The Backswing:
Part 1

Brian Gordon, PhD

In my last article I covered the first phase of the serve, the wind up. Now let's move on to the second phase, the backswing. It may look simple, but far and away, the backswing is the most difficult phase of the serve to comprehend. So let's examine its components and see what makes it so complex.

We'll also outline the conditions for maximizing the backswing in your motion. Since there are a lot of issues to cover, we'll analyze the backswing in two parts: part 1 in this article and part 2 to follow in the next issue.

Before we start, let's review what we've established so far. First we divided the serve into 4 parts or Phases.

In our previous article on Phase 1 or the Wind Up, we looked at various techniques for both the upper and lower body motions. We saw that there were pluses and minuses, but that varied techniques could lead to success in the later phases.

First, we saw that the path of the arm and racquet in the wind up can be viewed on a continuum from a pendulum, or semi-circular wind up, to a fully abbreviated motion.

Second we saw that there are various possible footwork patterns. There are two versions of the "platform" stance, either narrow or wide. There are also two versions of the "pinpoint" stance, either standard or lateral.

We saw that the choice of footwork has implications for several important biomechanical characteristics of the overall motion. These include the direction of the ground reaction force, the ending position of the server's center of mass, the generation of forward angular momentum, and the timing of the leg flexion and extension. (Click Here.)

The first article ends by describing several "conditions" or "benchmarks" that are important at the end of the windup. My conclusion was that all the options had potential to be effective if executed correctly, assuming that they were within physical capabilities of the player.

There are many ways to skin a cat, and in my coaching experience I can and have achieved higher benchmarks for different players with differing techniques. Making these determinations is where the data gives way to the art of coaching.

The Backswing

Now let's move forward to the second phase of the serve: the back swing. What are the conditions or benchmarks you want to achieve as a server at the end of the backswing phase? As an example, we'll use the serve of the same junior player we saw in the first article. Again, you can access the 3-D data on his motion through the interface below.

As I define it, the back swing begins at the start of the looping motion with the racquet, that is, when the racket starts to move downward, or possibly forward, toward the racket drop. The backswing concludes when the racquet face center reaches its lowest point, or the point closest to the court.

Determining the exact start of the backswing, or the transition point from the wind up to the backswing, is somewhat subjective. In the animation, Andy Roddick for example, looks very different at the start of his backswing compared to a player with a traditional motion like Mario Ancic.

But determining the end of the back swing is more concrete. This is because we can quantify the lowest point of the racquet path with our 3D measurements.

Why is the backswing the most complex phase? In the other phases, the actions of the upper and lower body are independent, at least to some extent. No such separation is possible in the back swing. The actions of the upper and lower body movements must be coordinated very precisely.

By this I mean coordinating the timing of the movement of the arm and racket and the timing of the leg drive. Our research shows that to maximize the benefit for the server, the arm and racket motion and the leg drive should start and end at the same time. This means that they must also have the same duration. This close synchronization is characteristic of all high level service motions.

The coordination of the backswing and leg drive is difficult to verify with the naked eye, or even with conventional video. If we look at the stick figure of our junior player in the interface, it may appear that his timing matches the top servers. But, if we look at the data, it reveals there is actually a discrepancy that we can measure.

For our player, the data shows the backswing and leg drive are slightly out of sync, with the timing of the arm and racquet motion slightly ahead of the leg drive. In other words, arm and racquet motion starts before the leg drive, and ends while the leg drive is still ongoing. The problem here is that if the arm motion starts before the leg drive, the player may not fully utilize the push against the ground in contributing to the early speed of the hitting arm and racquet.

This type of measurement is one of the advances in coaching that quantitative analysis can offer. It provides a very precise benchmark that players strive toward, plus a concrete way to validate when they have achieved it.

The question for our player, and many others with similar timing problems, is what causes the leg and arm to go out of sync? And how can it be corrected? This is where the complexity sets in.

To answer this we have to go back to something we addressed in Phase 1. This is the nature of the transition between the wind up and backswing phase.

In the last article we saw that the transition between the phases could be either a continuous transition or a hesitation transition. In a continuous transition, the racket moves from the wind up to the backswing at continuous or at an accelerating pace. With the hesitation transition, the racket slows down or even hesitates.

Either transition can be fine. It's a matter of how the transition affects the coordination of the timing of the arm and racquet with the timing the leg drive.

Timing problems here can propagate into the upward swing. This ripple effect is a real danger with some abbreviated motions in which the arm and racquet are positioned lower to the hitting side of the body at the start of the backswing.

We saw in the last article that our player had a continuous transition. But in his specific case, we found his transition caused his hitting arm to outrun the leg drive. His racquet is moving into the backswing at a speed which makes it difficult for him to synchronize his arm with the leg drive.

For our server, luckily, this timing issue is a relatively minor problem since he is only slightly out of phase. It can be solved by a slight adjustment in the speed of the racket at the end of the wind up or by manipulating the leg drive timing.

The timing problem can be more severe, however, when the leg drive ends before the arm and racquet reach their end of back swing position. In this case the influence of the leg drive on the hitting arm motion is reduced significantly.

This problem can be an unintended consequence when players adopt an extreme abbreviated windup on the model of Andy Roddick. The reason is the racquet must travel further from the start of the abbreviated backswing position. Roddick is able to pull this motion off, but if the player is unable to complete the backswing on time, the leg drive will run ahead of the racquet and arm.

The most direct result of this problem is that the latter part of the racquet drop occurs while legs are no longer pushing against the ground. The consequence is a loss of racquet head speed and a decrease in the depth of the racquet drop at the end of the back swing. The exact relationship between these benchmarks and the leg drive will be covered in the next article.

The Three Roles of the Leg Drive

So, again, maximizing the influence of the leg drive requires precise timing between the motion in the upper and lower body. Now let's get more specific about the nature of these upper and lower body motions and what they should accomplish from a mechanical perspective.

We'll start with the lower body or legs and address their role in the remainder of this article. Then we'll move on to the upper body action in the second backswing article next month.

The leg drive has three primary roles during the back swing. These are: the acceleration of the body in the vertical direction; the creation of forward angular momentum, and assistance in rotating the hips.

Vertical Acceleration of the Body

The role of the legs in vertically accelerating the body is the easiest concept to understand. If you push down on the ground with the legs, your body's center of mass will rise vertically in response. For most servers this means you will leave the ground.

The vertical acceleration of the body is proportional to the size of the vertical component of this ground reaction force. The stronger the leg drive, the greater the acceleration. (Assuming that the ground force is acting in the same direction in both cases.)

The ground force can be measured in terms of pounds of force. We can then compare this to the body weight of the player. Based on our database of elite servers, we believe that a ground force equal to twice the bodyweight of the player is a minimum goal. High level servers can generate ground force values between 2.5 -- 2.8 times their own bodyweight.

Again, this is where quantitative analysis can help players by supplying a measurable bench mark, in this case, a way to evaluate how changes in technique are affecting the amount of ground force the player generates.

So how is our subject doing? The estimate of the size of the vertical ground force he is generating can be found in the "Data Options" under "Angular Momentum". The table below the graph shows the highest measured value during the leg drive on line 3.

Our player is generating a ground force equal to 1.9 times of his bodyweight. So that value is slightly below our minimum.

In the wind up article, I explained the foot placement adjustment we made with this player. The goal of this adjustment was to attempt to elicit more vertical ground force reaction.

Specifically, I mentioned that we asked him to keep his feet closer together during his set up and leg drive (narrow platform). This suggestion was based on the intuition and experience of our coaches, and resulted in an improvement in the ground force profile, something we were able to measure and share with our player.

This change in the stance, combined with the proper timing and the amount of his leg flexion and extension, was the best technical solution for this player at this time. Since his value is still somewhat low, our belief is that more improvement will require power training in the weight room.

Why is this vertical acceleration so important? First, this acceleration can lead to a higher contact point, which improves the chances of directing a high speed serve into the service box.

But a second important reason is that the vertical acceleration of the body is linked to vertical acceleration of the hitting shoulder joint. This generates forces which are applied to the hitting arm and which turn out to be critical for an effective racquet drop. More on this in the next article.

Forward Angular Momentum

The second role of the leg drive is in the creation of forward angular momentum. Forward angular momentum means body rotation around an axis passing through the center of mass and parallel to the baseline. An important attribute of body angular momentum is that it can be redistributed to the various segments along the biomechanical chain, potentially increasing the speed of rotation of these segments about the same axis.

The most important redistribution of this momentum early in the back swing is into the trunk. This results in forward rotation of the trunk relative to the legs. This rotation may also be enhanced by the backward lean of the trunk at the end of the wind up. The trunk rotation at this point is often described as a "cartwheel" rotation because the body and trunk are generally oriented sideways to the net.

Our measurements show that the direct contribution of body forward angular momentum to racquet speed at contact is a modest 5 -- 10 %. However, there is a major indirect benefit. This is because the creation and redistribution of forward angular momentum has a beneficial influence on the segmental sequencing of the motion in the upper body, as we'll see in a later article. For this reason it is important to generate as much forward angular momentum as possible.

The lion's share of forward angular momentum generated during the serve is generated during the back swing. As the player pushes on the ground with the legs, a reactionary ground force then pushes back on the body. If this force can be directed behind the body center of mass (as seen from the side along the baseline), the player adds more forward angular momentum to that generated already during the wind up.

The direction of the ground force, the size of the ground force, and the position of the body center of mass during the leg drive are factors that determine how much angular momentum is generated. But to a great extent, the player's ability to generate these factors in the backswing is also related to what happened in the previous or first phase, the wind up.

We can see this in the data interface. By clicking "Angular Momentum" in the "Data Options" section of the left margin, you can access a line graph of body forward angular momentum. (You may notice I added the most recent measurements to the data as well.)

The graph shows that the steepest part of the curve occurs during the leg drive in the back swing. This curve increases to a maximum value of 16.4 units of angular momentum for our young player. If you look at the table below the graph you will notice a maximum goal value of 40 units. This is based on an average sized male adult, however. So if we scale the corresponding value to the size of the junior, the goal for our player should be about 20 units.

Obviously then we would like our player to generate more forward angular momentum, and this is something we continue to work to achieve. Here again we see the invaluable role of hard data in evaluating his current status and his potential for improvement.

If we look at our data over time, we can see how our coaching efforts have improved this benchmark to date. You can do this by clicking "Compare Trials" and selecting the next most recent date (11/11/06).

Now click the "Angular Momentum Graphs" (top left). This displays the comparison in body forward angular momentum. We were encouraged to see an increase, but it is more important to understand how and why this occurred.

By clicking "Vertical Ground Reaction Force" (bottom) it is apparent that size of the push against the ground was stronger throughout the back swing.

Based on the stick figures the foot positioning is similar and therefore it is also reasonable to assume that the direction of the ground force is similar between the two measurements.

This improvement can be attributed to decisions on stance. The narrower platform stance influenced the direction of the ground force by making it more vertical. We stayed with a version of the platform stance because of the improved effectiveness of the muscle contractions when extension immediately follows the flexion, compared to the pinpoint options.

Another reason for narrowing the platform stance was to allow the center of mass to be more forward. Click on "Enter Table Mode" (top left) and then "Center of Mass Position Data" (bottom) and look at line 3 of the table. You can that his body center of mass moved further forward by about 4.5 inches. These two factors--the improved vertical ground force and the forward movement of the center of mass--explain the increased forward angular momentum.

Hip Rotation Assistance

The third contribution of the leg drive is to aid hip rotation. The dominant angular momentum component on the serve is forward. We need to keep in mind, however, that rotation about the other axes is also important. Of particular interest is rotation about the axis passing through the center of mass that is perpendicular to the court. I will refer to momentum around this axis as "spin momentum".

Like forward angular momentum, spin angular momentum is also generated in the body by pushing on the ground with the legs to elicit a horizontal ground reaction force component. This spin angular momentum can also be redistributed to the other body segments. An important portion of this redistribution is to the hips, creating spin rotation of the hips.

Rotation of the hips on the serve is important primarily because this in turn impacts the rotation of the upper trunk or shoulders. Rotating the hips increases the overall range of shoulder rotation. Rotation of the hips also allows the upper trunk rotating muscles to contract in slower more favorable conditions.

The animation illustrates how this is accomplished. Both diagrams show 90 degrees of rotation by the shoulders. But in the first diagram, there is no hip rotation. In the second diagram the hips rotate 60 degrees and the shoulders rotate an additional 30 degrees relative to the hips.

This is still a total of 90 degrees of shoulder rotation, but in this second case, two thirds of the shoulder rotation is coming from the rotation of the hips.

If these rotations occur in the same amount of time, the upper trunk rotating muscles would be responsible for all 90 degrees of rotation in the first case. But in the second case, this is only 30 degrees. This indicates that the contraction speed in the first case was much faster. We know this because the muscles generated the additional 60 degrees of rotation in the same time internal.

In biomechanics, a basic fact is that a muscle produces less force with a faster contraction speed. So it is evident that from a mechanical perspective hip rotation is important. It allows the muscles that rotate the shoulders to contract at a more optimal speed.

The question becomes how do servers rotate the hips in order to do this? The diagram shows all the factors that go into hip rotation towards contact. Generally these include contracting the muscles crossing the hip joints, or using leg extension to force the pelvis to rotate.

Although more research is still needed in this area, my belief is that the back leg extension plays the major role in generating this hip rotation by forcing the pelvis to turn. It does this by pushing the right side of the pelvis forward. There are two different pieces of evidence we can examine that support this conclusion.

The first piece of evidence is that the center of mass is located more over the front foot than the back foot during the back swing. Because of this, the horizontal component of the ground force coming from the back foot will tend to be much more effective in creating spin momentum of the body, and therefore rotation of the hips. This is because it located further from the center of mass.

The second piece of evidence can be found by using the data interface of our sample junior on the page of the latest measurement. Study this by selecting "Leg & Trunk Data" from the "Data Options". The line graph (white line) shows the rotational speed of the hips. The bars on the left show the knee angle of the front (blue) and back (purple) legs where 180 degree is straight.

The data shows that the greatest increase in hip rotation speed occurs during the back swing phase. This increase occurs as the legs begin to extend from roughly the same angle (95 degrees for one leg and 96 degrees for the other.)

But the data shows that by the end of the back swing the back leg has extended more. Since the time of extension is the same, it follows that the back leg has extended more quickly. This is what causes the back foot to leave the ground before the front foot, something that can also be qualitatively verified by the leg action clips on the serve in the Stroke Archive. (Click Here.)

Why would the back leg extend more quickly? One possible explanation is that the server simply pushed harder with the back leg than the front. Another possible explanation is that the push was of equal strength but that the back leg was pushing against less resistance.

This would be the case if more of the back leg push was being used to generate spin momentum. The reason this is possible is that the body is much less resistant to rotation (inertia) around the "spinning" axis compared to the other axes.

Put simply, it is easier to spin the body and hips than it is to rotate the entire body forward and over in the "cartwheel" rotation. The back leg is pushing against the hips, encountering less resistance, and therefore extending more quickly and leaving the ground sooner.

Of course the time interval between the back foot and front foot leaving the court varies among the pro players. But I believe that the greater the premium placed on hip rotation speed in the motion of a particular player, the earlier the back foot will leave the ground relative to the front foot.

The role of the back leg push in generating hip rotation can be enhanced by changes in the stance, by placing the foot further behind the front foot in a wide platform, or more to the side of the front foot in a lateral pinpoint.

Does this mean that extensive hip rotation from use of the rear leg is the necessary key to a more effective serve? Not at all. And remember, we took the opposite tack with our junior player in narrowing his platform stance.

Hip rotation is only one of the myriad, interrelated biomechanical factors in elite serving. With our player we chose a different stance option to increase forward angular momentum, probably at the expense of spin angular momentum.

Our research shows that there are trade offs with many technical options in each of the phases of the serve. Quantitative analysis provides us with an evaluation process of the pluses and minuses of all these options for a given server.

This gives coaches the courage to follow their intuition because they can evaluate the impact of their various technical experiments. This process, applied over the long term, is likely to help players truly maximize their potential.

More in the next article on how the upper body interacts with legs in the backswing phase. Stay tuned.