There is a number that shows up, in slightly different costumes, in nearly every pitch for a tennis membership platform: hit a shot ten thousand times and it becomes automatic. Sometimes it is "300 reps and the muscle remembers." Sometimes it is a tidy "21 days to a new habit." The premise underneath the marketing is always the same — that improvement is a volume problem, and that the right library of tennis drills, delivered on a schedule, is the volume.
We went looking for where that belief came from. The short version is that it has a real source in motor-learning research, the source is genuinely interesting, and it has been bent almost beyond recognition on the way to the checkout page. This piece is about the gap between the two — and about what a membership platform can and cannot do for a player who hits once or twice a week and is tired of stitching together advice from strangers on the internet.
Do membership platforms actually make you better at tennis?
The honest answer: a structured library of tennis drills can help, but mostly by fixing two unglamorous problems — what you practice and in what order — not by adding magic to repetition. The platforms that work are the ones that organize your hour. The ones that don't are the ones that sell you a counter, a streak, and a feeling of progress that your match results never confirm.
That distinction matters more than the production quality of any drill video, and we will spend the rest of this article defending it.
A short history of the rep
The idea that skill is built by accumulating repetitions did not start with tennis, and it did not start with anyone trying to sell you anything.
It started, roughly, with Paul Fitts. In 1967, Fitts and Michael Posner published Human Performance, which laid out a three-stage model of skill acquisition: a cognitive stage (you are thinking hard about every part of the movement), an associative stage (the movement gets smoother, errors shrink), and an autonomous stage (the movement runs without much conscious attention). This is the framework — usually uncredited — behind the phrase "muscle memory." Fitts and Posner never said muscle. They were talking about attention. The skill moves out of working memory, not into the bicep.
That model was robust and is still taught. The trouble began when people tried to attach a dosage to it.
The famous "10,000" figure most people are quoting comes from a 1993 paper by K. Anders Ericsson and colleagues in Psychological Review, a study of violinists at a Berlin music academy. Ericsson's group found that the most accomplished students had accumulated more hours of what they called deliberate practice — focused, effortful, feedback-rich work at the edge of ability — by age twenty. The number that survived into popular culture was roughly ten thousand hours. The word that did not survive was deliberate.
This is the first place the belief gets thinner than its confidence. Ericsson spent the rest of his career objecting to the way his work was used. His point was never that volume produces skill. It was almost the opposite: that ordinary repetition — playing the piece you already know, hitting the forehand you already own — produces very little, and that the difference between a good player and a great one is the quality of the attention inside the hours, not the count of the hours. A 2014 meta-analysis by Macnamara, Hambrick, and Oswald in Psychological Science, pooling 88 studies, found that deliberate practice explained about 18 percent of the variance in performance across domains, and less in some. Practice matters. Practice is not the whole story, and undirected practice may be a small part of it.
So we arrive at a strange situation. The science says: attention and feedback are what move a skill from cognitive to autonomous, and raw volume without them is nearly inert. The marketing says: here are five hundred tennis drills, hit play, accumulate reps. These are not the same claim. They are barely related claims.
What actually happens when you drill
Walk through it in the order it occurs on the court, because the order is where the real finding lives.
You feed a ball to yourself, or a machine feeds it, or a coach feeds it. You hit a crosscourt forehand. You hit another. You hit forty more, all crosscourt forehands, same spot, same rhythm. Within a basket or two, the shot feels better. The contact cleans up. You leave feeling like you grooved something.
You probably did not, or not as much as it felt.
This is the most replicated and most ignored finding in the motor-learning literature, and it has a name: the contextual interference effect. The foundational work is Shea and Morgan, 1979, in the Journal of Experimental Psychology. They had people learn movement patterns under two conditions. One group practiced in blocked order — all of task A, then all of task B, then all of task C. The other practiced in random order — A, C, B, A, B, C, scrambled. During practice, the blocked group looked far better. They were smoother, more accurate, more confident. Then the researchers tested both groups later, on a different day. The random group — the one that had looked worse the whole time — retained and transferred the skill substantially better.
Read that again, because it is the quiet core of the whole subject. The practice that feels best in the moment is often the practice that teaches you least. Blocked repetition produces a fast, flattering improvement that decays. Random, interleaved practice produces slow, frustrating, error-strewn sessions that stick.
The mechanism, as best the field understands it, is that randomizing the order forces you to reconstruct the movement plan from scratch every time, instead of running the same plan you just ran. That reconstruction — the effortful retrieval — is the thing that builds the durable representation. The forehand you have to find again is the forehand you keep. The forehand you merely repeat is a forehand you are renting for the afternoon.
There is a related finding worth holding next to this. Research on feedback frequency — Winstein and Schmidt, 1990, and a body of work after it — suggests that getting corrective feedback after every repetition can actually impair long-term learning compared with getting it less often. Constant feedback turns into a crutch. The learner stops generating their own error estimate and starts waiting for the external one. When the feedback disappears, so does the skill.
Both findings point the same direction, and both cut against the volume model. Skill is not built by smooth, well-fed, high-rep, instantly-corrected sessions. It is built by messy, mixed, partially-blind ones, with the kind of struggle that does not feel like progress while it is happening.
So what is a membership platform actually selling
Now we can be fair to the product, because there is a version of it that is genuinely useful, and the useful version is not the one in the marketing copy.
A recreational player who plays once or twice a week and works without a coach has three real problems. The first is selection — not knowing which two or three things actually limit their game, so they spread effort thinly across everything. The second is sequence — not knowing how to order a practice session so it builds something. The third is feedback — having no external eye, so technical errors calcify for years.
A good membership platform can help with the first two. It cannot really do the third.
On selection: a well-structured library that says, in effect, "if your second serve sits up, here are the three causes and the drill that isolates each" is doing real work. It is converting a vague dissatisfaction ("I want to be better") into a specific target ("my toss drifts behind my head, which is why I can't pronate"). That conversion is most of the battle, and a scattered YouTube feed is actively bad at it, because the algorithm optimizes for the video you will click, not the flaw you actually have.
On sequence: this is where the contextual-interference research should change how these platforms are built, and mostly hasn't. The honest application of Shea and Morgan is that a session should interleave. Three minutes of crosscourt forehands, then switch — to a backhand, to a serve, to a transition ball — and come back to the forehand later, having lost the groove and having to rebuild it. Most platforms instead serve drills in tidy thematic blocks, because tidy blocks feel productive and feeling productive is what gets renewed. A platform built around the actual science would feel worse to use and would probably teach you more. We have not seen many built that way. It is a genuine tension between what retains subscribers and what retains skill.
On feedback: a video cannot watch your toss. Some platforms now bolt on a "send us your swing" video-review feature, and that is closer to the thing that matters — an external eye on your specific error. But the asynchronous, once-a-week version of it is a faint shadow of a coach standing next to you adjusting your grip in real time. We are not saying it is worthless. We are saying it is the part of the offering most likely to be oversold, and the part where the membership model is structurally weakest.
Where the volume model quietly fails the 1–2x-weekly player
There is a specific trap for the player this article is written for. If you play twice a week, your total practice volume is small, which means every session is precious, which means the cost of spending it on blocked, flattering, low-transfer repetition is high. The deliberate-practice literature has bad news here that the membership pitch tends to skip: the player with the least time has the least margin for inefficient practice, not the most. Volume-based framing implicitly assumes you have volume to spend. You do not. You have an hour, twice a week, and the question is not how many reps you can pack into it but how much retrievable skill survives until next time.
The part the marketing gets right by accident
We should not pretend the membership pitch is all error. There is one dimension it leans on heavily — belonging — and that dimension is real, even if it is doing something different from what the drill library claims to do.
A person who joins a community of other improving players, who posts a video and gets replies, who has a thread to be accountable to, will, on average, show up more. Adherence is the most underrated variable in the entire conversation. A 2014 review by Rhodes and colleagues on exercise adherence found that social accountability and identity ("I am someone who trains") are among the more durable predictors of sticking with a regimen — more durable than motivation, which is famously fickle. So the warm, peer-to-peer framing — connect, learn, improve alongside others — is not just sentiment. People who feel they belong to a group of players keep playing.
But notice the seam here. Belonging improves adherence. It does not improve transfer. Showing up more is necessary and not sufficient. If the sessions you show up to are blocked, over-fed, groove-chasing sessions, the community has simply helped you do the less-effective thing more consistently. The social layer and the skill layer are two different engines, and a good platform needs both running. Most sell the first and only deliver the second well enough to keep you paying.
An honest rule of thumb
If you are going to use a membership platform — or just structure your own hour — here is what the research supports, stated plainly.
Interleave, don't block. Spend your session on three or four shots rotated in short bursts, not one shot drilled to exhaustion. Accept that it will feel worse and trust that it sticks better.
That single instruction is better supported by the evidence than any specific drill in any specific library. The drills are interchangeable. The ordering is not.
A few more, in descending order of how confident we are:
- Pick two limiting flaws and ignore the rest for a month. Selection beats coverage. Doing four things badly is worse than doing two things deliberately.
- Get feedback sometimes, not constantly. After a cluster of attempts, check the model or the video, then go back to feeling your own errors. Continuous correction builds dependence.
- Treat the community as your adherence tool, not your coaching tool. Use it to make yourself show up. Do not mistake the encouragement for instruction.
Here is the same logic as a table, since the difference between what feels productive and what is productive is the entire point.
| What feels like progress | What the research favors | Why |
|---|---|---|
| 40 forehands in a row, same spot | A few forehands, then switch, then return | Contextual interference: retrieval builds retention |
| Correction after every miss | Correction after a cluster | Spaced feedback prevents dependence |
| A big library, a little of everything | Two targeted flaws for a month | Deliberate practice is narrow and effortful |
| A streak counter and daily reps | Fewer, harder, messier sessions | Volume without difficulty is largely inert |
The verdict, in the publication's terms
The claim that structured tennis drills improve your game is well-established in its weak form: organized, targeted, effortful practice with feedback beats scattered, undirected play. Nobody disputes that.
The claim that a membership platform's value is its volume — its rep count, its library size, its streaks — is folk wisdom dressed as science. It descends from a real result (Fitts and Posner, Ericsson) that has been stripped of the qualifier that made it true. The qualifier was deliberate, and the marketing dropped it on purpose, because "do fewer, harder, more frustrating reps" is a worse sales pitch than "unlock 500 drills."
The claim that community improves outcomes is plausible and partly supported — through adherence, not through skill transfer. Real effect, wrong mechanism in most of the pitches.
Back to the number
So return to where we started: hit a shot ten thousand times and it becomes automatic.
The number was never wrong, exactly. It was hollow. Fitts told us skill becomes automatic when attention is freed; Ericsson told us the hours only count if the attention inside them is deliberate; Shea and Morgan told us that the reps which feel the most automatic in the moment are often the ones that teach the least. Put those together and the ten thousand reps the platform is counting are the wrong reps — smooth, blocked, flattering, and gone by next week.
The number you actually want is much smaller and much harder. A few dozen reconstructed-from-scratch attempts, in mixed order, on the two things that are really holding you back, with feedback you mostly withhold from yourself — that hour, twice a week, will move your game further than ten thousand grooved forehands ever will.
A membership platform can hand you that hour, or it can sell you the counter. Buy the one that admits the difference.