Doubles Silver in Le Mans: When Scratch Pairings Beat Every Data Model

On the men's doubles table in Le Mans, Nigel Tarling (Brighton TTC) and Rob Cook (Leeds ParkyPING!/Community TTC) walked into their finals alongside two partners they had never once trained with. No tactical blueprint, no warm-up session together, no unspoken convention about service rhythm or the division of responsibility between the two halves of the table. Both still left the tournament with silver medals.

I logged that result into my tracking sheet close to midnight, and what made me stop was not the medal. It was the structure of the event that produced it. The second PingPongParkinson French Open, more than 80 players from 10 countries, not a single ranking point awarded, not a single prize dollar distributed. The dataset in front of me looked nothing like any sheet I have built in five years of working with table tennis data.

One thing has to be said up front: this event sits outside the ITTF and WTT systems. It belongs to the sport-for-health tier, built for people living with Parkinson's disease. The organisers run the draw under the C1, C2 and C3 classification system, grouping players by degree of motor impairment, following conventions close to para-sport practice. That is the central fairness mechanism of the event, and it is also the single most important organisational variable I have to feed into my reading model.

For elite table tennis I have three data layers ready: rally-by-rally scoring, service placement, return trajectory. Here I have none of them. No shot-level statistics, no frame-by-frame footage of individual strokes, no advanced metrics. All I have is match results, the classification system, and a few lines describing how things unfolded. Anyone familiar with my work knows that feeling: you choose between forcing a thin dataset into a story that sounds plausible, or admitting you are reading a map that was never finished.

I chose the second option. We do not hunt treasure, we hunt the way to read the map. And for that reason, the real value of this event lies somewhere else entirely, not in who beat whom.

In elite table tennis, doubles is a pure chemistry problem. Left-right hand combinations, style compensation, service-zone division, cross-court movement timing — all of it is polished across hundreds of hours of shared training. I once built a prediction model for doubles pairs on just three variables: shared training frequency, handedness offset, and a zone-compensation index. That model performed reasonably well at national-team level.

In Le Mans, that variable is deliberately neutralised. Both English silver medals came from scratch pairings, assembled on site, across different nationalities. What remains measurable is not rapport, but the speed of synchronising with a total stranger, and individual adaptability under real-time pressure.

That is why I read Rob Cook's singles results far more carefully than the doubles medal. In singles, Cook lost 3-2 to Wilco Jupil of France. The sequence is the part worth recording: Cook fell 0-2 behind, clawed his way back, and in the deciding game dropped to 3-10 before pulling it to 8-10. He still lost. But that run of consecutive points from a hopeless position is behavioural data, not technical data.

For a Parkinson's athlete, holding a run of points while already fatigued, while the hand begins to tremor, is a completely different problem from holding a run when the body is healthy. Tremor and bradykinesia do not disappear at the tenth point of the fifth game. They are there, and they tighten as the battery drains. The only indicator I can honestly extract from this tournament does not sit in the hand, it sits in the head: tolerance for the rhythm of the scoreboard.

In the men's doubles, the picture is structurally similar. The Cook/Tigellaar pair lost 3-1 to Spain's Alonso & Lobarinas in the final. In the other half, the final between Tarling/Duerr and Gallon & Lacassagne is described as tight, turning on a handful of key points. Two finals, two different outcomes, one shared signal: the skill gap inside each classification group is narrow enough that a few points decide the medals. Once the field has been flattened by the classification system, what is left of the margin belongs to small variables: one skewed serve, one wrong positional choice, one minute of lost focus.

If I had shot-level footage, I could point to exactly where each point landed. I do not have it. And rather than filling that gap with speculation, I record it as a data limitation. Numbers do not lie, they only keep secrets. Data cannot save a match, but it can point to why the match died.

There is another layer of information more notable than the results. Both English players are attached to club-level infrastructure: Brighton TTC, and Leeds ParkyPING!/Community TTC. The name ParkyPING! suggests a dedicated table tennis group for people with Parkinson's, sitting inside or tightly bound to a mainstream community club. That is the most important form of infrastructure in the sport-for-health model: patients are not separated from the club, they are embedded in it.

This is also where my analytical frame has to change axis. Every metric I normally use to assess a development system — age structure, generational conversion efficiency, coaching-staff stability — fails to apply. The real organisational unit here is the community club and the volunteer group, not a national team. And the existence of a named group like ParkyPING! is a pipeline signal at grassroots level, not at performance level.

There is a temptation anyone in the business of reading numbers has felt: turning a shortage of information into a counter-intuitive discovery. I could write that scratch pairings outperform trained pairs. But I have no basis for that. I have two cases, two silver medals, and one event-design decision.

What I do actually know is this: pairing on site is not designed to optimise results. It is designed to maximise encounters — an English player paired with a German, an English player paired with a Dutch player, strangers becoming teammates for two days. Here, a good doubles pair is not a carefully assembled machine, it is a conversation that starts on the right beat.

At the same time, there is an invisible data layer no statistics table can touch: sense of balance, reflex latency, and the confidence level of a person who knows their own body may betray them on the very next point. Every number is a recitation, every calculation a contemplation. But I have sat with data on an empty-court night, and I know where its limits lie.

Let me be blunt: if I applied elite analytical frameworks — fixture density, points-defence pressure, technical overhauls, generational gaps — to this event, I would produce a model that is beautiful and entirely wrong. That framework does not apply here. Admitting it is not a lack of expertise. It is the expertise.

What I will track next cycle is not the performance of Tarling or Cook. It is three other measurable things. Whether the number of participating countries rises above 10. Whether any peer-reviewed research confirms more strongly the effect of table tennis on Parkinson's symptoms. And whether the C1-C3 classification criteria, still a black box in the public record, are published transparently.

Do not ask data what the future holds, ask what the past is whispering. A second-edition event, with more than 80 people from 10 countries and not one ranking point at stake, is telling me a longer story than any medal. I will let it run a few more seasons before I write the conclusion.