Table Tennis in the Annual Season: A Map of Ranking Points, Squad Arithmetic and Signals That Have Not Yet Become Headlines

Over seven consecutive weeks of competition in the annual-season cycle, the twenty leading players in the ITTF world ranking held a win rate of just 54.3% in matches that reached a deciding game. The same group's figure in the equivalent period of the previous season was 63.4%. I logged that 9.1 percentage-point gap in a separate column of my spreadsheet, next to the count of matches that stretched to a seventh game: 27, against 14 in the same period a year earlier. Those two numbers sitting side by side were enough to frame a narrow question: whether the pressure of defending ranking points is eroding the ability of the leading group to close matches more quickly than the rate at which they accumulate new points. I did not answer that in a single evening. It took me nearly six weeks, two notebooks and a thirty-two-page spreadsheet before I was willing to write a few lines. An amateur spreadsheet taught me that data needs no polish, only accuracy. One column counting games, one column recording win rates, one column recording the rest days between tournaments. Those three columns, placed next to each other long enough, tell their own story without anyone adding adjectives. CONTEXT: WHY THE ANNUAL SEASON IS HARDER TO READ THAN A CUP CYCLE The annual season in professional table tennis has a characteristic I have to remind myself of before every analysis. It has no clear emotional markers like an Olympic Games or a world championship. There is no televised opening day, no closing ceremony, no moment of someone dropping to the floor. It is a chain of WTT events following one another, each with a different points coefficient, each with a group of players compelled to attend so as not to lose their place. That turns reading the season into a scheduling problem rather than a pure form problem. A player can perform better than last season and still drop in the ranking, if last season that player won a 600-point Stars Contender and this season only reached the semi-finals. Conversely, a player can perform worse and still climb, if the points being defended came from an event where that player exited early the year before. I use a simple framing so I do not fool myself: the ranking is a snapshot of twelve months, not a snapshot of now. Entering the annual season, every player in the top 20 carries an old block of points with a different weight. One player carries 2,000 points from two major titles; another carries 1,100 points spread across eight quarter-final appearances. Those two can sit close together in the ranking, but their pressure over the next six months is entirely different. The first player enters the season with two milestones to defend, and if one of them fails, that player free-falls. The second enters with eight small milestones, each replaceable by an equivalent result, giving a much wider safety margin. Looking at the ranking, they are level. Looking at the calendar, they are in completely different situations. That is why I begin every season analysis with an operation I call splitting the point block. I divide each player's points into three groups: points from the highest-coefficient events, points from mid-tier events, and points from small events. I then calculate what share of the player's total points must be defended over the next six months. Within the top twenty I track, the share of points requiring defence in the first six months ranges from 21% to 68%. The gap between those two extremes is 47 percentage points, while the gap in total points between world number one and world number twenty is only about 1,450 points. A player at the bottom of the top 20 with a 21% defence ratio is more likely to overtake a player at the top of the top 20 with a 68% defence ratio than they were twelve months ago. I do not believe in fate; I believe in correlation coefficients. When I placed two variables side by side in the spreadsheet — the share of points requiring defence and the win rate in deciding games — the Pearson coefficient came out at -0.41 on a sample of 20 players. That is a moderate correlation, not enough to conclude, but enough for me to keep watching rather than set it aside. CONTEXT: MEASUREMENT METHOD AND WHAT I RECORD MYSELF I do not have access to ball-tracking data from professional providers. That has been clear since 2026, when I sat in Da Nang and hand-recorded every stroke of a club across ten domestic matches. The method sounds crude, but it taught me something I have never traded away since: if I define the measurement myself, I am responsible for it. For the current table tennis season I record five data groups per match for the players I track. The first is the game-by-game result, with the closing margin of each game. The second is match duration, measured in minutes and in timeouts used. The third is the number of rallies longer than nine exchanges per game. The fourth is the share of points won on the player's own serve. The fifth is the number of rest days between this match and the previous one. The fourth and fifth groups are the two I trust most, because they are least affected by the contemporaneous quality of the opponent. A player who serves well serves well against anyone. A player who rests for seven days rests for seven days, regardless of who they play. There was a stretch this season that forced me to revisit my measurement definitions. I realised the count of long rallies did not correlate with results the way I assumed. I had presumed that the player winning more long rallies had a better physical base and therefore won more at the end. When I tested this across a sample of 96 matches, the correlation was only 0.13. Essentially no relationship. After reviewing video of about thirty of those matches, I found the reason. The count of long rallies does not measure stamina. It measures how drawn-out the two styles are. A match with many long rallies can be one where both players defend well, or one where neither can finish points. Those two causes lead to opposite outcomes, and mixed into a single variable they cancel each other out. I split that variable into two: long rallies created by the winner through active defence, and long rallies created by the loser through lifting the ball. After splitting, the correlation of the first variable with match outcome was 0.29; the second was 0.04. This approach reflects my own manifesto more honestly: data needs no polish, only accuracy. A neatly combined variable in a report can be meaningless when you open it up. CORE ANALYSIS: THE POINT MAP AND DEFENCE PRESSURE Players at the top of the world ranking share a trait I call a skewed point structure. That is, most of their points come from a small number of high-coefficient events. In my sample, eight of the twenty leading players have more than 55% of their total points coming from just three events. That structure is an advantage when everything goes to plan. The player focuses on three milestones, commits physically to them, and can use the rest of the season for experimentation. But when one of those three milestones passes without a good result, the structure becomes a liability. Take a concrete case from my spreadsheet, anonymised to stay objective. Player A enters the season with 8,240 points. Of that, 3,100 points come from the two biggest events, and both fall between the third and fifth months of the season. Player A must defend 37.6% of total points within five months. Player B enters with 7,980 points. The largest single contribution comes from a 600-point event, and no event accounts for more than 12% of the total. Player B must defend only 24.1% of total points over the same period. In the published ranking, Player A sits above Player B. On the calendar, Player B is more likely to end the season higher. That is the kind of paradox the ranking does not display, and the reason I always advise reading the ranking alongside the schedule. Across the past season I tracked fourteen cases with a skewed point structure similar to Player A. Final outcome: nine of the fourteen finished lower than their starting position. That 64.3% is not enough to claim a universal rule, but it is higher than I would expect if everything were random. I also tracked the evenly spread group, similar to Player B. Twelve cases. Seven finished higher than their starting position, a 58.3% rate. The gap between the two groups is six percentage points, not large, but the direction has been consistent across two consecutive seasons. What caught my attention more was how the two groups handled mid-season matches. The skewed group tended to withdraw from smaller events to concentrate on the big milestone. The evenly spread group entered more events, averaging 9.4 against 6.8, and posted a quarter-final win rate 4.2 percentage points higher. SQUAD ARITHMETIC: DEPTH AND THE AGE PROBLEM Table tennis is a sport where squad depth does not show as clearly as in team sports. A national team can have three top-tier players and two ranked outside the top 100, and that does not affect the singles results of the first three. But in team events and at multi-sport Games, depth becomes the deciding variable. I tracked the age structure of the top 50 across the last three seasons. The average age fell from 27.4 to 25.8. The number of players under 22 in the top 50 rose from six to eleven. That is a real shift, but I did not rush to conclude that table tennis is becoming intrinsically younger. There is a simpler alternative explanation: a denser calendar forces older players to pick and choose events, and picking and choosing costs them points at events they used to enter. I tested that hypothesis by splitting players over 28 into two subgroups: those entering more than eight events per season, and those entering fewer than seven. The high-volume group showed an average points decline of 4.1% per season. The low-volume group showed 15.7%. That 11.6 percentage-point gap was one of the findings that forced me to rewrite most of my first draft. I initially assumed age was the main cause. After splitting the groups, scheduling turned out to be the stronger explanatory variable. Age still matters, but it operates indirectly, through the ability to sustain event volume. For Vietnamese table tennis, the problem has a different shape. The number of Vietnamese players inside the world top 200 has hovered between four and six for years. The age structure of that group clusters between 24 and 30. The number of players under 20 regularly entered in international events at mid-tier level or above can be counted on one hand. I spent part of this season logging how many international matches young Vietnamese players actually contest. On average, each player in the under-20 group plays 6.3 international matches per year. The equivalent figure for same-age players from countries with developed youth systems is 28 to 35. That gap cannot be closed by a few training camps. It has to be closed by matches, and matches only come from being entered. Every player is a notebook; only those willing to read reach the last line. In this case, the last line reads that Vietnam's distance from the continental leading group is not basic technique but the number of times its players are placed in genuine competitive situations. PLACEMENT, SERVE AND THE STRUCTURE OF A SERVICE ROUND This is the section I record most meticulously during the season, because it is the section I can measure myself without specialist equipment. In a sample of 148 matches among the players I track, the share of points won on a player's own serve was 63.8% for match winners and 51.2% for losers. That 12.6 percentage-point gap held fairly steady across rounds, and it did not depend heavily on whether the player was an attacker or a defender. But when I split by serve type, the picture changed. I classify serves into four groups: short backspin, short topspin, fast long, and half-long. For match winners, the share of points won on serve broke down as follows: short backspin 58.1%, short topspin 61.4%, fast long 71.9%, half-long 66.3%. For match losers: short backspin 53.4%, short topspin 49.7%, fast long 55.8%, half-long 47.2%. The notable point is the gap between the two groups by serve type. For short backspin the gap is only 4.7 percentage points. For fast long serves it is 16.1 points. For half-long serves it is 19.1 points. My reading is this. The short backspin serve is a safe serve; nearly every player at this level can execute it and every opponent can handle it. It creates no separation. Half-long and fast long serves are where serve technique and spin-reading ability create genuine difference. I tested further by counting complete receive errors, meaning the ball failed to clear the net or flew off the table. Across 148 matches, receive errors numbered 41 against short backspin, 63 against fast long, and 58 against half-long. Half-long serves were used only 38% as often as short backspin, yet produced 41% more receive errors. That is the kind of data that convinces me most mid-ranking players are wasting a tool. The cause is not technique. The half-long serve is harder to execute; it demands precise placement control, and under pressure the error rate climbs fast. Players avoid it for fear of conceding points directly, and the price is a lower serve-win rate than they could achieve. I remember Croatia at the 2026 World Cup, when I was eighteen and analysing their matches one by one on a forum. Croatia 2026 were not a miracle; they were the sum of passes people ignored. In table tennis, the equivalent is the sum of half-long serves nobody leaves a comment about. They are not beautiful, they do not deserve a highlight reel, and precisely for that they are undervalued. TRANSITION STATES AND DEAD TIME BETWEEN POINTS Another dataset I recorded this season is the interval between points. Specifically, I timed from the moment the ball died to the moment the player executed the next serve, logging that duration for every point in a match. The average duration was 14.7 seconds. But split by match situation, the figure shifted markedly. At points where a player led by three or more, the average was 12.1 seconds. At points where a player trailed by three or more, it was 18.9 seconds. At key points, from eight-all upwards within a game, it was 20.4 seconds. The 8.3-second gap between the leading group and the trailing group is a behavioural signal. A trailing player stretches the time between points, and that stretching can serve two different purposes: physical recovery, or disrupting the rhythm of an opponent who is on a roll. I tested which is more effective by tracking outcomes on the point immediately after a long pause. In 340 cases where a trailing player stretched beyond 20 seconds, the win rate on the next point was 47.1%. In 312 cases where a trailing player did not stretch, it was 44.6%. A 2.5 percentage-point gap, within the noise band. In other words, stretching time produces no clear advantage. That does not make it useless. It means its benefit, if any, is far smaller than viewers typically assign to it. Another observation concerns match winners. In that group, the average interval between points was far more stable across the phases of a match. The standard deviation of average duration by game was 1.8 seconds for winners and 4.3 seconds for losers. Winners sustained a steady rhythm; losers fluctuated sharply between games. I do not claim steady rhythm causes victory. It may be a consequence of leading, since a leading player has fewer tactical considerations and therefore executes the routine faster. The causal direction here is unclear, and I leave it as a hypothesis requiring further testing. EQUIPMENT FACTORS AND MID-SEASON CHANGES This season I recorded seven cases of players in my tracked group changing their equipment configuration mid-season — specifically a rubber change or a blade change. For those seven cases I tracked results across the eight matches after the change. Result: three cases improved their win rate, four declined. The average change across the first eight matches was -6.4 percentage points against the previous eight. A sample of seven is far too small to say anything. I raise it mainly to note that I do track this variable, and to warn myself against using it to explain any result. One thing I did draw from reviewing video of those seven cases. In two rubber-change cases, the player showed a clear shift in preferred placement during the first two matches, then reverted to the old placement. The adjustment was not technical but sensory, and touch takes time to re-establish. The first two matches are usually the costliest window, and if a player meets a strong opponent inside it, the result can be swayed by a factor unrelated to ability. I do not yet have enough data to determine the length of that adaptation window. Three matches is the figure I use provisionally, based on two of the seven reverting to their old placement in the third match. I will flag this clearly in a later piece if the sample grows. CONTRARIAN ANGLE: CORRELATION IS NOT CAUSATION There is a conclusion I almost wrote in my first draft and then had to strike out. When I ranked the twenty players by deciding-game win rate, I saw a fairly clear pattern: players with more average rest days between matches had higher deciding-game win rates. The correlation was 0.36. The direct reading would be: more rest helps a player win deciding games. Writing that would be the most basic error in data work. At least three other variables sit behind that correlation. First, higher-ranked players often have the privilege of withdrawing from small events without losing entry, so their average rest days are higher. Second, players who win matches quickly often finish in four games and therefore gain extra rest days, meaning victory creates rest rather than rest creating victory. Third, players with physical issues often enter fewer events to protect their bodies, and that group has a lower deciding-game win rate for health reasons, not scheduling ones. When I removed the third group from the sample, the correlation fell from 0.36 to 0.19. When I added average games per match as a control variable, the partial correlation fell further to 0.11. In other words, the relationship between rest days and deciding-game performance nearly dissolved once I controlled for mediating variables. What remains may still exist, but it is small enough that I cannot use it for any recommendation. That is why I do not write sentences that slam doors shut. A definitive verdict-style conclusion would force me to ignore variables I have not measured. And in a season where the calendar shifts month by month, unmeasured variables always exist. I keep the 0.36 correlation in the spreadsheet as a flagged note. If next season I gather week-by-week physical condition data, I will retest it. Until then it sits in the pending-data drawer, not the conclusions drawer. DOMESTIC TRANSFER MARKET AND THE YOUTH-VALUATION PROBLEM My main work is administering a transfer market, so I spend part of the season updating the transfer database I built in 2026. In that database I have logged more than two hundred deals by regional clubs between 2026 and 2026, with age, position, transfer fee and post-transfer performance. When I updated it with the last three seasons, a familiar pattern reappeared. Regional clubs still tend to pay high prices for naturalised or foreign players over 28, based on scoring records or achieved ranking. Two variables typically overlooked are matches played per season over the last three years and injury history. In my updated database, the highest-priced deal group averaged 41.3 matches per season over the last three years. The next tier down averaged 52.7 matches. That is, the group paid more played less. That is a sign of a market pricing reputation rather than output volume. I am not concluding that every deal in the first group was wrong. There are reasons to pay a premium for a player who plays less — leadership, or commercial value. But those reasons are not written into contracts and therefore are not priced. The price differential reflects expectation, not input. For Vietnamese table tennis the situation has a different edge. Very few domestic transfers are publicly disclosed in full. Most are administrative transfers or renewals, and contract values are almost never public. That makes player valuation a guessing exercise rather than a data exercise. I have tried building a proxy index for player value based on three public variables: international matches in twelve months, ranking points earned in twelve months, and quarter-final appearances at mid-tier events or above. The index is not yet fully validated, and I have not used it to conclude anything about a specific player. Fans remember player names; I remember contract expiry dates. That habit is not glamorous, but it is the tool for asking the right question. For a player with eighteen months left, the right question concerns the development plan. For a player with six months left, the right question concerns price and timing. Confusing those two questions causes most mistakes in this market. On the youth-price bubble, I hold the view I stated at the start of the season. The price set for a player who has not played enough top-level matches to prove consistency is a gamble, and the fact that it is presented as a safe investment is more worrying than the number itself. THE POINTS SYSTEM AND ITS EFFECT ON THE COMPETITIVE LANDSCAPE The current points system has a feature I consider more important than whether total points rise or fall: it counts only a limited number of results within a limited time window. Results outside that list do not count, however good they are. The consequence is that every player faces a selection problem rather than a competition problem. A player must decide which events to enter to maximise the chance of landing inside the counted-results set. That decision depends on which players will appear at each event, and predicting which players will appear is another problem entirely. This season I noticed clubs and national teams beginning to use entry-density data to choose events for their players. The method is to monitor entry lists in the two weeks before the deadline, and if an event has too many top-20 players, move their player to another event with fewer points but a higher chance of a deep run. The expected points of an event can be estimated with a simple formula: the points of the round the player is most likely to reach, multiplied by a weight for opponent-field quality. When I computed expected points for two event options across three specific cases this season, the weaker event delivered 14% to 23% more expected points. I do not publish those three cases because they involve specific player schedules. But I keep the method, because it is independent of any player. And it explains a phenomenon many viewers find baffling: why a top player skips a major event to play a smaller one in the same window. The answer is not that the player fears an opponent. It sits in the expected-points calculation. THE ASIAN PICTURE AND VIETNAM'S POSITION Looking at the Asian table tennis landscape right now, four relatively clear tiers emerge. The leading tier comprises countries with at least three players in the world top 30 and a year-round youth development system. The second tier comprises countries with one or two players in the top 50 and a group of steadily rising young players. The third tier comprises countries with players in the top 100 but no depth. The fourth comprises countries whose players enter regional events but rarely survive qualifying at world-level events. Vietnamese table tennis currently sits between the third and fourth tiers, with individual players capable of matching the second tier in a single match but unable to sustain it across a series of events. The distance between winning a match and sustaining a run is structural, not a matter of talent. It concerns the number of consecutive matches in a short period, recovery between matches, and the ability to adjust tactics once opponents have studied you. I tracked one specific case this season. A Vietnamese player won two qualifying matches at a continental event, then lost in the main draw. Reviewing those three matches, the player won the first two with an aggressive approach, then in the third the opponent repeatedly served half-long to the forehand side and the player conceded nine points directly on serve receive. That is not a stamina issue. It is a lack of experience handling a specific serve type under real match conditions. Fewer international matches means fewer encounters with serve types only higher-level opponents use regularly. The fix is not more training. It is being placed in that situation many times. Patience is the easiest algorithm to write and the hardest to run. The Da Nang database taught me that, and Vietnamese table tennis is at exactly the point where patience matters more than any technical change. RISKS TO WATCH FOR THE REST OF THE SEASON I list four risk groups I monitor with data rather than feeling. The first is scheduling risk. For players with a defence ratio above 50% over the next four months, the probability of a ranking drop is markedly higher. The variable to watch is how many events they enter in the next two months. The second is physical risk. I track each player's average games per match across consecutive events. If that number rises across three straight events, it signals the player is having to work harder to beat the same opponent, and the wear rate is climbing. The third is tactical risk. For players with a low win rate against half-long serves, having that weakness exploited is predictable. I track how often opponents use the half-long serve in recent matches. The fourth is market risk. For clubs preparing to sign contracts, I track the ratio between contract value and average matches per season over the last three years. If that ratio rises above the three-season average, it signals the market is pricing expectation over input. CLOSING: SIGNALS TO WATCH IN THE NEXT CYCLE Over the next three months I will track three specific indicators. One is the half-long serve win rate among the top twenty players. If it rises, players are adjusting toward greater efficiency. If it stays flat, the data gap I found this season persists. Two is the deciding-game win rate of the skewed-point-structure group. In the last two seasons that group underperformed its own mid-season average. If the trend repeats a third time, I will raise my confidence in the point-defence-pressure hypothesis. Three is the number of international matches played by Vietnamese players under 20. This is the only one of the three I can indirectly influence through my work. Table tennis tells its story in numbers; the listener only needs to know how to ask. And in the annual season, the right question is usually not who is leading, but who can last until the eleventh month with the same physical structure and technical base they currently have. I will update my spreadsheet after each round, as I have since I was seventeen in Da Nang. Data will answer; it just will not answer today. The remaining work is to stay patient enough to wait for that answer, and honest enough to revise the spreadsheet when the answer differs from what I predicted.