The Gap Between Practice and Performance

A hitter comes into the facility with a clearly identified swing flaw. We assess their swing, establish a plan, and spend several weeks or even months working toward a specific mechanical change.

The progress is there. The hitter begins moving more efficiently, the data is trending in the right direction, and the changes become more and more consistent. Eventually, they can demonstrate the desired change against batting practice or a pitching machine. By almost every measure available to us in training, the intervention appears successful.

Then the hitter returns to competition, and nothing changes. Their production looks remarkably similar to what it was before the training. They struggle with the same issues, or simply fail to produce the improvement we expected.

I've encountered variations of this situation throughout my entire coaching career. It's one of the more frustrating aspects of player development because it forces us to confront an uncomfortable possibility: we might have successfully improved something about the hitter without actually making them a better hitter.

That distinction has become incredibly important in how I think about developing hitters.

What are we actually trying to improve?

In the baseball world, we’ve gotten pretty good at being able to measure things.

We can evaluate batted ball profiles, kinematic sequencing, bat tracking data, and countless other characteristics that were difficult or impossible to quantify not long ago. These measurements provide valuable information about how hitters move and what might be limiting their performance, but having the ability to measure something doesn't necessarily establish its importance.

We can identify mechanical characteristics associated with successful hitters and we can develop interventions intended to improve those characteristics. From there, we can measure whether those interventions produce the desired changes. What becomes more difficult is establishing whether those changes actually improve the hitter's ability to perform in competition.

One of the primary problems is that we have limited evidence showing that deliberately changing a specific mechanical characteristic will improve a hitter's performance in games. We can identify characteristics that correlate with desirable outcomes, but correlation alone doesn't establish that deliberately changing those characteristics will improve an individual hitter.

It's also a problem of training. Even when we've correctly identified an area that needs to be improved, the environment in which we develop a solution can be substantially different from the environment in which the hitter actually needs to use it.

Those are two separate challenges, and both matter a lot.

For me, the purpose of improving a hitter's mechanics has always been to improve their performance in competition. If a hitter is having success in games, I don't care whether the data we’re tracking falls into the ranges we consider ideal.

The mechanical characteristics we track help us understand and develop performance; they aren't the performance objective itself.

The difference between performing a skill and learning it

One of the most important distinctions in motor learning is the difference between performance during practice and learning that persists beyond practice. A hitter might perform exceptionally well in a familiar training environment because they've become accustomed to the task. They know what the pitch speed is going to be, how that pitch is going to move, and the general area where it will be located.

The environment allows them to organize their swing with little to no uncertainty.

That doesn't mean the practice is useless. It may be exactly what the hitter needs at that point in their development, but success in that environment tells us relatively little about whether the improvement will be there when the demands change.

Motor-learning researchers have long distinguished improvements observed during acquisition from retention and transfer. Kantak and Winstein (2012) discuss how performing well during practice doesn’t necessarily mean something has been learned. Looking at whether those improvements last or carry over to different conditions can tell us more. This distinction matters a lot when it comes to hitting because the competitive environment rarely allows a hitter to reproduce the conditions under which they learned something.

Pitch speeds change. Shapes and spins change. Locations change. Even things like lighting and atmospherics change. All of this means timing demands change. The hitter has to perceive all of that information, make a decision to swing or not to swing, and then organize an appropriate movement within a very short amount of time. A movement that seems effective when the hitter knows exactly what to expect may become much more difficult to express when the uncertainty of competition is introduced.

This raises an important question:

Are we developing a movement the hitter can demonstrate, or a capability the hitter can use?

The environment changes the movement

I don't believe training that addresses swing mechanics and representative practice need to exist on opposite sides of a philosophical divide. Traditional drills have demonstrated their value over time. Simplifying a movement or reducing task demands can help a hitter explore a new solution, or develop a physical trait that may be too difficult to address in a more demanding training environment.

The problem arises when success in those drills becomes the primary, or only evidence that the hitter has improved.

Consider a hitter working to change a characteristic of their swing specific to their needs, like counter-rotating less during their load or lowering their attack angle. Against a stationary ball or predictable pitch, they might consistently demonstrate the desired change. But when exposed to different speeds, spins, and angles, their movement begins to vary.

Personally, I don't necessarily view that variation as a problem.

Hitting is highly reactive. How a hitter moves is heavily influenced by the information available to them and the demands of the pitch. Gray (2020) found that improvements in batting performance were associated with changes in how hitters coordinated their swings. They showed more variability in ways that helped them meet the timing demands of the task, and less variability in ways that interfered with it. To me, that’s an important distinction.

What I'm more interested in is whether the hitter can still do what we've been working on when the pitch becomes more challenging. I don't expect every swing to look the same, but I do want to see whether the changes we've made hold up when the hitter has less time to react or has to adjust to something unexpected.

This is one reason I like combining traditional drill work with training environments that more closely resemble competition. We can preserve a specific training objective while exposing the hitter to relevant velocities, spins, and angles. The drill doesn't have to disappear simply because we introduce more realistic demands.

For example, we might use open-stance or closed-stance drill work to address direction or rotation while having the hitter face fastballs from a machine at velocities more representative of competition, rather than relying exclusively on flips.

Pinder and colleagues (2011) discuss this idea in their research on representative learning design. Essentially, we want to preserve the parts of competition that matter for what we’re trying to train, so we can see whether the hitter can use what they’re learning under those demands.

Of course, hitting off a pitching machine isn't necessarily reproducing the entire competitive environment. The hitter may still be missing information from a pitcher's delivery, the uncertainty of pitch selection, or the consequences of an actual at-bat. Representativeness isn't an all-or-nothing thing.

The more useful question is which parts of competition do we want to include, and which parts can we eliminate. We can then determine whether those choices make sense for the qualities we're trying to develop.

More difficult doesn't automatically mean better

There's a temptation to assume that if competition is unpredictable, practice should be as unpredictable and difficult as possible.

That's definitely not what we’re going for.

A practice environment can be difficult without being productive. A hitter who is consistently overwhelmed may have very few opportunities to explore or stabilize the adaptations we're trying to develop. Likewise, an environment can be variable without meaningfully representing the demands of competition.

Research on contextual interference offers an interesting example.

In a 1994 study involving 30 collegiate baseball players, Hall and colleagues compared extra batting-practice sessions with fastballs, curveballs, and changeups presented either in blocks or in a random order. After six weeks, the random-practice group performed better on controlled transfer tests than the blocked-practice group.

That's encouraging evidence for considering practice variability, but the study didn't establish that randomized batting practice improves offensive production during games. And the broader evidence is less straightforward than that one result might suggest. A 2023 review of sport-based studies found no clear overall advantage for random over blocked practice on acquisition, retention, or transfer. A separate 2024 meta-analysis found an overall transfer advantage for random practice, but the advantage in applied settings was smaller and not statistically significant.

I think this is an important reminder that principles from motor learning should inform coaching decisions without becoming universal prescriptions.

Guadagnoli and Lee (2004) discuss something similar in their Challenge Point Framework, which suggests that how difficult a task is depends partly on the skill level of the person performing it.

The appropriate amount of difficulty depends on the hitter and their goals, and the stage of the learning process. A hitter attempting to establish a substantially different movement pattern may benefit from more predictable practice initially. Another hitter who can already demonstrate the movement consistently might need greater variability to determine whether the adaptation remains functional.

The wild part about hitting is that sometimes the opposite is true. A hitter who's just starting to make a swing change might need more variability, while a hitter who's already consistent might benefit from more predictable practice. Those needs can even change within the same session.

The objective shouldn't be to make practice as difficult as possible. It should be to create an environment that gives the hitter an appropriate opportunity to develop the intended adaptation.

What happens when the improvement doesn't transfer?

Suppose we successfully improve a hitter's mechanics in training but their production in games, where it actually matters, remains unchanged.

Perhaps the mechanical characteristic wasn't actually limiting their performance. We identified something that looked different from ideal, but changing it didn't address the problem that mattered most.

Perhaps the change was useful, but the hitter couldn't express it under competitive demands.

Perhaps the mechanical improvement introduced a trade-off. The hitter gained something in one area while losing something elsewhere, or perhaps the improvement is real, but the competitive results haven't reflected it yet.

These possibilities are difficult to distinguish.

At the major league level, stadium-mounted camera systems provide us with ball-flight data, swing-tracking data, and information about how hitters move during games. But even with those measurements, we don't necessarily know whether a mechanical change caused an improvement in performance. Comparable tracking information also isn't consistently available at other levels.

It doesn't mean we should abandon interventions directed at improving mechanical flaws or stop using objective measurements. It just means we should be careful about what those measurements allow us to conclude.

If a hitter improves a swing metric, we've demonstrated an improvement in that metric under the conditions in which it was measured.

If they can maintain that improvement against increasingly challenging pitching, we've learned something more.

If they begin producing better outcomes in competition, we have stronger evidence that the overall development process may be working.

Even then, we need to consider the many factors that influence offensive performance. There isn't always a clean, immediate relationship between an intervention and its intended result.

Developing better hitters

I don't think the gap between practice and performance can be eliminated entirely, especially when it comes to hitting.

Competition introduces demands that are difficult to reproduce perfectly in training. Individual hitters respond differently to the same interventions, and the evidence available to coaches rarely provides the certainty we'd like. Those limitations certainly shouldn't prevent us from improving how we approach development.

First and foremost, we need to be more deliberate about identifying what actually limits performance. Once we’ve established those limitations we can design training environments that develop specific qualities while accounting for the demands of competition. After the training has taken place we can then evaluate progress without assuming that improvements in isolated metrics automatically mean an improvement in hitting.

For me, this entire process is iterative.

Assess the hitter. Identify and prioritize the limitations most likely to affect performance. Design a training intervention. Evaluate how things go. Reassess as new information becomes available.

Sometimes the intervention works exactly as expected (though rarely). Sometimes it changes the intended characteristic without improving performance, and sometimes the results challenge our original understanding of the problem.

All three outcomes can give us some great insight, provided we're willing to examine them.

Ultimately, the objective isn't to develop hitters who are exceptionally good at demonstrating what we've taught them during practice; it's to develop hitters who can use what they've learned when the game demands it.

Research references

• Hall, K. G., Domingues, D. A., & Cavazos, R. (1994). Contextual interference effects with skilled baseball players. Perceptual and Motor Skills, 78(3), 835–841. https://doi.org/10.1177/003151259407800331
• Pinder, R. A., Davids, K., Renshaw, I., & Araújo, D. (2011). Representative learning design and functionality of research and practice in sport. Journal of Sport and Exercise Psychology, 33(1), 146–155. https://doi.org/10.1123/jsep.33.1.146
• Guadagnoli, M. A., & Lee, T. D. (2004). Challenge point: A framework for conceptualizing the effects of various practice conditions in motor learning. Journal of Motor Behavior, 36(2), 212–224. https://doi.org/10.3200/JMBR.36.2.212-224
• Kantak, S. S., & Winstein, C. J. (2012). Learning–performance distinction and memory processes for motor skills: A focused review and perspective. Behavioural Brain Research, 228(1), 219–231. https://doi.org/10.1016/j.bbr.2011.11.028
• Gray, R. (2020). Changes in movement coordination associated with skill acquisition in baseball batting: Freezing/freeing degrees of freedom and functional variability. Frontiers in Psychology, 11, 1295. https://doi.org/10.3389/fpsyg.2020.01295
• Ammar, A., Trabelsi, K., Boujelbane, M. A., Boukhris, O., Glenn, J. M., Chtourou, H., & Schöllhorn, W. I. (2023). The myth of contextual interference learning benefit in sports practice: A systematic review and meta-analysis. Educational Research Review, 39, 100537. https://doi.org/10.1016/j.edurev.2023.100537
• Czyż, S. H., Wójcik, M., & Solarská, A. (2024). The effect of contextual interference on transfer in motor learning: A systematic review and meta-analysis. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2024.1377122