Every pitching coach has seen it. A pitcher spends a month cleaning up a long arm swing, the bullpens look great, and in the first tight inning of the season the old arm action is back. It is tempting to call that a mental problem or a lack of effort. The dynamical systems view of movement offers a more useful explanation, along with a plan.
What is an attractor in pitching mechanics?
Picture every possible way you could throw as a landscape of hills and valleys. A valley is a pattern your body falls into easily and returns to when something bumps it. Scientists call these stable patterns attractors, an idea developed in Scott Kelso's work on self-organizing coordination and applied to sport by Davids, Button and Bennett [1, 2]. Gray explains that attractors make coordination resistant to disturbance: after a bump, you get pulled back into the well [3].
Years of throwing one way carve a very deep valley. That is great for consistency, because a deep attractor is what lets you repeat a delivery thousands of times. It is also why changing it is hard. A new pattern starts life as a shallow valley next to a deep one.
Why does pressure bring old mechanics back?
Under stress or fatigue, a pitcher tends to slide toward the deepest valley nearby, which is usually the old habit. Two lines of research support this.
Pressure changes pitching kinematics. Gray, Orn and Woodman tested experienced baseball pitchers under pressure. Both groups hit their targets less often. Pitchers given only a target showed changes in expertise-related kinematics, while pitchers also shown an "avoid" zone threw more pitches into that zone [4]. Pressure, and what a pitcher is told to look at, changed how they moved.
Explicit rules can break down under pressure. Rich Masters' reinvestment research found that skills learned through explicit technical rules were more likely to break down under pressure, as performers reverted to conscious control of the movement [6]. Gray uses this work to argue that a delivery built mostly on verbal body-part rules may be the most fragile when stakes rise [3].
Coordination research adds detail on why some changes stick more easily. Learners bring their own tendencies, and new patterns either cooperate or compete with them. When they compete, change tends to be abrupt and harder to hold [5, 11]. A change that fights years of habit is exactly the kind most likely to slip when you stop paying attention.
Sliding back under pressure is a predictable property of attractor dynamics. Treat it as information: the new pattern is still shallow, and practice should keep deepening it before the stakes rise.
How do you make new mechanics stick?
A coach can reshape the landscape in two directions at once. A constraint can temporarily fill in the old valley by making the old pattern fail the task, and good practice digs the new valley deeper, rep by rep [3].
Fill in the old valley
The idea is to design a task the old pattern cannot solve. Gray's connection ball for forearm flyout is one example: the pitcher has a single goal for the ball tucked under the arm, and the old pattern fails it [3]. A target the old arm path consistently misses, or a rubber position that breaks the old timing, can do the same job. These are coaching ideas built on the theory, so treat them as experiments and watch the results.
Dig the new valley deeper
Repetition still matters, but the kind of repetition matters more. Rubicon Performance describes the goal as a valley that is deep and has a wide mouth, so you fall into it from many different starting points. In their words, there are many more ways to fall into "a canyon than into a manhole" [7]. A new pattern practiced only from one comfortable setup, at one tempo, on one mound, is a manhole. Seifert and colleagues describe expert movement as combining stable outcomes with the flexibility to reach them from varied conditions [8], and Hristovski's boxing work shows athletes switching flexibly between actions near critical points without being locked into one rigid solution [9].
Practice that tests the pattern
Build Command days into your week
Pitch AI builds dedicated Command days into your throwing program, so the work of deepening a new pattern has a place in your week alongside velocity and recovery days. Every session you log builds your workload progressively.
How to test whether a new pattern is game-ready
A truly learned pattern should survive small bumps. Before you take a change into games, try to knock it out of its valley on purpose:
- Slide step and leg lift. Alternate them within the same bullpen.
- Pitch-clock pace. Work quickly between pitches for a full simulated inning.
- A different mound. Throw off an unfamiliar slope or a portable mound.
- Late in the bullpen. Check the pattern near the end of a session, with mild fatigue, inside your normal workload limits.
If command collapses under small perturbations like these, the valley is not deep yet. Keep building before you take the pattern into games [3, 8]. Our article on why you get worse after changing your mechanics covers how long that building phase can take.
A staged path from bullpen to game
Because pressure pulls toward the old habit, raise the stakes gradually. A practical ladder looks like this: drills, then bullpens, then bullpens with a stand-in hitter and called counts, then live at-bats and simulated games, then low-leverage innings, then high-leverage innings. Move up a rung only when the pattern holds at the current one. This ladder is our application of the research and has not been tested as a protocol. It follows directly from the idea that a shallow valley needs to deepen before it can resist big bumps.
Representative practice also helps. A bullpen with a hitter in the box and counts on the line exposes the new pattern to some of the information it will face in a game, which is the subject of perception-action coupling in pitching, the next part of this series.
What to think about on the mound
In games, keep cues outcome-focused. "Hit the glove" and "finish the count" give your attention somewhere useful. Body-part rules like "keep your elbow up" invite the kind of conscious control that Masters linked to breakdown under pressure [6]. And give yourself a target instead of an area to avoid: in the Gray, Orn and Woodman study, the avoid-zone instruction backfired [4].
The Rubicon podcast host, a former yips sufferer, frames the extreme version of this problem as losing your relationship with ball flight [10]. That is a practitioner's view, and it fits the research above: the way back runs through reconnecting your attention to where the ball goes.
Respect the habits that work
Not every deep valley is a problem. A stable, effective delivery is an asset, and digging up a pattern that is getting outs has a cost. Only destabilize a habit when there is a clear performance or health reason. Our guide to pitching mechanics that actually matter can help you decide whether a change is worth it, and how to progress a throwing program shows how to fit the work into a sensible week.
Frequently asked questions
Why do my mechanics fall apart in games but not in bullpens?
A newly learned pattern is a shallow attractor. Under pressure or fatigue, movement tends to slide back to the deepest nearby pattern, which is usually the old habit. Research on experienced pitchers found pressure changed their kinematics.
How long until new mechanics hold up in games?
There is no fixed timeline. A good sign is that the pattern survives small perturbations in practice, such as slide steps, pitch-clock pace, a different mound and throwing late in a bullpen. If command collapses under those, keep building.
Should I think about mechanics during a game?
Keep game cues on the target and the outcome, such as hitting the glove. Research on reinvestment suggests explicit body-part rules are more likely to break down under pressure, and avoid-zone instructions backfired in a pitching study.
Is going back to old mechanics a mental weakness?
No. Sliding back under pressure is a predictable property of how movement patterns stabilize. It means the new pattern is still shallow and needs more varied, game-like practice.
Make the new pattern stick
Train the pattern you want to see in games
Pitch AI builds dedicated Command days into your throwing program and logs every throwing session, so you can stage a new pattern from bullpens toward game-like work without spiking your workload. The daily Wellness Check-In tracks sleep and readiness, two of the things that make old habits come back.
Sources and further reading
- Kelso JAS (1995). Dynamic Patterns: The Self-Organization of Brain and Behavior. MIT Press.
- Davids K, Button C, Bennett S (2008). Dynamics of Skill Acquisition: A Constraints-Led Approach. Human Kinetics.
- Gray R (2021). How We Learn to Move: A Revolution in the Way We Coach & Practice Sports Skills. Perception Action Consulting & Education. Chapters 6 and 7.
- Gray R, Orn A, Woodman T (2017). Ironic and reinvestment effects in baseball pitching: how information about an opponent can influence performance under pressure. Journal of Sport & Exercise Psychology 39(1):3-12. https://doi.org/10.1123/jsep.2016-0035
- Kostrubiec V, Zanone P-G, Fuchs A, Kelso JAS (2012). Beyond the blank slate: routes to learning new coordination patterns depend on the intrinsic dynamics of the learner. Frontiers in Human Neuroscience 6:222. https://doi.org/10.3389/fnhum.2012.00222
- Masters RSW (1992). Knowledge, knerves and know-how: the role of explicit versus implicit knowledge in the breakdown of a complex motor skill under pressure. British Journal of Psychology 83(3):343-358. https://doi.org/10.1111/j.2044-8295.1992.tb02446.x
- Crider B (2026). Skill Adaptation. Rubicon Performance, July 25, 2026. Practitioner article. https://www.rubicon-performance.com/education/articles/skill-adaptation/
- Seifert L, Button C, Davids K (2013). Key properties of expert movement systems in sport: an ecological dynamics perspective. Sports Medicine 43(3):167-178. https://doi.org/10.1007/s40279-012-0011-z
- Hristovski R, Davids K, Araújo D, Button C (2006). How boxers decide to punch a target: emergent behaviour in nonlinear dynamical movement systems. Journal of Sports Science and Medicine 5(CSSI-1):60-73. https://www.jssm.org/jssm-05-CSSI1-60.xml
- Rubicon (2026). Casting The Die, Episode 2, with Steffen Simmons. YouTube, September 19, 2026. Practitioner podcast. https://youtu.be/PkRCz3dhYn0
- Zanone PG, Kelso JAS (1992). Evolution of behavioral attractors with learning: nonequilibrium phase transitions. Journal of Experimental Psychology: Human Perception and Performance 18(2):403-421. https://doi.org/10.1037/0096-1523.18.2.403