer can now out-shoot an opponent across a match and lose it. Any series running across 2011 would be joining a measure of accuracy to a measure of who won more sets, and those are not two values of one quantity. What could exist, and does not here, is a series of ranking-round totals — those are still cumulative and still comparable, subject to the 2014 arrow diameter limit. That would be a real chart. It has not been assembled from a cited source.

DanceSport's scores cannot form a series, and the reason is the same one this site records for gymnastics — arrived at from a stranger starting point. For most of the sport's history, judges did not score couples at all. They placed them, one against another, and the placements were resolved by a counting rule borrowed from figure skating. A result under that system was a statement about a particular field on a particular floor: it could say a couple was third, and could not say by how much, or how they would compare with a couple in another final. The WDSF's absolute judging system replaced this in its own competitions with marks against defined components. That is a scale change in the strictest sense — before it, there was no scale, only an order — so nothing on either side of it can be placed on one axis. Gymnastics moved from a closed ten-point scale to an open-ended one in 2006 and severed its own history. DanceSport moved from no scale at all to a scale, which is the same severance with even less on the earlier side of it. There is a third obstacle. Judging remains split between federations, so two couples of comparable standing may be marked under different systems by different bodies in the same season.

A Trivium result is a percentage split between two dancers in one battle. It is not a score either of them received. Gymnastics awards a difficulty value and an execution mark against a published standard, so a routine has a number of its own. Diving multiplies a judged mark by a tabulated degree of difficulty. Breaking does neither: a judge moves a slider towards whichever dancer they thought was better on each of five criteria, and the aggregate says who won and by how much. That means there is nothing to chart. A dancer who wins 62–38 in one battle and 55–45 in another has not produced two comparable performances; they have beaten two different opponents by two different margins. Comparative judging cannot generate a series by construction. This is a deliberate design decision rather than an oversight, and a defensible one. Breaking had no absolute scale to borrow, because forty years of battles had been judged holistically by practitioners. Inventing one would have meant asserting a standard the form did not have. The site records it as impossible rather than unpublished for that reason. There is no quantity here that anyone declined to publish.

Karate scores two different things and neither produces a comparable series. Kumite awards points for controlled techniques — strikes delivered with good form, distance and awareness, deliberately not landed hard. A score is a panel's judgement that a strike would have worked, which is a counterfactual rather than an observation. It is closer to squash's let than to taekwondo's sensor reading, and it is why karate cannot be instrumented the way taekwondo was. Scoring values were also re-weighted in 2019, so the units changed. Kata was judged by flags — a comparative decision between two competitors, with no number attached — until 2019, when it moved to a points score out of a stated scale. A flag result and a points score are not two values of one quantity, in exactly the way breaking's Trivium split and gymnastics' execution mark are not. So there is no chart, and the two disciplines fail to produce one for opposite reasons: kumite because its scores are judgements about strikes that did not land, kata because the sport replaced a comparative system with an absolute one.

A skateboarding score is an impression out of 100, awarded holistically, on a course that is different at every competition. There is no difficulty table. That is not an omission — it is a decision. Every other judged sport on this site prices its elements: gymnastics assigns difficulty values, diving tabulates degrees of difficulty, taekwondo prices head kicks and spins, climbing prices holds. Skateboarding declined, because its culture does not accept that a trick has a fixed value independent of where and how it is performed. So an 88.6 in one park final and an 88.6 in another are not two values of one quantity. They are two panels' impressions of two runs on two different courses. The criteria have also been revised between every Olympic cycle the sport has contested, which means the scale itself has not held still across the two Games it has appeared at. This is the same shape as breaking's entry and for the same underlying reason: a culture with holistic contest traditions was asked to produce a transparent scale in a hurry, and chose not to invent one it did not believe in.

Sport climbing has had three different Olympic scoring systems in three Games, and none of them produces a number comparable with the others. In 2020, one medal covered all three disciplines and the result was the product of a climber's ranks. In 2024, speed separated off and boulder and lead were added on a shared hundred-point scale. The 2020 score is a dimensionless product of positions; the 2024 score is a points total. They are not two values of one quantity, and neither is comparable with whatever comes next. There is a second problem inside the first. Multiplying ranks means the value of a position depends on the other results — improving in one discipline can lose you ground overall — so even within 2020 the score is not a measurement of climbing. It is arithmetic about placings. The honest thing to say is that this sport's Olympic results are three separate experiments. Speed climbing is the exception and it is charted nowhere here only because the times have not been assembled from a cited source: because the route has been identical worldwide since 2007, those times are genuinely comparable, which is more than most sports on this site can say.

Croquet's scores are bounded and its competitions are handicapped, and between them there is nothing to chart. An association game is won at twenty-six points and cannot exceed it. A scoreline of 26–14 and one of 26–8 are both wins, and the losing figure reflects how many turns the loser got rather than how well they played — in a game decided by breaks, a strong player may win 26–0 without the result meaning they played better than in a closer one. Handicap play removes what remains. Bisques are free turns taken at a moment of the recipient's choosing, so a handicap result depends on when they were spent as much as on the play. And handicaps are recalculated as players improve, so a given player's results are measured against a moving entitlement. The two codes rule out the rest. Association croquet and golf croquet share a lawn, hoops and balls, and are scored differently, played differently and contested by increasingly different people. A number from one says nothing about the other. What could be measured — the proportion of games won from a completed triple peel, say — is recorded by enthusiasts and has never been assembled into anything cited.

Soft tennis's numbers are not comparable with tennis's and were never meant to be, which this site notes only because the two sports share a court, a net and a name. A soft tennis game is four points, not fifteen-thirty-forty. There is no set. The serve is not a weapon, so there is no service-hold structure and nothing that corresponds to a break point. Rallies are much longer, because a thirty-gram ball cannot be hit through anybody. Every rate statistic in the sport describes a different game. Internally, the sport has had no scoring rewrite of the kind that severed badminton, volleyball, table tennis and fistball between 1999 and 2006. Its four-point game predates international codification and has not been changed — so a continuous series is genuinely conceivable here, which is more than most racket sports on this site can say. Nothing is entered because nothing has been taken from a cited source. The competitive record lives in Japanese, Korean and Taiwanese domestic competition and has not been assembled internationally.

Softball's statistics are not comparable with baseball's and were never meant to be, which this site notes only because the Olympic programme treats the two sports as one entry. A softball game is seven innings, not nine. The pitch is underhand from a flat circle at a shorter distance. The ball is larger, the bases are closer, and the field is smaller. Every rate statistic in the sport describes a different game. Internally there is a break too. The international tie-breaker, from 1987, starts each extra half-inning with a runner on second, which raises scoring in exactly the innings where games are decided. Bat performance standards were tightened in 2018 after composite bats had raised batted-ball speeds. The figures are not entered because they have not been taken from a cited source. What can be said without them is the thing worth recording: baseball adopted softball's tie-breaker in 2020, thirty-three years late, and severed its own extra-innings statistics doing so.

Polo's scoreboard does not start at nil, and that alone prevents any series. In handicap competition the difference between two team handicaps is awarded to the weaker side before play begins. A match may start at two and a half goals to nothing, and the final score is a compound of goals awarded and goals scored. Counting them together measures a rating decision as much as a performance. Matches are also scoped by handicap level. A twenty-two-goal tournament and a twelve-goal tournament are different competitions with different players, different horses and different scoring rates, and results do not cross between them. Then the horses. The string is widely reckoned to account for the majority of a player's performance, strings vary enormously in quality and cost, and since the 2010s some of the best animals in the sport have been clones of one mare. A goals-per-match figure would be measuring an asset market. This is the same shape as golf's problem and the opposite arrangement. Golf's handicap adjusts a score after the round, leaving the elite game's scratch scores comparable. Polo's awards goals before it, so nothing is left uncontaminated — the handicap that makes the sport playable makes its numbers unreadable.

The FAI ratifies more records than any other body on this site, and none of them belongs on a single line. The weather sets the ceiling. A gliding distance record requires a specific meteorological phenomenon — mountain wave downwind of a range, usually — that occurs reliably in a small number of places on earth. The great distance records are set in the Andes and in a handful of comparable ranges, and they measure the availability of a weather system as much as the skill of a pilot. Competitive gliding acknowledges this openly by scoring tasks relative to what the rest of the field achieved in the same air. The classes multiply. Records are held by aircraft class, by discipline, by category of flight, and by whether the aircraft was launched under its own power. A record is a fact about a category rather than about the sport. The remit exceeds the sport. The federation certifies absolute aeronautical performance — altitude, speed, distance — for flights that were never competitions, and its definitions extend beyond the atmosphere altogether. That is why the accepted boundary of space at 100 kilometres is a sporting body's definition, adopted for the purpose of deciding which records were aeronautical and which were astronautic. A chart cannot hold that. What this space records instead is the oddest fact about this federation: it was founded in 1905 to arbitrate performance rather than to run contests, and it has been arbitrating things far outside sport ever since.

Pelota holds the fastest ball speed recorded in any sport, and it is not a series. The figure usually cited for cesta punta is around three hundred kilometres an hour. It is a measurement taken on particular occasions, by particular methods, rather than a ratified record with a progression behind it — there is no equivalent of a governing body homologating ball speeds the way athletics homologates a throw, and no continuous list to chart. The specialties make anything broader impossible. A ball struck by a bare hand, a solid wooden pala and a metre-long wicker basket are three different projectiles, and the balls themselves differ — leather over a wound core for the hand, rubber for paleta. Speeds across specialties are not comparable and were never intended to be. Nor are results. Each specialty is contested separately at world championships, on courts of different lengths, and the professional hand game — the sport's largest audience — is run by private promotion companies whose competitive records are not the federation's. So the sport's most striking number sits here as a single fact rather than a line, which is the honest way to hold it.

No continuous scoring series exists for flag football, and the reason is that the sport is younger than most of its own rule changes. The five-a-side international format dates from 2002. Before that, flag football was a family of local codes — four a side, seven a side, eight a side, with and without contact blocking, on fields of varying size — and figures from those codes describe different games. Since 2002 the format has been stable but the competition record is thin: a world championship every two years, with a field whose strength varies more between editions than the scoring does. A rate statistic drawn from it would measure who entered, not how the game is played. What can be said is structural. There is no punt and no field goal, so every possession ends in a score or a turnover on downs, and scoring rates are high by the standards of any football code. Whether they have moved is not something a cited source currently supports.

Bridge scores are relative by construction and cannot be placed on any axis that runs between events. A duplicate result is a comparison. In pairs, a matchpoint score says how you did against everybody else holding those cards on that day, in that field. In teams, an IMP margin says how you did against one other pair at one other table. Neither number means anything outside the session that produced it, and both are bounded by the field's size rather than by performance. A top on a board is a top whether you beat the field by one trick or five. Two identical percentage scores in different events describe different achievements against different opposition. This is not a defect. It is the mechanism that makes bridge fair: comparing results across identical deals removes the luck that would otherwise decide the game, and the price of that comparison is that the number only exists inside the comparison. What the sport does have is masterpoint and ranking systems that accumulate across events, but those measure participation and placing over a career rather than strength on a common scale — they are not the instrument chess has, and the federation does not present them as one.

Orienteering cannot have a progression of times, and it is the only sport on this site that has legislated that outcome on purpose. Two rules make it impossible. First, the terrain is different every time and is embargoed beforehand, so no course is ever run twice by anybody. Second — and this is the unusual one — the rules instruct course setters to build to a target winning time. A sprint is planned so the winner takes about fifteen minutes; a middle distance, about thirty-five. The winning time is therefore an input to the course design rather than a measurement of the athlete. A chart of winning times over the years would be a chart of how well course setters hit their targets. It would tell you nothing about whether anybody got faster, and it would look like a flat line drawn by the rulebook. What the sport does measure, since electronic punching arrived in the 1990s, is splits between controls within a single race — the richest per-race data of any sport covered here, and comparable only inside the race that produced it. Set this against athletics, where the whole apparatus of the sport exists to make a time in one stadium comparable with a time in another fifty years later. Two measured sports, opposite answers.

Handball changed which sport it was in 1966 and how it attacks in 2016, and both moved scoring substantially. Before 1966 the international game was eleven a side, outdoors, on a football pitch. Afterwards it was seven a side indoors on a court, and the scoring rate multiplied. Those are different sports sharing a name, in the way rugby fifteens and sevens are. The 2016 change is subtler and this site finds it more interesting. Removing the goalkeeper-substitution restriction let teams attack seven against six with an empty net, and elite teams do it constantly. That raises the attacking team's conversion rate and simultaneously concedes long-range goals into an empty net, so both teams' scoring moved and in the same direction. The series is scoped to international tournaments because domestic leagues adopted the 2016 rule at different points and the IHF publishes no continuous series for the sport as a whole. The figures are not entered because they have not been taken from a cited source. When they are, the chart appears here with two segments and a caption noting that the second break was an accident — a procedural tidy-up that changed how the sport is played.

Billiards sports cannot produce one series because they are not one sport. A carom average, a pool rack count and a snooker break are three different quantities produced on three tables of different sizes with different balls and different numbers of pockets — none of which a player of another code could compete on. The confederation that unites them was formed in 1992 for institutional reasons and has never claimed otherwise. Inside snooker, one number does hold. The maximum break of 147 is fixed by the rules and has not changed, and unlike bowling's 300 it has not been devalued by equipment: cloth and balls have improved, and a maximum remains rare enough that each one is still counted. That makes snooker's perfect score one of the very few bounded maxima on this site that still means what it meant. What prevents even that becoming a series is where the record lives. Snooker's elite competition, its rankings and its statistics belong to a professional body formed in 1968, separate from the confederation representing billiards to the Olympic movement. The numbers are excellent and they are not the recognised federation's.

Baseball keeps better records than any sport on this site and has severed them more often than almost any. Runs per game, batting average, earned run average and every rate statistic built on them respond directly to rule changes, and this sport has made three large ones. The mound was lowered and the strike zone shrunk in 1969, both designed to make hitting easier. The designated hitter arrived in one league in 1973 and the other in 2022, which means that for forty-nine years the two halves of the same competition were not comparable with each other, let alone across time. The pitch clock and shift restriction in 2023 moved the numbers again. The series is scoped to one league because there is no such thing as baseball's runs per game: the international game, the professional leagues of Japan and Korea, and MLB all differ, and MLB itself was two rulebooks for half a century. The figures are not entered because they have not been taken from a cited source. When they are, the chart appears here in segments — and it will need more of them than any other chart on this site.

Cricket did not replace a format. It added two, and kept the first. A batting average in Test cricket measures how many runs a player scores before being dismissed, over five days, with no limit on how long they may take. The same statistic in a Twenty20 match measures something else entirely, because a batter who occupies twenty overs carefully has failed at the job. The sport knows this, which is why it keeps three separate sets of records and has never once claimed they are comparable. That makes cricket the clearest voluntary case on this site. Everywhere else, a governing body severed a series by accident or under pressure and then had to decide what to do about the record book. Cricket created the discontinuity deliberately, in 1971 and again in 2003, and handled it correctly from the start by never merging the lists. So there is no chart here, and there should not be one. A combined batting average across Tests, one-day internationals and Twenty20 would be an arithmetic operation on three different quantities — and unlike every other absence on this site, nobody in cricket has ever wanted it.

Finswimming keeps world records and they are, within a discipline and a fin type, genuinely comparable — which makes this one of the few absences on this site caused only by the figures not having been assembled from a cited source rather than by anything in the sport. What the sport does have is a record book divided more finely than most. Surface, apnea, immersion and bi-fins are four physically different events; a monofin time and a bi-fins time over the same distance are two different measurements and are never pooled. The reason this page exists rather than a chart is worth stating, because it is the site's central question answered a third way. Swimming faced polyurethane suits in 2009, banned them, and let the records stand — which is why several remain unbeaten. Athletics faced the javelin in 1986, redesigned it, and reset the book. Finswimming has never been able to consider either, because an athlete without a fin is not competing in this sport at all. So it caps the fin by dimension, leaves materials broadly open, and splits the book by fin type. Equipment selection remains part of the sport, openly, and the records carry that with them.

Hockey changed the ground under its own game in 1976, removed offside in 1998, and stopped the clock in 2015. Each of those moved scoring, and the first two moved it substantially. A goals-per-game figure spanning them would be measuring three different sports. On grass the ball did not run true; on turf it does. With offside, attackers could not wait behind the defence; without it, they can. With a running clock, a leading team could waste time; with a stopped one, it cannot. The series is scoped to Olympic tournaments because, as in football, there is no body that keeps a continuous quantitative series for hockey as a whole and domestic competitions adopted these changes at different times. The figures are not entered because they have not been taken from a cited source. When they are, the chart appears here in segments — and the caption will have to say that the 1976 break is a change of surface rather than of rules, which is a distinction nothing else on this site has had to draw.

Kickboxing cannot produce a comparable series, and the obstacles are stacked three deep. The disciplines are different sports. Point fighting halts after every scoring technique and awards one, two or three points; the ring codes run continuously and are scored round by round on a ten-point must system. A point-fighting total and a full-contact decision are not the same kind of number. The codes disagree on targets. Full contact bars kicks below the waist; low kick permits them; K-1 style adds knees. Results across them describe contests with different permitted techniques, which is the same objection this site records between amateur and professional muaythai. The federations disagree on everything. Several international bodies sanction world titles, IOC recognition went to one, and no mechanism reconciles their records. As in powerlifting, there is no single authority to say what a world championship result is. What could be charted — the minimum kick counts actually thrown in full contact, for instance — is not published in any cited form. It would be genuinely interesting, because it would measure how hard the rule is working.

Before 1979 every field goal in the NBA was worth two points. After it, some were worth three. Points per game, field goal percentage, offensive rating and the value of a possession are all measured against a different scale on either side of that line, and no rescaling recovers the difference — a pre-1979 shot from 25 feet was worth two points and there is no record of how many were taken. The series is scoped to one league on purpose, and this is a stronger version of the football problem. Basketball does not have one rulebook: FIBA adopted the three-point line five years after the NBA, at a different distance, and moved it again in 2010. A figure for "basketball" would be averaging across competitions whose courts are different shapes. The figures are not entered because they have not been taken from a cited source. When they are, the chart appears here in two segments with 1979 between them, and the caption will still have to say that the NBA changed the distance twice in the 1990s and changed it back.

Powerlifting cannot have a world record progression, and it is the only sport on this site whose obstacle is not one severing event but a permanent condition. There is no single world record because there is no single federation. The sport split, repeatedly, over drug testing, and the federations that resulted differ in what they test for, what supportive equipment they allow, and how strictly they judge squat depth. Each keeps its own book. A number from one cannot be placed on a line with a number from another, because the conditions that produced them differ in ways the kilograms do not carry. Inside the IPF alone there are two more divisions: raw and equipped, kept apart deliberately since 2012 because a bench shirt returns stored elastic energy and the totals are not the same measurement. And the bodyweight categories were redrawn in 2011, which detached the records attached to the old ones — the same instrument weightlifting used in 1993 and again in 2018. So the honest rendering is this space. Powerlifting is a measured sport in which the measurement is unambiguous and the comparison is impossible, which is the reverse of most of the absences recorded on this site.

Nothing is entered for the underwater team sports, and the honest reason is that nobody has compiled it. These are small sports. Underwater hockey and underwater rugby have world championships, national leagues in a handful of countries, and no professional structure anywhere. Match statistics exist in tournament records and have never been assembled into anything continuous or cited. There is also a facility problem that would distort any series if one existed. Underwater rugby requires a pool at least three and a half metres deep, and depth varies between venues within the permitted range — so the volume of the playing area, which is the sport's defining variable, is not constant between competitions. What can be said without numbers is that scoring in both sports is low, and for a reason worth recording: possession is limited by breath rather than by a clock or an opponent, so an attack that would succeed given ten more seconds simply ends when the attacker has to surface. It is the only sport covered here where the limiting factor on a move is the athlete's need for air.

There is no series to draw here, and the reason is structural rather than archival. A pétanque match ends at thirteen points. The number is bounded, it has not changed, and a final score of 13–7 in 1960 means what it means in 2026. There is nothing to progress and nothing to break. What could in principle be measured is shooting accuracy — the rate at which an elite *tireur* displaces a target boule — and the precision shooting discipline generates exactly that number. It is not published in any cited, continuous form, and it would in any case be scoped to that discipline rather than to the sport. The umbrella structure rules out anything broader. Pétanque, boule lyonnaise, raffa and volo are separate codes with separate federations, separate rules and separate championships, united by a confederation formed in 1985 for institutional reasons. A number from one says nothing about another, and the confederation has never tried to make one.

Dragon boat times cannot form a progression, for the reason rowing gives and refuses to call its marks records: the water decides too much. A 500 metre time depends on the course, the current, the wind, the water temperature and the boat. Dragon boats are not one-design in the way an Olympic class dinghy is, and a heavy festival boat and a modern racing hull are different craft. A time set on one lake says very little about a time set on another. Crew composition compounds it. Standard and small boats, open, women's, mixed and BCS divisions all race the same distances, and the sport's mixed divisions specify the number of each sex aboard — so a mixed time is a function of a rule about who is in the boat. There is a governance obstacle as well. Two international federations run dragon boat competition in parallel, and neither ratifies the other's results, which puts the sport in the same position as powerlifting: not short of numbers, short of an authority to say what they mean. What the sport has instead is finishing order, which is all it claims.

There is no such thing as football's goals-per-game figure. Football has no central record book and no body that keeps a continuous quantitative series for the sport, so any rate statistic belongs to one named competition and means nothing outside it. That is why this series is titled with its competition and not with the sport, here and everywhere else it appears. Even scoped, this particular series has a confound that deserves stating rather than smoothing over. The back-pass rule took effect in July 1992. The English top flight was reconstituted as the Premier League in February 1992 and began play the same month the rule did. A change in goals per game across that line is a change across a rule amendment and a competition reorganisation at once, and this site will not present it as measuring the rule. The figures themselves are not entered because they have not been taken from a cited source. When they are, the chart appears here in two segments with the 1992 line between them, and the confound stays in the caption.

Muaythai runs two scoring systems for the same sport, and they are not translations of each other. International amateur competition uses a ten-point must system, scored round by round and added up — the same additive structure as boxing. A fighter who loses two rounds must win the rest. Thai professional scoring assesses the bout as a whole, weights the later rounds most heavily, and rewards dominance, balance and visible effect rather than counting strikes. A fighter can concede the opening rounds deliberately, find the range, and take the fight in the fourth and fifth. That is a recognised tactic in Thailand and it is unavailable under an additive system, because the arithmetic will not allow it. So a score from one code says nothing about a performance in the other, and the two coexist rather than one having replaced the other. There is no break to draw, because there was never a single line. The equipment differs too — the amateur code requires headguards, shin guards and elbow pads and restricts elbows by category — which means the two systems are not even scoring the same permitted techniques.

Duathlon times cannot form a progression, and the sport's own rules give two independent reasons. The courses differ. A duathlon is raced on roads, and every course has its own profile, surface, altitude, number of turns and weather. Two standard distance races are the same only in the numbers on the entry form. This is the position road running and cycling are in, and both handle it by treating a time as belonging to a course. The formats differ more. Elite short-course racing is draft-legal and most other racing is not. A draft-legal bike leg is a bunch race in which sitting in a slipstream saves a large fraction of the effort; a non-drafting bike leg is a time trial. The bicycles are different, the tactics are different, and the race is decided by different disciplines. A time from one says nothing about a time from the other. What could be compared is the split — how much slower the second run is than the first — because that ratio is the sport's distinctive physiological signature and is course-dependent in a smaller way. It is collected by teams and is not published in any consistent cited form, so nothing is entered.

Ju-jitsu cannot produce one series because it is not one competition. Fighting scores points across three phases and can end in a submission. Ne-waza scores positional advancement and can end in a submission. Duo is marked by a panel of judges on the realism and control of a cooperative demonstration. These are three different quantities wearing the same word. A Duo mark and a Fighting score have less in common than a gymnastics D-score and a boxing decision. One is a judged aesthetic-technical assessment of two team-mates; the other is a count of things one person did to another. There is also nothing stable underneath. The federation added Ne-waza as a discipline in 2015, which changed what the world championship contains, and the largest grappling constituency in the world competes under other bodies whose scoring is different again. So this space records what the sport is rather than what it measured: a federation running three competitions that share a mat, a gi and a name.

Cheerleading scores cannot form a series, for the reason every judged sport on this site gives and one that is particular to it. The general reason: a routine is marked against a scoresheet that is revised, and difficulty values change with it. A score is a statement about one performance against one season's rubric. The particular reason is more interesting. Cheerleading's scoresheet is bounded by its safety rules, and those rules are what change most. The permitted stunt classes, pyramid heights, released skills and toss forms are restricted by age and level and are tightened when injury evidence warrants it. So the ceiling on difficulty is set by a safety code rather than by what athletes can do — and when the code changes, the maximum achievable score changes with it. No other sport here has its scoring ceiling set by a medical argument. Gymnastics removed its ceiling in 2006 to let difficulty run open-ended; cheerleading keeps a ceiling deliberately, and moves it for reasons that have nothing to do with sport. The two divisions compound it. A Cheer score and a Performance Cheer score measure a stunting routine and a dance routine, and are not comparable at all.

Rugby's Olympic sport and rugby's principal sport are different games that share a rulebook. Fifteen a side runs eighty minutes with contested scrums of eight, lineouts, and long periods of possession. Sevens runs fourteen minutes with three-man scrums and enormous space, and produces scoring rates several times higher. A points-per-game figure spanning them would be meaningless, and nobody in rugby has ever suggested otherwise. This is now the third sport on this site to send a reduced format to the Games while its main competition is played in another — after cricket, which goes as Twenty20, and lacrosse, which invented Sixes for the purpose. In all three cases the discontinuity is deliberate and handled correctly, by keeping the records separate. Rugby's version is the starkest because the Olympic history is split across it. The sport was contested as fifteens from 1900 to 1924 and returned in 2016 as sevens, so its own Olympic record spans two different games with a ninety-two-year gap in the middle.

Lacrosse now runs four versions of itself: men's field, women's field, box — played indoors on a hockey rink — and Sixes, which was invented in 2021. They differ in team size, field dimensions, equipment, contact rules and scoring rates, and none of their statistics travels to another. Sixes is the one that matters for this site. It is six a side on a smaller field with a shot clock and no specialist positions, and it produces far more goals in far less time than field lacrosse. A goals-per-game figure spanning the two would be measuring two different sports that share a name. What makes it worth recording rather than merely noting is why Sixes exists. It was designed in 2021 to fit an Olympic athlete quota and a broadcast window, and the sport was added to the programme for 2028. Cricket added Twenty20 for commercial reasons and kept three separate record books; rugby adopted sevens for the Games; lacrosse invented a format from scratch for admission. In all three cases the discontinuity was deliberate and is handled correctly, by never combining the numbers.

Lifesaving keeps world best times for its pool events, and this site does not chart them, for a reason particular to this sport: the technique they measure is not allowed to be stable. Competition technique must reflect current rescue practice. When the underlying medical and rescue guidance changes — how an airway is managed, how a casualty is held, what a resuscitation sequence looks like — the rules follow, and a method that was correct and fast becomes non-compliant. The clock is honest; what it is timing has been redefined more than once by people outside the sport. The 1993 merger compounds it. Pool and ocean events came from two different federations and two different traditions, and ocean conditions cannot be standardised at all — a surf race time says as much about the sea that day as about the swimmer, which puts it in the same position as rowing's times, and rowing declines to call its marks records for exactly that reason. Nothing is entered here because nothing has been taken from a cited source with the rule version attached. A time without the technique standard it was set under would be the kind of number this site exists to avoid.

A golf score is strokes taken on a particular course, on a particular day, in particular weather. Two rounds of 68 are not the same achievement unless the courses and conditions were the same, and they never are. That would make golf the least comparable sport on this site — except that golf noticed first and built a machine for it. The handicap takes a score and adjusts for the difficulty of the course, the tees played, and since 2020 the conditions on the day, and produces a number that genuinely can be compared across players, courses and countries. So this entry is here for the opposite reason to every other absence in this collection. There is no chart of raw scores because raw scores were never comparable and golf has known it for a century. What exists instead is a formal system for making them comparable, which nothing else here has attempted. The elite game, which does not use handicaps, keeps the problem. A winning score at one championship and another are not comparable, which is why golf's records are things like number of titles rather than numbers of strokes. The 2028 ball change adds a further break at the top end: the object itself will fly shorter, so distances and scoring records set before it stand in a different game.

There is no chart here, and rowing is the reason this site should have expected that before it started. Every other federation in this collection publishes records and then has to decide what to do when they stop being comparable. Athletics reset the javelin list. Swimming kept the supersuit marks. Weightlifting annulled everything twice. Cycling split its hour record in two and then merged it back. Rowing declined the premise. It recognises world best times, not world records, and its stated reason is exactly the argument this site makes everywhere else: a 2000-metre course's wind, current and water temperature cannot be standardised, so a time set on one day and a time set on another are not measurements of the same thing. A chart of best times would therefore be a chart of which crews happened to race in a tailwind, and the federation says so itself. The times exist, they are published, and they are not a series. This is the only entry on this site where the absence of a chart is not a gap in the research or a consequence of a rule change. It is the federation's own position, held since before the problem arose.