We Know What Happened. We Still Don't Know Why.

In the first essay, I argued that junior tennis had already built the athlete pipeline investors are starting to notice. We just never thought of it as one thing. A player enters the sport, starts training, plays tournaments. Results pile up, ratings move, the family spends money at every turn, and somewhere down the line there's recruiting, or college tennis, or the quiet end of one chapter. Facilities, tournaments, coaches, technology companies, and governing bodies have each watched their own slice of that. The player and the family are the only ones who live the whole of it as one continuous thing.

That left me with a question I couldn't put down. If we're going to connect the pipeline, what should actually move through it?

The obvious answer is data. And that's where it gets interesting, because tennis is already drowning in data. We know, in enormous detail, what happened to a player. What we know far less well is why it happened. Those turn out to be very different kinds of knowing, and the difference is the whole subject of this essay.

We Got Good at Recording Outcomes

If I wanted to know what happened to a junior player last weekend, I could find out fast. The draw, the score, the opponent, the rating before and after, maybe some video. Follow the player long enough and a whole competitive history takes shape. That's real information, and it's worth having. It's also badly incomplete, and the easiest way to see how incomplete is to take a single line of it seriously.

A player loses 6-3, 6-2. What does that tell me?

It tells me she lost. That's about the end of it. Maybe the other kid was simply better. Maybe our player was in the middle of a change we'd been working on and got temporarily worse before she got better. Maybe she threw out the game plan after three games. Maybe she executed the plan perfectly and the score hides real progress. Maybe she panicked on every big point, or served the best she ever has and still couldn't return a ball. Maybe she played beautifully. Maybe she played terribly. The score knows none of that. The rating doesn't either. And even good analytics can miss the one thing that matters if nobody wrote down what the player and coach were trying to do in the first place.

That's the distinction I keep coming back to. Outcome data tells us what occurred. Developmental information tells us what it meant. Tennis has gotten very good at the first and left the second mostly to memory.

A Match Isn't a Test Unless You Know What You're Testing

I've written before about the trouble with treating competition like a final exam while a kid is still learning the material. But the problem underneath that one is simpler. An exam only tells you something if you know the question.

Say we spend three weeks asking a player to use height and shape when she gets stretched wide, instead of what she's always done, which is try to rip a winner off a ball she has no business attacking. At the next tournament, she loses. If the result is all I have, I might decide nothing changed. But suppose I was there, and I watched her get pulled wide seven times, and six of those times she lifted the ball, bought herself time, repositioned, and stayed in the point. Three weeks earlier she'd have gone for all seven and missed most of them. The scoreboard didn't move. The player moved a great deal.

That's obvious the moment someone says it out loud. Yet most of the machinery around junior tennis is built to keep the score and throw away the context. We save the answer and lose the question. Then Monday comes, the player's back at practice, the coach may or may not have seen the match, the parent has a version, the player remembers pieces, and everybody carries some secondhand account of the weekend into the conversation. There's rarely any durable record of what the intention was going in, what the player experienced during it, what an observer actually saw, and what should change as a result. So we start over.

The Space Between Practice and Competition Keeps Vanishing

This is the part of development I've gotten most interested in. There's a gap between training and competition, and things fall into it.

A player can understand an idea perfectly in a lesson and never recognize the moment to use it in a match. Another recognizes the moment but picks the wrong response. Another picks right but can't execute under speed. Another executes fine until the score starts to matter. Those are four completely different problems, and they can all produce the same missed backhand. If the only thing we record is the missed backhand, we haven't recorded much at all.

This is why I think coaches have to watch their players compete, not just train. Competition surfaces information you can't fully manufacture on a practice court, because a player's perception, nerves, and decisions all change once the result is real. But watching by itself isn't enough. If the information disappears by Tuesday, a coach can spend an entire weekend gathering something valuable and let almost all of it evaporate. The watching only pays off if what you saw makes it back into the work.

The Player Knows Things the Coach Doesn't

There's another hole in the data we usually collect. Almost all of it looks at the athlete from the outside. The score, the ranking, the rating, the video, even the coach's own observation, all of it is external. Every one of those matters, but development is also being lived from inside the player, and the systems we've built are strangely poor at capturing that part.

A coach might decide a player went passive because she was scared to miss. Ask the player and she might tell you she wasn't scared at all, she just couldn't tell which ball she was supposed to attack. Those are different problems with different fixes. A coach thinks the player abandoned the game plan; the player thought she was running it. A parent is sure the kid lost confidence; the kid says she was confused, not shaken. Every one of those disagreements is information, and it's information you only get by asking. For most of tennis history the systems we've built have treated the athlete as the thing being measured while collecting almost nothing about how the process feels from where she's standing. If we're trying to develop a person, that's a strange thing to leave out.

Why I Started Using Cards

This is part of what pushed me toward the Training Cards and Match Cards I've been working with. They aren't sophisticated. That's the point of them. A Training Card asks a player to name what she intends to work on, then reflect on what actually happened. A Match Card carries the same habit into competition.

Underneath both is a loop I call IEDE: Intention, Experience, Debrief, Evolution. Set an intention, have the experience, debrief what really happened, use it to change what comes next, then run it again. No single step is new. Coaches have set goals and debriefed matches forever. The difference is making the loop visible and repeatable enough that the information stops leaking out between cycles. The player says what she was trying to do. The match says what happened. The coach adds what he saw. Someone else might add something different. And the next practice gets to start somewhere smarter than the last one did. That's a developmental record. Not a complete one, but it holds onto things a score and a ranking never will.

A Second Set of Eyes

It's also where our faith in objective data can quietly mislead us. Human observation is biased, of course it is. The primary coach knows the player, and that knowledge cuts both ways. It creates insight, and it creates expectation. We start seeing a kid through the story we've already written about her. He gets tight. She won't defend. He starts slow. Sometimes the story is true. Sometimes it's just a filter we can no longer see through.

A second coach watching the same match may catch what the first one misses, precisely because he isn't carrying the same history. That doesn't make him objective. It hands us a different bias, and a different bias is useful. The record I'm describing wouldn't try to scrub human interpretation out of the picture. It would hold several views side by side and let patterns show up across them. The player says one thing, the primary coach sees another, the tournament coach notices a third, the video shows a fourth, the result adds a fifth. No single one of them is the whole truth. Together they start to get interesting.

From Data to Developmental Memory

This is where AI starts to matter to me, and not for the reason it usually comes up. I don't much care that cameras are getting better and better at measuring racquet angle and movement. It will, and plenty of companies will build fine products around it. What interests me is memory.

Human development systems forget. Coaches forget, players forget, parents forget, and programs forget everything the moment a coach who kept it all in his head takes another job. The notebook goes in a box. Tuesday's practice gets buried under Wednesday's problem. The thing somebody noticed at a tournament six months ago rings a faint bell and nothing more. Now imagine that information simply doesn't vanish. Imagine a player with three years of real developmental history behind her: not just scores and ratings, but intentions, coach observations, her own reflections, what people saw at tournaments, what we tried, what repeated, what looked important and turned out not to be, what looked trivial and surfaced again half a year later.

Then imagine asking that record a question. When have we seen this before? What has changed in how she describes this problem over the last year? Do the coaches see in practice what the observers see in matches? We've worked on this three times now, what happened after each try? And the one I care about most: are we solving the problem we think we're solving? That's not really analytics anymore. It's closer to memory, the kind a system can hold across the years and the staff changes that human attention never manages to keep.

Memory Isn't Judgment

There's a danger in that, and it's worth naming plainly. If a system remembers everything, people will start to believe it understands everything. It won't. A record can preserve context without supplying judgment, and holding more information may actually make judgment matter more, because someone still has to decide what any of it means. Someone still has to know that a twelve-year-old who grew four inches this year will look, for a while, like a worse athlete than he is. Someone has to see that an ugly technical fix can be the functionally right one, to notice when the real problem isn't tennis at all, to tell when a parent needs less information rather than more, and to sit across from a kid and read the difference between fear, confusion, fatigue, and an ordinary bad day. The point of developmental memory isn't to turn a child into a database. It's to give the adults responsible for her a better memory to reason from. That's a different thing.

Deciding What to Remember

Here's where this essay and the last one meet. I argued before that the real question isn't who owns the pipeline, it's what the pipeline is taught to care about. The same holds for the data inside it. Wiring every result, rating, measurement, and video frame into one big system doesn't make development better on its own. It can just make the system more confident. Feed it nothing but outcomes and it will get brilliant at finding patterns in outcomes. Point it at engagement and it will get brilliant at spotting which families are likely to buy another tournament. Point it at ranking and it will optimize for ranking. Sometimes those line up with a kid's development. Sometimes they don't. The risk was never that the technology fails. The risk is that it works beautifully on a question we never should have asked.

Which is why what goes into the record matters so much, and why it shouldn't be everything. More data isn't better data. A useful developmental record needs to remember what the athlete was trying to learn, what she experienced, what trained eyes noticed, what we tried, what happened next, and where the same thing showed up in different places. It also needs to remember what we didn't know. "We're not sure yet" is developmental information. So is disagreement. So is the fact that something looked important once and later didn't. And it has to let us change our minds. Development is recursive: we form a read on a player, we test it, we learn, we revise, we test again. A record that doesn't let us revise our understanding of an athlete isn't developmental at all. It's just a permanent file of old assumptions, and those get more dangerous, not less, once a machine starts handing them back to us with the appearance of certainty.

We know more about what happens to junior tennis players than we ever have. The next step is learning to remember what it meant. And once that kind of memory can follow the player instead of living in one coach's head or one academy's filing cabinet, an old assumption starts to wobble. For as long as I've been doing this, we've treated the academy as the place where developmental knowledge lives. I'm no longer sure the academy of the future is mainly a place at all.

That's where I want to go next.

*This is the second in a four-part series that began with a question Lisa Stone put to the ParentingAces community after sharing a Sports Brief article: what are the implications of this for junior tennis?*

Written with Reggie, Claude, Gemini, Otter, and 64+ trips around the sun.