The Physics behind the Flipo Flip Toy
Authors: Patrick Chen, Yu-Hsuan (Niels) Chien, En-yu (Lucien) Lo
The three authors contributed equally to this work and share first authorship.
Editor's note
This work investigates “Flipo Flip,” Problem 6 of the 2026 International Young Physicists’ Tournament (IYPT). The original problem statement, available on the official IYPT website, is:
A Flipo Flip toy can roll for multiple turns even though its shape is not circular. Investigate how its motion depends on parameters such as geometry and the initial release conditions.
Abstract
We investigated the motion of a Flipo Flip, a non-circular toy that can perform repeated flips after an initial push. A golden-colored metal toy was tested on a wooden shelf, and phone recordings were analyzed using Tracker. Representative light, medium, and hard pushes produced incomplete flipping, repeated flipping, and irregular motion with an apparent bounce, respectively. We also examined the energy, geometry, and surface comparisons shown in the figures. The observations suggest that repeated flipping depends on launch conditions and surface contact, rather than translational speed alone. The results remain qualitative because force-time histories, trial counts, and measurement uncertainties are not reported due to difficulties with the experimental setup.
Keywords: Flipo Flip; rigid-body motion; video analysis; surface friction.
Introduction
The study of rotational motion is essential in understanding how forces influence the movement of objects in physics and engineering. From the design of mechanical components to the operation of rotating tools and machinery, the relationship between applied force, shape, and rotational motion plays an important role in determining how an object moves. In this research, we investigated the motion of a Flipo Flip, a metal toy capable of flipping despite its non-circular shape. The toy consists of three parts, including two symmetrical components with a two-dimensional profile resembling a bicylinder. We chose this experiment because we were curious about how an object with an unconventional shape could rotate and how different magnitudes of applied force would affect its motion. Most people are familiar with circular objects, such as wheels, which rotate in a relatively predictable manner. However, a Flipo Flip presents a more unusual case because its shape may influence how its mass is distributed and how it interacts with the surface during motion. This raises the question: How does the magnitude of an applied force affect the motion cycle of a non-circular object such as a Flipo Flip? To investigate this question, we applied three different levels of force - soft, normal, and hard - under consistent experimental conditions and used a force gauge and Tracker video analysis software to collect and analyze data. By comparing the results, we aimed to understand the relationship between applied force and the rotational behavior of the Flipo Flip.
Experimental Setup and Methods of Analysis
In the Flipo Flip experiment, we used a phone to record the motion of the toy, a force gauge to measure the applied force, a golden-colored metal Flipo Flip (41 mm high when standing and 13 mm high when lying down), a wooden shelf as the testing surface, and Tracker video analysis software to analyze the recorded videos. These materials helped us collect accurate data and observe how different amounts of applied force affected the motion of the Flipo Flip.
For the geometry comparison, we used three 3D-printed Flipo Flip prototypes with the same overall height of 40 mm. The circular arcs on the two side profiles subtended central angles of 60°, 120°, or 180°, respectively; the 180° arcs were semicircles. The curvature radii were adjusted for each configuration while keeping the overall height fixed. These angle labels refer to the central angles of the circular arcs, not the toy's orientation during motion.

For the force comparison, we used the same golden-colored metal Flipo Flip for every trial. We also placed it on the same wooden shelf each time to keep the testing surface consistent. Using the same toy and surface helped reduce the effects of other variables that could influence the results. We recorded each trial using a phone, making sure that the motion of the toy could be clearly seen in the video. The videos allowed us to review the movement more carefully instead of relying only on our observations during the experiment.
Before each trial, we used a force gauge to measure the force applied to the Flipo Flip. We then observed and recorded how the toy moved after the force was applied. By conducting multiple trials, we could compare the results and identify patterns in the toy's motion under different conditions. After recording the experiments, we imported the videos into Tracker video analysis software. Using Tracker, we tracked the position of the Flipo Flip frame by frame over time to obtain quantitative motion data. These measurements helped us analyze the toy's movement and calculate important quantities, such as its velocity.
For the energy calculations, we used the center-of-mass height and speed obtained from Tracker. We calculated gravitational potential energy as U = m*g*h꜀ₘ, where m is the toy's mass, g is the acceleration due to gravity on Earth, and h꜀ₘ is the height of the toy's center-of-mass. We calculated the translational kinetic energy as Kₜ = 0.5*m*v꜀ₘ², where m is the toy's mass and v꜀ₘ is the magnitude of its center-of-mass velocity. Their sum U + Kₜ represents the translational kinetic energy plus gravitational potential energy, rather than the total mechanical energy. Rotation about the center of mass contributes an additional kinetic-energy term, Kᵣ = 0.5*I꜀ₘ*ω², where I꜀ₘ is the toy's moment of inertia about its center of mass (around the relevant axis of rotation) and ω is its angular velocity. This contribution cannot be determined from center-of-mass speed alone [1].
Finally, we organized the collected data so that the results from different trials could be compared. By keeping the materials and testing surface consistent, measuring the applied force, recording the motion, and analyzing the videos with Tracker, we aimed to make the experiment as fair and reliable as possible. This method allowed us to investigate how the applied force affected the motion of the Flipo Flip using measurable evidence rather than observations alone.
Figure 2 shows the experimental setup and an example of the video analysis in Tracker [3, 4].

Three Types of Motion
Figures 3 - 5 plot the motion of the Flipo Flip under three initial force conditions using a dual-axis format. The x-axis represents time in seconds. The primary y-axis (left) tracks the center-of-mass height in centimeters (red line with circular markers), while the secondary y-axis (right) tracks the measured velocity in meters per second (dashed blue line with cross markers).
Analysis of the Three Motion Types:
Light Force (Failed Flipping): The recorded velocity reaches 0.5 - 0.6 m/s, while the plotted height drops from 3.2 cm to 0.65 cm by 0.13 seconds. This suggests that the available energy is insufficient to sustain rotation. After 0.30 seconds, height fluctuations shrink, and the toy gradually settles mostly below 0.5 cm, failing to establish a continuous rolling cycle.

Medium Force (Successful Flipping): In this trial, the recorded velocity reaches approximately 0.72 m/s and supports repeated flipping. The toy enters a sustained rhythm, completing approximately five distinct height oscillations. The center of mass repeatedly peaks at 1.9 - 2.1 cm with a regular period of 0.17 - 0.20 seconds between major peaks. The repeated oscillation continues until 0.80 seconds before losing amplitude.

Hard Force (Failed Flipping and Bounce): In this trial, the recorded velocity reaches approximately 0.7 m/s, but the launch appears to disrupt the flipping motion. The height drops sharply to 0.04 cm near 0.12 seconds and immediately spikes to 1.67 cm, consistent with the observed bounce rather than a regular flip. By 0.20 seconds, the motion becomes irregular, and subsequent movement is reduced to minor settling fluctuations below 0.3 cm.

Motion Comparison Summary
The data suggest that successful operation depends on a suitable range of applied force. An inadequate push may not supply enough energy to raise the center of mass through another flip, while a hard push may lead to loss of surface contact and energy loss during impact. Future iterations of this experiment would benefit from recording exact force applications in newtons (N) together with push duration and application point to determine a reproducible launch condition.
Energy Conversion and Curvature
Figure 6 shows how gravitational potential energy and translational kinetic energy change as the Flipo Flip moves. These changes provide a qualitative view of the energy associated with the center-of-mass motion during flipping. Potential energy is given by U = m*g*h꜀ₘ, while rigid-body kinetic energy includes both translation and rotation [1]. These contributions must be distinguished when interpreting an energy plot.

Video analysis was used to examine changes in height and motion. The geometry comparison used the three 3D-printed prototypes described in the Methods, separately from the main experiments with the golden metal toy. The external dimensions of the golden toy were approximately 41 mm when standing and 13 mm when lying down; these dimensions should not be confused with center-of-mass height.
As the toy rises, gravitational potential energy increases; as it falls, that energy can be converted into translational and rotational motion. Surface impacts and sliding can dissipate mechanical energy. The torque about the center of mass also depends on the contact forces and their lever arms [2].

The experiment also investigated how curvature affected the Flipo Flip's motion by comparing the three 3D-printed configurations with side-arc central angles of 60°, 120°, and 180° at a fixed overall height of 40 mm. Figure 7 shows the center-of-mass height over time for these three angle configurations. The 60° configuration displays the largest downward changes in height, falling to approximately 0.024 m, while the 120° configuration shows moderate variations. The 180° configuration generally maintains a higher center-of-mass height, reaching approximately 0.040 m, with fewer extreme downward drops. The 180° configuration shows smaller downward excursions in this recording, whereas the 60° configuration shows greater fluctuations. The differences may be related to how the curvature affects the toy's rotation and contact with the surface. However, height variations alone cannot fully establish stability, so additional measurements, such as the number of successful flips and the consistency of flipping time, would help confirm this conclusion. The central angle and curvature radius varied together under the fixed-height constraint, so this comparison does not isolate the effect of either geometric parameter independently.
By comparing the motion under different force and curvature conditions, the experiment aimed to understand how these factors influence the flipping behavior of the Flipo Flip. The video analysis provided quantitative evidence for examining the relationship between applied force, curvature, and motion.
Surface Friction and Maximum Velocity of Flips
Surface friction is an important factor that can affect the motion and maximum velocity of the Flipo Flip. When the Flipo Flip moves across the wooden shelf, contact forces act between the toy and the surface, affecting how it slides, rotates, and flips. The amount of friction depends on the contact between the two surfaces and the type of motion occurring. Sliding friction can transfer some of the toy's mechanical energy into thermal energy, which may reduce its speed. However, the effect of friction on maximum velocity also depends on how the applied force causes the toy to move and rotate. In ideal rolling without slipping on a stationary rigid surface, static friction does not dissipate energy at the contact point.

The main velocity comparison used the same golden-colored metal Flipo Flip and wooden shelf, with phone recordings analyzed in Tracker. Figure 8 presents an additional comparison of static-friction coefficients for four toy variants on seven surfaces. The plotted coefficients vary with both the toy and the surface, so these results should be distinguished from the main launch comparison.

The relationship between friction and maximum velocity is not always straightforward because the Flipo Flip can move in different ways, including sliding, rotating, and flipping. Friction may oppose sliding, but it can also provide the torque needed for rotation. Therefore, friction does not always simply reduce the toy's speed; its effect depends on the type of motion. Analyzing the position and velocity data helps us better understand these changes and explain why the Flipo Flip may reach different maximum velocities under different conditions. Figure 9 compares measured maximum initial velocity and flip count on the shelf, notebook, and table. The panels use different velocity scales, and higher speed does not consistently correspond to more flips. The shaded contours summarize the distribution of measurements; their extension below zero is a smoothing effect rather than a physical negative flip count. These comparisons alone do not isolate friction from other differences between surfaces.
Discussion and Limitations
The results suggest that repeated flipping depends on both the initial push and the toy's contact with the surface. However, the applied force may have varied between trials, and video tracking may introduce errors in position and velocity. The number of trials and measurement uncertainties should be reported to assess the reliability of the results. In addition, height variations alone cannot confirm stability, and translational velocity alone cannot determine the toy's total kinetic energy. More repeated trials with controlled pushes and orientation tracking would help clarify these relationships. The height data and tracked point also need to be defined before the plotted height can be interpreted quantitatively as the center-of-mass height. The light, medium, and hard labels are qualitative because numerical force-time histories are not reported here. The energy analysis is a simplified description based on center-of-mass motion. Because rotational kinetic energy is not included, the plotted curves do not provide a complete mechanical-energy balance, and changes in their sum cannot be attributed solely to dissipation.
Conclusion
This investigation provides an experimental answer to our main question: How can the launch condition, geometry, and surface contact be adjusted to obtain more flips? The representative trials show that pushing harder does not necessarily produce more flips. The medium-push trial produced repeated flipping, whereas the light-push trial did not sustain the motion and the hard-push trial showed irregular motion with an apparent bounce. A useful direction for increasing the flip count is therefore to find a launch condition that initiates repeated rotation while avoiding the observed failure modes, rather than simply maximizing the applied force or translational speed.
The motion can be interpreted as a coupling between translation, rotation, and changes in center-of-mass height as the toy moves through successive surface contacts. During planar motion without slipping, the contact point acts as an instantaneous center of rotation. As the center of mass rises and falls, gravitational potential energy is exchanged with kinetic energy, while sliding and impacts can dissipate energy. The three 3D-printed geometries, all 40 mm high, produced different height trajectories. The 180° side-arc configuration showed smaller downward excursions than the 60° configuration in the displayed recording. This comparison demonstrates that geometry changes the path of motion and the associated gravitational-potential-energy variation. It does not establish a numerical fraction of energy dissipated by each geometry, because the plotted energy excludes rotation about the center of mass.
The surface comparisons also show that friction depends on both the surface and the toy variant, and that higher measured initial velocity does not consistently produce more flips. Measuring friction therefore helps characterize the contact conditions relevant to selecting a suitable push and avoiding sliding or irregular motion. Friction measurements and launch observations must be considered together; the static-friction coefficient alone does not determine a unique required force, and an overly strong push may enter a failure regime rather than improve performance.
The contribution of this study is to connect observable flipping outcomes with launch conditions, geometric design, and surface contact in a single experimental investigation. The recorded trajectories, geometry comparisons, friction measurements, and velocity–flip-count distributions provide a practical reference for operating the toy and selecting combinations of shape, surface, and push that favor repeated flipping. They also provide an experimental basis for other students and researchers investigating similar non-circular rolling and flipping systems.
Acknowledgments
The authors thank Hsuan-Yu Yeh (National Yang Ming Chiao Tung University), their senior mentor, for guiding them through the experimental exploration and preparation for this project.
References
[1] Moebs, William, et al. “10.4 Moment of Inertia and Rotational Kinetic Energy.” University Physics Volume 1, OpenStax, 19 Sept. 2016, https://openstax.org/books/university-physics-volume-1/pages/10-4-moment-of-inertia-and-rotational-kinetic-energy.
[2] Moebs, William, et al. “10.7 Newton’s Second Law for Rotation.” University Physics Volume 1, OpenStax, 19 Sept. 2016, https://openstax.org/books/university-physics-volume-1/pages/10-7-newtons-second-law-for-rotation.
[3] Open Source Physics. “Getting Started.” Tracker Help, https://opensourcephysics.github.io/tracker-website/help/gettingstarted.html. Accessed 18 Oct. 2026.
[4] Open Source Physics. “Tracker Video Analysis and Modeling Tool for Physics Education.” Tracker, https://opensourcephysics.github.io/tracker/. Accessed 18 Oct. 2026.
AI Disclaimer
ChatGPT (OpenAI) was used to assist with language editing, organization, figure captions, and drafting the abstract and discussion from the supplied manuscript and figures. AI assistance was not used to generate experimental measurements. The authors remain responsible for checking the final wording, interpretations, references, and scientific content before submission.
Author Information and Biographies
Author 1: Patrick Chen
Email: patrickcheeeen@gmail.com
Affiliation: Hsinchu American School
Bio: Patrick Chen is a student at Hsinchu American School. He is interested in physics and enjoys understanding physical concepts through experiments.
Author 2: Yu-Hsuan (Niels) Chien (submitting author)
Email: niels.jian@gmail.com
Affiliation: Hsinchu American School
Bio: Niels Chien is a student at Hsinchu American School. He enjoys exploring physics and wants to deepen his understanding of science and engineering.
Author 3: En-yu(Lucien) Lo
Email: lo.enyu100@gmail.com
Affiliation: Hsinchu County American School
Bio: Lucien Lo is a student at Hsinchu County American School. He enjoys exploring different areas of physics, particularly sound, and is also interested in conducting experiments and learning more about engineering.

.png)