The Week-Eleven Breaking Point: Shoulder Injury Data From 43 Vietnamese Swimmers

On July 14, 2026, in lane 4 of a 50-metre pool, a nineteen-year-old swimmer abandoned a 200-metre butterfly set at the 130-metre mark. There was no shout, no collision, no fall for a camera to replay. He simply stopped, put both hands on the wall, and told the coach standing at the head of the pool that his right shoulder could no longer push water. Three weeks earlier, on my tracking sheet, he sat in the group of four athletes with the lowest risk scores.

The column I was watching was not a pain column. It was the column for distance per stroke cycle. Over eleven consecutive weeks that distance shortened: 2.11 metres per cycle to 2.03, then 1.98, then 1.91. At the same time, stroke rate rose from 31 to 34 cycles per minute. The swimmer was not getting weaker. He was compensating with speed for the propulsion his body had lost. Nobody on the coaching staff checked that column, because before I built it, the column did not exist here.

The body is a closed system, but data is the key that opens it.

Swimming is a sport where injury almost never arrives in a single moment. There is no studs-up challenge, no bad landing, no contact. In the lane, every session is four to seven thousand metres repeating one motion, and the bill is paid at compound interest. A swimmer's shoulder takes somewhere between twelve hundred and two thousand revolutions per session depending on event, multiplied by six days a week, across forty weeks a year.

In the international literature, the share of competitive swimmers who have experienced shoulder pain ranges from 40% to 80% depending on the study and on how shoulder pain is defined. A breaststroker's knee takes valgus load in the kick, the same mechanism as a medial collateral ligament injury. A butterfly swimmer's lower back takes repeated extension load. Those three regions rarely rest at the same time, which is why an athlete entered in four events carries four different risk curves running in parallel.

In Vietnam the picture has extra layers. One coach typically handles ten to fifteen swimmers, doubling as technique coach, load manager and psychologist. Dedicated sports-medicine staff for swimming at provincial level are close to non-existent. Load monitoring, which many swimming nations have treated as a minimum standard for more than a decade, is still largely done on paper and from memory. The domestic calendar then compresses into a few months, while a training cycle needs four to six uninterrupted months.

I once sat at a training centre, counting laps and writing them into a notebook, then copying them into a spreadsheet when I got home. There is nothing glamorous in that. Only patience.

Across the 2026-2026 season I tracked 43 swimmers in four groups: six butterfly specialists, eleven sprint freestylers, sixteen middle-distance freestylers, and ten individual medley swimmers. The system recorded four families of measures. External load covered total metres per session split across four intensity zones, plus dryland volume. Internal load took the session rating of perceived exertion multiplied by minutes, then a seven-day rolling average. Technical measures covered distance per stroke cycle and stroke rate, taken from four standard 50-metre swims in the main session. The last family was status: shoulder pain score, internal and external rotation range, and rotator cuff strength.

The results after twenty-nine weeks. Forty-one of the forty-three swimmers recorded shoulder pain above some threshold in at least one week. That rate sounds enormous, and it is, but it sits inside the published range. The interesting part comes next: twenty-seven crossed the threshold that forced at least one missed session, and nine showed structural tendon change on ultrasound.

The distribution over time is what concerned me most. Eighteen of those twenty-seven threshold cases, or 67%, fell between week nine and week thirteen of the training cycle. Not week one, not week nineteen, not the peak week before a meet. The dead zone sits in the middle, when accumulated volume is high and the programme has not yet entered its taper. I call it the breaking point at week eleven, and it has nothing to do with whether the swimmer had a bad session that day.

What happens in that window? Weekly volume rises from a baseline of roughly 38 km to 58-62 km, while the share of work at threshold and VO2 intensity climbs from 18% to 31%. At the same time dryland volume rises, with lat pulldowns and supine pressing. Together, the shoulder absorbs three rising load sources inside the same three-week window. None of them alone would cause injury. Combined, they produce a breaking point.

I built a correlation table on this exact cohort and came away with three thresholds that have practical value.

A week in which pool volume rises more than 12% above the four-week average: risk of crossing the shoulder pain threshold rises roughly 2.4 times within the following ten days.

Volume swum with hand paddles exceeding 25% of total pulling volume: risk rises roughly 3.1 times. That is the number that surprised me most, because paddles are the cheapest and most widespread tool in any pool, and the first thing a coach reaches for when trying to build propulsion quickly.

Distance per stroke cycle falling more than 5% below an individual baseline and staying there for ten days: risk rises roughly 3.8 times. That is the strongest single indicator of the three, and the one almost nobody measures.

Why does distance per cycle signal so early? Because it is the combined output of things a swimmer cannot hide. When the supraspinatus tendon is overloaded, a protective reflex makes the swimmer shorten the front-end catch, release the hand earlier, and lose part of the propulsion. The swimmer does not feel it for the first two or three weeks. He only notices he is more tired, that his in-training 50-metre time is a little worse, and he attributes it to poor sleep or under-eating. By the time the pain surfaces, distance per cycle has already been down three to seven percent for weeks.

I checked this against video. In the early-decline group, the elbow angle at the catch changed from roughly 110 degrees to 95-100 degrees over three weeks. That gap is invisible within one session, but obvious when two clips twenty days apart are laid side by side.

The measurement does not require expensive equipment. A phone on a tripod at the pool wall, filming one lane at a high frame rate. Count stroke cycles over a 25-metre segment, divide 25 by the number of cycles, and that gives distance per cycle. Error is around 2% if the camera angle is right. Record once a week, same swimmer, same lane. After four weeks the trend line draws itself.

Internal load gives a signal too. In the group tipping into overload, session rating of perceived exertion for the same set rose from 6 to 8 over roughly twelve days, while the volume of that set did not change. In other words, the body quoted a price before the tendon structure was damaged. The problem is that the signal sat in a notebook nobody read.

By stroke, the butterfly group had the highest threshold rate: four of six. Middle-distance freestyle was lowest: five of sixteen. Individual medley sat in between. The reason is not psychological. Butterfly demands an overhead pull with the largest shoulder range of any stroke, combined with hip rhythm and a double propulsion phase. Add the fact that butterfly sets often carry heavy paddle volume, and you have a formula that funnels load into one joint.

Nguyen Huy Hoang's 1500-metre freestyle is the counter-example. The absolute volume is enormous, but it is distributed evenly across the stroke cycle, and freestyle lets force spread across more muscle groups. Total metres is never the most important number. How those metres divide by stroke and by intensity zone is what decides.

Tran Hung Nguyen races the individual medley, which means his body absorbs four different load patterns in the same week. Pham Thanh Bao in butterfly absorbs the most concentrated pattern. Two athletes of comparable level, comparable weekly volume, and very different risk. A monitoring sheet has to reflect that instead of using one team-wide threshold.

Dryland programmes in many places are built for strength targets rather than shoulder protection. Supine pressing, behind-the-neck lat pulldowns and shoulder presses appear in almost every plan I have ever seen. In the group with behind-the-neck pulldowns in the week, the shoulder threshold rate ran about 1.9 times higher than in the group without. That lift puts the shoulder in maximal external rotation under load, directly into the inflamed structure. Switch to front-of-chest pulldowns, keep the load identical, and nothing is lost in strength terms.

In the fourteen-to-sixteen age band there is another variable. During rapid growth, arm bone length increases faster than tendon and ligament adaptation. Same set, same volume, a fifteen-year-old carries higher risk than a twenty-year-old. In my cohort, the threshold rate among the under-seventeens ran about 1.7 times higher than among the over-nineteens. That is why the 12% volume-increase threshold should drop to around 8% when applied to youth squads.

In sessions without a coach at the head of the pool, including solo sessions, make-up sessions and voluntarily added work, technical markers drifted about 1.6 times faster than in main sessions. Empty stands, the golden rule bends, and the body pays. In swimming, an empty stand is not an empty grandstand. It is unmeasured time. A swimmer adds eight hundred metres of butterfly because he feels it was not enough, and nobody records those eight hundred metres anywhere.

Most analysis of shoulder injury in swimming stops at technique. Slow-motion video shows a dropped elbow, a hand entering too wide or too narrow, insufficient body rotation. All of that is correct. And all of it is expression, not cause.

Technique is the only thing a camera sees. No camera records the sequence of volume increases across a four-week block. No footage shows week seven up 11% and week eight up 13%. When a swimmer's shoulder hurts, the clinic has illustrative video, while the case conference has a load chart that does not exist. So technique gets convicted, because it is the only defendant present at the scene.

My data points elsewhere. Same swimmer, same technique, same coach. Change the loading sequence and the outcome changes. In the second season we kept the entire technical programme intact and only changed the volume ramp: splitting the increase into two smaller steps ten days apart instead of one large step. Threshold cases in the sixteen middle-distance swimmers fell from five to two. No new technical drill was added.

The second counter-intuitive angle concerns treatment. The first reflex when a shoulder hurts is rest. Three days off, the shoulder settles, the swimmer returns to the same set at the same volume. Within seven to ten days the symptoms return, usually worse. The problem is not the rest. The problem is that the road back was never divided into steps. Recovery is a process, not a decision.

The protocol I propose at grassroots level has four stages, each at least three sessions, twelve days minimum in total. The first stage is easy swimming with no equipment, volume at 40% of baseline, capped at 1,500 metres. The second adds band pulls and range work, pool volume to 60%. The third reintroduces pulling, but removes hand paddles, volume to 80%. The fourth returns to the normal programme but holds volume steady for two more weeks before any increase. Skip a stage and the relapse risk rises accordingly.

Over two seasons of applying it, the monitored group had 6 relapses in 31 cases within six weeks, against 14 in 29 in a comparison group that returned via short rest and immediate full training. That is not enough to call high-grade evidence, but it is enough for me to keep applying it and keep measuring.

One thing needs stating plainly about reliability. Forty-three swimmers over twenty-nine weeks is a small sample. The 12%, 25% and 5% thresholds are working thresholds, not thresholds confirmed on a large sample. A single season cannot eliminate confounders such as a coaching change, a change of pool, or a change of school schedule. What I can assert is only this: in this cohort, in this season, loading sequence was the variable most strongly correlated with threshold cases, stronger than technique, stronger than age, stronger than sex.

Every fall has a graph, and every graph has a breaking point. In the lane, the fall makes no sound. It is a curve running flat in silence, then dropping at week eleven.

What I carried out of those twenty-nine weeks was not a new drill or a new device. It was a question I would ask every swimming coach to answer before entering the next cycle: what does your week eleven look like on paper?

If the answer is a blank space, the body will write into it. And it always writes in the most expensive language there is.

Numbers stay silent, but their sequence always knows how to tell a story.