This is a very lengthy article, so I decided to post the pictures first, before the text explanations. A word of warning. This is an article that focuses on pronation in the serve. I agree with the pronation point of view in many respects. But I am still free to disagree with Anatoly Antipin as far as the statement that turning the body parts other than the forearm and upper arm do not contribute much to the EFFECTIVENESS of the serve.
Sampras' serve is said to have the weight of a bowling ball. I believe the weight comes from the degree he throws his weight behind a very wide coil. He is also said to have the RPM of 4600+ revolutions!
Replies to This Discussion
Well, the idea behind it is to increase the amount of possible twist by creating more "room" by way of a larger angle of attachment. I don't use it because the problem this creates is that the racquet head may end up driving the ball long, instead of snapping down vertically enough.
If you look at Sampras, he uses less of a severe angle of attachment (hand to handle), yet he gets plenty of pronated finish on his second serve.
Figure 2.4. Federer serve
Different Serves and Arm’s Actions
2.2 Different Serves and Arm’s Actions
There are a lot of speculations about the wrist movement (the wrist snap, the
wrist whip effect) during the last second before impact. Some of the tennis
specialists (Vic Braden etc) say no such thing occurs. But, others (Brian
Gordon and so on) insist the wrist motion is very important. But what do the
arm and the wrist in particularly really do? When I’m serving I feel like my
wrist is doing something very essential, however maybe my feelings mislead me.
But, the pictures never lie.
The Kick Serve with Effective Pronation, Wrist Ulnar Deviation, and no Wrist
This time, I’m going to analyze the pictures from Figure2.10 and Figure 2.11.
They show the set of the video’s frames taken during the Lleyton Hewitt and
Andy Roddick kick serves. Let’s pay attention to the wrist motion
Figure 2.10. The wrist at the last second before impact Lleyton Hewitt’s kick serve
Figure 2.12. The wrist at the last second before impact Andy Roddick’s kick serve
These players employ the Continental grip. From above pictures we can see the
racquet string plane has practically constant vertical orientation. On all
pictures, the hand is almost straight out and there is no extension or flexion
of the wrist before and for the duration of the impact.
At the same time the wrist ulnar deviation directs the racquet upward very
fast. The pictures show this movement takes place in the plane which coincides
with the racquet string plane. Hence, it can produce the brushing boll motion
only. But, the brushing motion mostly responsible for the ball rotation, not
for the ball speed. The wrist ulnar deviation might be used to create different types of the spin serves and
this is a very good option. Nevertheless, it also can decrease pronation
efficiency, the most important part of the tennis kick serve.
OK, so far I don’t see any wrist actions in the kick serve which could add any
real speed to the tennis ball, but only the spin!
Figure 2.7. Stosur spin serve
Definition: Racquet efficient length Rel is the distance between player’s hand (point O on the Figures
(2.4-2.7) and the ball during impact. I think Rel = 25” (63.5
cm) in the most occasions.
Definition: Arm efficient length Ael is the distance between shoulder joint and player’s hand.
Since everybody have different arm size, I guessAel = 25” (63.5
cm) as average length.
On the pictures above,RV is the radius of the
arm and the racquet vertical rotation,RH – the radius of the racquet horizontal
RV = Ael + Rel × cosβ =25” × (1+ cosβ), where:
Notation: pronation angle β is the angle between long axis of the racquet and axis of the forearm/arm (Figure 2.4-2.7).
RH = Rel× sinβ = 25” × sinβ
RV can vary from Ael to Ael + Rel (or from 25” to 50”) because cosβ has
range from 0 to 1, depending on the β magnitude. RV can never be equal to
zero, because Ael or the arm efficient length is constant and equal to 25”.
RH can vary from 0 to Rel (or from 0” to 25”)
because sinβ has range from 0 to 1. RH can be equal to zero
and therefore linear speed would be zero! It can be very big problem for the
tennis player. Maintaining the proper magnitude of the angle β before
impact is absolutely crucial for pronation! On
figures from 2.4 to 2.7 the best players keep β from 35° to 45°depending
on the serve type.
How they are able to do that I described in step 2.2.2.
Notation: |VLV| - Linear speed of the
racquet in the vertical plane; |VLH| - Linear speed of the racquet in the horizontal plane. VLV and VLHare corresponding
velocities. Reminder: the linear speed = radius × angular speed. In the last formula the angular speed
should be expressed in radians. The angular speeds in degrees (from Figure 2.3)
were: ΩV=2°/ft, ΩH=20°/ft. In radians they are ΩV=(π/90)/ft, ΩH= (π/9)/ft.Then linear speeds in
the vertical and horizontal planes can be calculated according to the following
|VLV|= RV× ΩV= 25” × (1+ cosβ) × (π/90)/ft = 25” × (1+ cosβ) × (π/90) × 300/sec
|VLH|= RH× ΩH= 25” × sinβ ×
(π/9)/ft = 25” × sinβ ×
(π/9) × 300/sec
The sum of the linear racquet speeds would be |VLV|+ |VLH|. The results of the
calculation are presented on the Figure 2.8
Figure 2.8. Linear speeds of the racquet in vertical |VLV|and horizontal |VLH| rotations and their summation
The data on Figure 2.8
demonstrate, if the angle β ≥
12° the linear speed of
the pronation |VLH| begins to prevail over
the linear speed of the vertical rotation |VLV|.
It should be noted, unfortunately, the calculated above pronation linear
velocity determines mostly theoretical potential maximum. In reality, this
speed may be slower even in case when the pronation angle beta has appropriate
value. I’ll explain this phenomenon later, in the step 2.2.3.
Since, RH = 25” × sinβ, then we can calculate the pronation efficiency
according to following formula
Efficiency = sinβ×100%.
The results of the calculation are presented on the Figure 2.9.
Figure 2.9 Pronation’s efficiency as function
of the angle β
OK, it appears I found the proof! In case of the kick serve, the
pronation can really provide much bigger linear speed of the racquet than
others body limbs (except the wrist) altogether! But, if the pronation angle β=0°, the pronation produces
nothing at all, just the proper racquet string bed orientation.
That’s why I repeat
again, the best tennis players keep the pronation angle β around 30° - 45°
(Figure 2.4 -2.7). Maintaining the proper magnitude of the angle β before
impact is absolutely crucial for pronation! If the pronation angle has the
proper magnitude the pronation would be the most important and effective
contributor to the powerful kick serves!
The arm pronation can produce angular speed around Ω = 6000°/sec, trunk rotates around 500°/sec, arm rotates in vertical plane around 600°/sec, etc. What motion is the most important? There is no uncertainty, at least for me, it should be pronation and
Seriously About Flat and Kick Serve
2.2.2 The Flat Serve with Intensive Pronation, Wrist Flexion, and Restrained Wrist Ulnar Deviation
Figures below give us an idea about arm actions during typical flat serve.
Figure 2.15. The arm’s actions during Marat Safin's flat serve
Figure 2.16. The arm’s actions during Kevin Anderson’s flat serve
In case of the typical flat serve the main components of the racquet speed are: the fast arm
pronation, the fast wrist flexion and slow vertical arm rotation. If players
are employing Continental grip, all these motions mostly create the flat
component of the racquet speed and practically no spin. The most important
would be the arm pronation.
It follows from these photos, during the pronation phase of the flat serves both players keep their
right elbows in bend position. In all previous analyzed spin serves, most
pros maintain their arms during pronation phase practically straight.
Question: Why during the pronation phase, these players keep elbows in bend position?
Answer: When we swing the racquet upward our shoulder joint brings the upper arm in vertical and
the forearm in horizontal position. After that, by using fast elbow
extension, the forearm moves upward and the pronation angle β equal 90 degrees (Fig. 2.15.1). If
the elbow unbends completely, it brakes and inevitably the racquet starts
moving upward by using inertia and very fast wrist ulnar deviation. This
motion reduces the pronation angle beta (Fig. 2.15.4 β=45°) and can kill pronation
component of the racquet speed. To prevent it from occurrence, even during
the impact, the elbow joint should be bent. Next pictures (Figure 2.17)
illustrate Andy Roddick’s arm action during flat serve and confirm last
Figure 2.17. Andy Roddick's arm actions during flat serve
Pay attention on the angle between axis of the upper arm and the forearm. This angle is never less
than 30 degrees. Andy constructs the motion which often called as elbow snap.
Maybe this is the main secret of his so successful flat serve. For instance,
Marat Safin during impact keeps his elbow straight (Fig. 2.15.4). That’s why,
perhaps, his serve is slower than Roddick’s. But, Marat straightens his arm
at the very last moment before impact by using the elbow extension. This
motion doesn’t have enough time to decrease significantly pronation angle
beta, and produces just spin (not flat component). Hence, it enhances the
In everyday life, when we “pronate/supinate” a screwdriver, for example, we always unconsciously
keep elbow in bend position to increase a force applied to the screwdriver’s
handle. The same natural motion Andy applies to the racquet’s handle. He uses
bend elbow as some kind of “force multiplier”. I think, it doesn’t make any
harm if this force would be very active during impact. This screwdriver
approach could be very handy to overcome the inertia’s resistance of the
tennis ball and hence, increase its speed. To increase the flat component of
the racquet speed all the best servers also apply the wrist flexion.
I believe, we can keep the bend elbow before and during impact even in case of the kick serve.
It could help to curb excessive activity of the wrist ulnar deviation and
hence keep the appropriate amount of the pronation angle beta (not less than
There is one more way to keep pronation angle β in proper range 30° - 45°
We may also use slightly modified Continental grip. Next picture shows what kind of the grip Taylor
Dent brings into play for flat serve.
Figure 2.18. Taylor
Dent uses modified Continental grip
Taylor keeps his finger knuckles parallel
to the long axis of the tennis racquet. With this type of the grip it is
physically impossible to align the long axis of the racket in parallel to the
axis of the forearm/arm and hence, impossible to kill the pronation component
of the racquet’s speed. In fact, the pronation angle cannot be less than 30°.
I would recommend this grip for any type of the serves, except the slice
Originally Posted by LeeD
Are you sure you can use the same grip, pronate on flat serves, then pronate also on heavy topspin serves swinging UPWARDS??????
Figure 1. Andy Murray
Take a look at Andy Murray. He definitely is going to hit topspin serve. His grip is continental,
because the racquet string is perpendicular to the court ground. He is going to
use the wrist fast ulnar deviation to produce heavy topspin. If we employ
eastern backhand grip, the racquet would be semi closed. And automatically
around 50% of the racquet kinetic energy will be wasted to produce topspin
component of the ball velocity and hence only 50% on the linear ball speed. In
this occasion, we practically don’t need the wrist ulnar deviation. This grip
makes it easier to generate topspin, but we lose around 50% of the boll speed.
I think, tiw Nadal switched his grip toward continental/eastern grip and got
much more powerful serve.
Anatoly Antipin - one of the most delicate tennis players in the world.
Feel free to ask me anything about the technical aspects in this article.
this is all interesting to a point but those who need this have problems with their serve and those of us lucky ones its just natural.
Maths means nothing to me Rhythm is the most important, recently i just had a couple of decent serves beat me on the "T" prob bout 90 mph, Sampras, Fed etc only ever serve big at 120mph but hit position best maybe we should talk about that??
Also the big hitters are meat and drink to the great recievers, the best servers for me someone like Stich had no pretensions. Kick is great but it also puts the ball in a great hitting space, Stich hit slice/swerve which was a lot more problematic. Look at Goran not the fastest server but one of the best.....
I totally agree with you that placement is more important than power. I can rely on my slice serve, which can pull someone into the next court or side fence, for an easy volley. It is never more than 95 mph.
However, there is a need for the kicker when the returner is tuned into the flat and slice. I'm for variety to throw off returners.
I am not against a math or physics explanation. I believe it reveals alot into where the racquet head speed comes from. For most, the serve is the hardest stroke to learn. If for you, it is rhythm and naturalness, you are indeed lucky.
But I have seen two things from that. I have seen "natural" servers (former baseball pitchers) hit the heater, but never could develop spin for consistency.
I also have seen high school coaches insist their players hit with no pronation and hit serves with the patty cake method. For me, it is important to be knowledgeable.
Yes, Stich was a great server. I liked his motion alot. He did look very natural and basic, and he never rushed his serve.
As for Goran, I could not stand watching him double fault as much as he did. I believe it all stemmed in how open his stance was, making it impossible to turn enough to develop a strong kick serve. Everything was either flat or slice. He even admitted in an interview that he had no kick serve.
I think Becker summed it up right in his recent serve clinic I saw. He said he need to stand not too open because it gave him no angle to the ball, and not to closed because it hurt his lower back.
Tim post Tim!! This is a great point to hitting a proper kick serve. Couple things...
If legs and core don't matter to power or to generate kick then why jump or twist at all?
Also, I think that using your whole body will help generate the 80% in the arm. Using the kinetic chain of energy (starting w/ your feet all the way to your wrist) I think that each part of the body helps generate the 80%. What do you think?
Again great post!
It is obvious to me that it is much better to use the whole body in assisting the pronation, as you say. To hit with pronation alone is like sprinting on a bicycle only from the sitting position, when more power can be had from using body weight.
Sampras really turned alot, more than 180 degrees. He also pulled his right elbow behind his body, as if aiming across his chest back towards the baseline. He rocked from the very beginning of his toss, and he went into a full squat if he wanted.
Becker is another example of the same thing. Enormous shoulder and leg work, think of that squat! He just did a more "arms up together, down together" circular takeback instead of the scissors take back Sampras does.
I personally am against the Braden, "pronation only" school. And I went to his academy for five days!
I did actually try the Ivanisevic serve motion for myself. I never hit faster balls than that in my entire life, using a lower toss and hitting the ball on the way up to the peak. The super open stance position does enhance the racquet drop in the back by a large margin. But there is not enough time to get sideways enough to get topspin.
I tried it the Ljubicic stance method with essentially the same motion, but with a stance more closed. I got more hips involved and weight, along with topspin angle from the stance being more sideways. But then from there, I said to myself, why not just go more sideways like Sampras does and get the maximum weight transfer, using the entire body?
The minor problem from there is to figure out the toss and not have it shift because of the sideways stance. I figured out that after about a couple of weeks. ;)
Just to update. I really experiment with the ideas in this article in the last week. I found myself tonight figuring out the true import of this article. I have never hit harder serves in my life and have easily increased my serve speed by ten percent. In every department: flat, slice, top, kick.
Using the body is important as the frame, but now I see a much different way of serving, where I emphasize the snap and whip motion more. My practice session was so easy on my body. I hit bombs for an hour and half, whereas before I would be tired after an hour concentration on rotation, twisting and arching hard. You can twist and arch hard, use counter rotation with the pinpoint, or bend both your legs deep. But a full pronation is the most dynamic way to hit the serve period, while using the other elements as the frame to direct the pronated arm strike.
Don't take my word for it. I had nine people from the next courts watch me hit my practice serves tonight.
Here's an article I wrote on the serve:
Biomechanical Analysis of the TennisService
The service isarguably the most important stroke in tennis; starting off every point and being the only stroke over which the player exerts full control. Without an effective service it is very difficult for a tennis player to succeed, as illustrated by the 2008 US Open men’s final, in which runner-up Andy Murray was only able to win 49% of the points on his serve, compared to 65% by his opponent,Roger Federer (www.atpworldtour.com).
The service is also the most complex and difficult stroke to master, requiring sequential co-ordination of virtually every body segment. Despite this complexity, professional tennis players are regularly able to exceed service velocities of over 200 km/h (Girard et al., 2005), with the current world record, achieved by Andy Roddick during a 2004 Davis Cup tie, standing at 249.4km/h (Collins, 2008).
From a biomechanical perspective, the service can be viewed as a kinetic chain, involving the transfer of both linear and angular momentum (Bahamonde, 2000). This chain begins with the vertical ground reaction forces generated by the players through their footwork (Girard et al., 2007) and ends with the racquet head connecting with the ball. In the split second between these two events momentum is transferred upwards and forwards through the legs, trunk, upperarm, forearm and wrist (Ivancevic et al.,2008).
This article is a review of the literature pertaining to the biomechanics of the tennis service and the relative contributions of the different body segments to it.
The Lower Limbs
In the tennis service, the maximum velocities of body segments increase gradually in aproximal-to-distal sequence from the knee to the racquet (Elliot et al., 1995; Van Gheluwe &Hebbelinck, 1986). As such, research on the lower limbs, the starting point formomentum transfer, has been an important area of focus.
Girard etal. (2005) observed that elite players produced greater ground reaction forces (GRF) in the vertical direction and also greater changes in vertical GRF than novices, both of which indicate a more forceful leg drive amongst the elite performers. This resulted in greater contact heights and faster service velocities.
They also argued that the combination of eccentric and concentric muscle action during the leg drive are an example of the stretch-shortening cycle; a process by which more force can be generated when a concentric contraction is preceded by an eccentric contraction. The leg drive can be likened to a counter-movement jump: elastic energy is stored inthe lower limbs during knee flexion and if this knee flexion is followed quickly enough by extension then the stored energy will assist in a greater leg drive.
A 2007 study by Girard et al. utilised the restraint paradigm to investigate the effect of restricting knee motion during the service. Subjects performed serves in two conditions: firstly, as they would do normally, and secondly, with their legs kept outstretched by splints, allowing only the trunk and upper limbs to contribute to velocity.
It was found that restricted knee motion compromised vertical GRF, as well as the change in vertical GRF. This led to service velocities approximately 15% slower across all abilities tested. A strong relationship between the change in vertical GRF and post-impact ball speed was also noted.
Knee flexion, though, does not just serve to generate GRF. It also increases the distance the racquet must travel to the ball, thus providing a greater distance over which racquet head can accelerate and velocity can be generated (Elliot, 2001).
Both the 2005 and 2007 studies by Girard et al. used participants of all standards. This range of abilities, from novice to elite, provided insightful comparisons and information on why the serves of elite players are quicker; in this case the reason being a greater change in vertical ground reaction force, leading to a more powerful leg drive. It is also worth mentioning that neither study reported significant differences in lower limb muscular power between the different abilities, suggesting that the differences observed in the kinetic variables were due to the subjects’ differences in technical expertise.
The extension ofthe knees in the leg drive provides upward linear momentum, transferring the ground reaction force from the lower limbs to the trunk (Elliot et al., 1995). According to Elliot et al. (1995), the combination of trunk rotation and the forward drive from the lower limbs contributes 9.7 % of the impact velocity of the centre of the racquet head. Yet despite this significant contribution by the trunk muscles, specific research on them is sparse.
Based on previous research, Bahamonde(2000) noted that angular momentum is derived from the rotation of the trunk, and that this rotation also helps elevate the racquet shoulder and lower the opposite shoulder, allowing players to reach a greater contact height.
Chow etal. (2003) monitored EMG signals from the left and right rectus abdominis, external oblique, internal oblique and lumbar erector spinae. Their results showed that muscle activity peaked after the racquet reached its highest point and before it reached its lowest point behind the trunk. Muscle activity was greatest in the left internal and external obliques (for a right-hander server).
Little EMG activity was observed until after ball release; leading to the conclusion that minimal trunk motion during ball release leads to a more consistent ball toss.
This study, however, provided little scope for generalisation, due to the limited sample size and use of only highly skilled players; a fact acknowledged by the researchers themselves.
There are seven major anatomical rotations at the upper limb alone (Knudson, 2008), making the swing up to the ball the most difficult phase of the service to co-ordinate. Most studies (e.g. Elliot et al., 1995; Fleisig et al., 2003; Tanabe & Ito, 2007) implicate shoulder internal rotation and wrist palmar flexion as the movements contributing most to final racquet head velocity.
Fleisig et al. (2003) were the first researchers to quantitatively analyse the first-serve kinematics of some of the world’s topmale and female professional players; which they did at the 2000 Sydney Olympic Games. They reported a rapid kinematic chain of rotations in the short time between maximum shoulder external rotation and ball impact, consisting of forward trunk tilt (280 °/s), pelvis rotation (440 °/s), upper torso rotation(870 °/s), shoulder internal rotation (2420 °/s for males and 1370 °/s forfemales), elbow extension (1510 °/s) and wrist flexion (1950 °/s).
The only significant differences found between the male and female players were in shoulder internal rotation and service velocity. The faster service velocities of the male players were at least partially attributed to their greater shoulder internal rotation velocities.
Further evidence for the importance of shoulder internal rotation was found by Tanabe & Ito (2007), who observed that for a fast server (ball speed 45 m/s), the contribution of shoulder internal rotation was 54.0 %, while for a slow server (ball speed 30 m/s) shoulder internal rotation contributed only 25.9 %. They also noted a significant positive correlation between the contribution ofshoulder internal rotation exhibited and linear racquet head velocity atimpact.
They concluded that shoulder internal rotation and wrist pronation were the final components of the service action and major contributors to final racquet headspeed. Based on their results, they advise slower servers (who tend to developracquet head speed by horizontal flexion of the shoulder joint) to generate a greater angular velocity of shoulder internal rotation.
One of the most highly regarded studies on the tennis service is that by Elliot et al. (1995), who found mean peak angular velocities of 2091 °/s for shoulder internal rotation and 1719 °/s for wrist flexion. These values are similar to, but slightly lower than, those reported by Fleisig et al. (2003), though the higher values from the Fleisig et al. study can be attributed to the fact that their sample consisted of top professional players.
They also found that shoulder internal rotation contributed 54.2 % ofthe mean linear velocity of the racquet head, a finding consistent with Tanabe & Ito (2007). Their value of 31.0 % for the contribution of wrist flexion, however, was higher than the 26.0 % reported by Tanabe & Ito (2007) for their fast servers. Furthermore, mean linear racquet head velocity was 31.0m/s; slower than the 38.1 m/s recorded by Tanabe & Ito (2007), a difference that could be attributed to the differing wrist flexion contributions, with too much wrist flexion actually compromising service velocity.
The large shoulder internal rotation and wrist flexion contributions recorded in all ofthe studies reflect the significance of these movements in generating racquethead velocity during the service action.
The importance of shoulder internal rotation is further highlighted by Chandler et al. (1992), who, in testing 24 college tennis players, noted a 25 % higher value for shoulder internal rotator muscle strength and power inthe dominant arm compared to the non-dominant arm.
Further to this, Mont et al. (1994), in an intervention study, found that both isokineticconcentric training and isokinetic eccentric training of the shoulder internal and external rotator muscles produced strength gains of 11 %, which led to an 11 % increase in serving velocity. The trained subjects were also subsequently able to maintain their service velocity over a greater period of time than those in a control group.
Research on the service has indicated that kneeflexion followed closely by extension is a prerequisite of serving effectiveness and forms the basis of linear momentum transfer in the service.
This linear momentum is then transferred from the lower limbs to the trunk, the rotation of which enables the server to reach a greater impact height and from which angular momentum derives.
The final phase of the kinetic chain isthe rotation of the upper limb segments and subsequent generation of racquet head velocity, with shoulder internal rotation and wrist flexion contributing most.
The studies conducted have provided much useful information on the biomechanics of the tennis service. The majority of them, though, are limited by their use of players already possessing a fast service. Studies involving players of all abilities, such as those by Girard et al.(2005; 2007), allow comparisons to be made, enabling researchers to better quantify the different mechanisms used by fast and slow servers. Future research should, therefore, focus on variations in serve kinematics, joint kinematics and ball velocity amongst players of different skill levels.
In conclusion, the research conducted on the tennis service has clearly demonstrated the significant individual contributions of the lower limbs, trunk and upper limbs in the generation of racquet head speed. There is, however, no single component which makes for an effective service; rather, it is thesequential co-ordination of each and every body segment that determines howeffective a service is.
Very interesting information here. I just believe that each part of the kinetic chain has to be thought in a hierachical sense of importance and yet all in a synchronous whole.