The toss: parabolic, or straight up and down?

[AlexR]

Many pros toss the ball in a parabola, both into the court, and often right to left along the baseline, almost behind the server for a topspin serve. And this is taught at all levels of instruction. It’s also common advice to keep the toss arm parallel to the baseline.

For a long while I tried following all these tips on the serve, and I guess I’m just not coordinated enough to toss the ball consistently into the court and into a comfortable spot for topspin from that position. It gives you two axises to mess up.

So, from recent experience, I submit that the straight up and down toss is perfectly valid, and that “toss arm parallel to the baseline” is more of a suggestion than a rule.

Recently I’ve just started tossing with my arm more perpendicular than parallel to the baseline. Maybe the acute side of the angle would be 75 or 80 degrees out from the baseline. Straight up. Because my stance is pretty upright, this is about a foot into the court. It’s not going behind me for that natural topspin, but I can still put plenty on.

I flow right from this toss into making sure the toss shoulder goes up and the racquet shoulder goes down. Then i swing. A very simple and repeatable motion.

I’m not twisting my back in an extreme way to wring out an extra few mph. but im landing serves in the solid 90s so much more consistently. I believe Delpo’s toss is like this so that validates it a bit for me.

 

[SystemicAnomaly]

Parabolic is the norm. Not sure that straight up and down is really feasible or optimal for a high-level serve. Tossing toward the net is not ideal either. It is difficult to achieve much of a coil on the serve without a pretty high toss.

But then, a very high toss would unlikely be straight up and down. Not a fan of Stanimal's toss (toward the net). He does not yet very much of a coil with his toss direction. This and other elements of his serve appear to put more stress on his right shoulder. Note than Stan's toss will often arc backward... not straight.

For students who are not comfortable lifting the tossing arm parallel to the baseline, I suggest something between a 30 degree and 45 degree lift direction wrt the basline. More than 60 degrees is usually not conducive to a good service coil.

 

[SystemicAnomaly]

DelPo's toss isn't really straight up and down. It travels forward quite a bit... 2 feet or more? His toss is pretty high. He doesn't get a whole lot of coil on his service motion but he makes up for it, in part, with his size and his considerable leg drive.

I have to wonder if the lack of coil on his serve contributes to all his time on the DL. Wrist and other injuries have kept him from competing quite a bit

 

[AlexR]

Not a fan of Stanimal's toss (toward the net). He does not yet very much of a coil with his toss direction. This and other elements of his serve appear to put more stress on his right shoulder. Note than Stan's toss will often arc backward... not straight.

I was watching Stan to see what you mean and happened to realize that Djoker basically does what I’m talking about. His toss looks straight up and down, and his toss arm isn’t parallel to the baseline at all. FWIW these guys are both very good servers by the numbers. Not the very top servers but they get a good amount of aces and unreturnables. I don’t believe something should be considered incorrect if multiple slam winners, including one of the all-time greatest players, do it that way.

 

[SystemicAnomaly]

I was watching Stan to see what you mean and happened to realize that Djoker basically does what I’m talking about. His toss looks straight up and down, and his toss arm isn’t parallel to the baseline at all. FWIW these guys are both very good servers by the numbers. Not the very top servers but they get a good amount of aces and unreturnables. I don’t believe something should be considered incorrect if multiple slam winners, including one of the all-time greatest players, do it that way.

Two important things to consider. The anatomy of the elite players you've mentioned might allow them to get away with certain idiosyncrasies in mechanics that will not necessarily work for you and your anatomy. Usually better to see what most top servers have in common and see how you can adapt the various elements to suit your own anatomy and skill set.

Another crucial part of this equation is the predisposition or risk of injury. Sure, Stan has been able to serve some huge heaters. But he has also been plagued by shoulder injuries. When I studied his serve mechanics some 4-5 years back, I predicted that he would likely develop arm/shoulder problems, even tho he appears to be a very strong guy. Sure enough, he had experienced significant shoulder problems in 2015. He may have also had some issues issues around 2013/14 IIRC. In the past few years, he has continued to be plagued with shoulder problems.

Don't know what version of Novak's serve you are referring to. He has changed it several times in the past decade. Actually, some of the most significant changes came around 2010. These changes were instrumental in his stellar 2011 year. However, it appears that he had changed it again around 2014/15. And then again more recently. Elbow problems were a large part of the reason for changes in the past few years.

Novak has served some fairly big serves in the past (130 mph?) but never as quite big as Federer, Murray and others. Placement/depth on 2nd serves has not always been as good either. For the serves that I looked at (from a few years ago), Novak was lifting his left arm at about 45 degrees wrt the baseline -- as I had suggested. It was not directly toward the net.

On his 2nd serves, I saw a very distinctive right to left arc for his serve toss. On his 1st serve toss, there was a lot less deviation in this direction... a very narrow parabola (which might be interpreted as nearly straight up and down). But even with this toss there was a significant forward arc to his toss. He released it a bit in front of the baseline but contacted it quite a bit forward of the baseline.

 

[Wurm]

How overtly the feeling of going "up" towards the ball is defines the difference between a first and second serve. I'm assuming that's the same for everyone.

The difference in production of my ball toss for the different serves comes from a subtle difference in how much of my body weight is leaning into the court as I release the ball. I might be atypical in the way I do this but it works for me.

 

[SystemicAnomaly]

Prior to 2010/2011, Novak's right elbow position during his trophy phase dropped too low (wrt his shoulder line). This is something I mentioned on these pages back in 2009. Appeared to result in some timing problems. He had a couple of other issues with his serve at that time as well. These flaws are things he corrected in 2010. Some changes in the past decade:

 

[Dragy]

When discussing elements we tend to sometimes leave out of equation the complexity and integrity of the motion. We want toss to be the boss and execute it as almost isolated task, but under that “almost” we hide several important circumstances: we want to be in a posture to coil and load for the serve, we want to follow the timing for the swing (unless we apply an extremely high toss leaving enough time to get a cup of tea before actually starting service motion), we might want to “toss into swing” rather then “swing into toss”, and so on.
With due consideration of the above and taking into account optimal range for arm action (I’d say between 90deg against shoulder line and 30-45deg) we can figure out how to toss the ball. If you want to coil back like textbook techniques prescribes, you’d settle somewhere between parallel to the baseline and 45deg in. Take into account it also depends on deuce/ad side - around 30deg difference to initial setup orientation.
A “cheat” is available for step-up pinpoint servers: if you prepare in wide stance sitting on your back leg like Kevin Anderson or Felix Auger-Aliassime, you can toss close to straight up and then get yourself under that toss as you step into pinpoint.
Now you can surely limit your coiling and therefore have ability to toss straight into the court, but I wouldn’t trade it for just the seemingly easier toss. You end up with either weaker swing or bigger strain against longer rotation from more coiled prep.

 

[SystemicAnomaly]

Novak elbow too low (prior to 2010)

Optimal elbow alignment:

 

[heninfan99]

Many pros toss the ball in a parabola, both into the court, and often right to left along the baseline, almost behind the server for a topspin serve. And this is taught at all levels of instruction. It’s also common advice to keep the toss arm parallel to the baseline.

For a long while I tried following all these tips on the serve, and I guess I’m just not coordinated enough to toss the ball consistently into the court and into a comfortable spot for topspin from that position. It gives you two axises to mess up.

So, from recent experience, I submit that the straight up and down toss is perfectly valid, and that “toss arm parallel to the baseline” is more of a suggestion than a rule.

Recently I’ve just started tossing with my arm more perpendicular than parallel to the baseline. Maybe the acute side of the angle would be 75 or 80 degrees out from the baseline. Straight up. Because my stance is pretty upright, this is about a foot into the court. It’s not going behind me for that natural topspin, but I can still put plenty on.

I flow right from this toss into making sure the toss shoulder goes up and the racquet shoulder goes down. Then i swing. A very simple and repeatable motion.

I’m not twisting my back in an extreme way to wring out an extra few mph. but im landing serves in the solid 90s so much more consistently. I believe Delpo’s toss is like this so that validates it a bit for me.

Of course the straight toss with a straight arm is perfectly valid and you're right it works excellently for Delpo. There's more than one way to do the ball toss, forehands, backhands & etc. Soderling also had a great, simple toss.

Many "high level" players use this straight style.

Straight is much simpler to learn than the Murray-style J-toss but the player should do whichever feels more natural.

 

[SystemicAnomaly]

As already mentioned in post #3, DelPo's toss wasn't actually straight up and down. And his chronic injuries (wrist and other) make me wonder if his lack of service coil hasn't been a factor.

And I'm not a fan of the Soldering serve as a model for most players. As a huge, 6'4" (193 cm) player, he could get away with a rather odd style (see below) that would not be optimal for other players. But even then, his career seemed like a blip. On the tour for a while but he didn't make much of a splash until 2009. Played his last pro match in July 2011 (even tho his announced retirement was 2015). Stopped play in 2011 cuz of injuries and mono.

https://tt.tennis-warehouse.com/index.php?threads/soderling-serve-analysis-video.319568/#post-4509903

 

[Kevo]

So, from recent experience, I submit that the straight up and down toss is perfectly valid, and that “toss arm parallel to the baseline” is more of a suggestion than a rule.

I am not sure exactly what you mean by straight up and down. I frequently teach my students a toss that uses a bent elbow and tries to lift the ball straight vertically into the air. This solves a lot of problems for some people since you are not dealing with the ball moving from left to right. I really don't care how they toss as long it makes it to the right location for contact and that happens consistently. For some people it seems impossible for them to do it using a straight arm and swinging from the shoulder, but I've yet to find someone who can't learn the bent arm lift the ball straight up method, so I use it quite a lot.

 

[AlexR]

Two important things to consider. The anatomy of the elite players you've mentioned might allow them to get away with certain idiosyncrasies in mechanics that will not necessarily work for you and your anatomy. Usually better to see what most top servers have in common and see how you can adapt the various elements to suit your own anatomy and skill set.

Another crucial part of this equation is the predisposition or risk of injury. Sure, Stan has been able to serve some huge heaters. But he has also been plagued by shoulder injuries. When I studied his serve mechanics some 4-5 years back, I predicted that he would likely develop arm/shoulder problems, even tho he appears to be a very strong guy. Sure enough, he had experienced significant shoulder problems in 2015. He may have also had some issues issues around 2013/14 IIRC. In the past few years, he has continued to be plagued with shoulder problems.

Don't know what version of Novak's serve you are referring to. He has changed it several times in the past decade. Actually, some of the most significant changes came around 2010. These changes were instrumental in his stellar 2011 year. However, it appears that he had changed it again around 2014/15. And then again more recently. Elbow problems were a large part of the reason for changes in the past few years.

Novak has served some fairly big serves in the past (130 mph?) but never as quite big as Federer, Murray and others. Placement/depth on 2nd serves has not always been as good either. For the serves that I looked at (from a few years ago), Novak was lifting his left arm at about 45 degrees wrt the baseline -- as I had suggested. It was not directly toward the net.

On his 2nd serves, I saw a very distinctive right to left arc for his serve toss. On his 1st serve toss, there was a lot less deviation in this direction... a very narrow parabola (which might be interpreted as nearly straight up and down). But even with this toss there was a significant forward arc to his toss. He released it a bit in front of the baseline but contacted it quite a bit forward of the baseline.

Hm, I’ll concede that I’m sacrificing max power by using less coil as a result of a straight toss with the toss arm extended into the court. However. I don’t agree that there’s inherently more injury risk.

Maybe trying to hit 130 mph without coiling is the injury risk. All I’m trying to do is make good contact with decent pop. I’m conscious of not overstressing any joints. For me a good first serve hits the back fence about 1 foot up on a typical public court, with slice. I don’t really hit flat. People tell me it’s probably mid 90s or 100. I don’t think I need 104 at the expense of landing fewer. Actually, I feel I’m stressing my tendons less because the more consistent toss means more swings with the optimal arm slot that’s smooth but also has power. Trying to make slight corrections for little variations in the toss was a source of muscle tweaks for me. i know you can just catch it but nobody catches every off toss.

 

[Chas Tennis]

Your toss will depend on where you choose to impact the ball with your particular serving technique.


Here are two different tossing techniques examined by Toly.

What the ball toss is better?

See also corresponding videos -

For the kick serve when you release the toss always consider that pro servers move their head forward before they impact the ball. Your toss will depend on where you choose to impact the ball with your serving technique. If you look at videos from the side view and watch you will see the head move forward between toss release and impact in high level serves.

For high level serves there is forward movement of the head from toss release to impact. You can hit a serve without moving your head forward but you are using your own technique. To check this, search for side views of kick serves and watch the head positions at ball release and impact.

The instruction to 'Toss the ball over your head for a kick serve' is simple but very misleading. Toss the ball to where you want to impact it and move yourself forward and to the side so that your head is under the ball for the kick serve impact. That description fits what is seen in high speed videos.

 

[Mountain Ghost]

I have a basic toss formula that I start EVERYONE with ... and then they can migrate to their own "solution" (or not) once they get better. THAT formula is that the tossing arm rises on the same directional line as the hitting arm travels during the forward swing. On the deuce court this would be to the right net post. This eliminates one potential inconsistency from the mix (left and right) ... leaving only height and depth to deal with. Another thing that helps is to START the toss from the bottom of the tossing arm path only ... NOT to start somewhere in the middle ... go down ... and then back up. This LACK of a change of direction in the tossing arm helps consistency a lot. Tossing arm should be straight throughout ... but not "stiff" ~ MG

 

[AlexR]

I made a video of what I mean. Serve is probably mid 90s with top and side spin. The ball hit the fence a foot up on 1 bounce which is basically the best I ever did when I was trying to do a coil and a parabolic toss into the court and back toward me, me as is coached. As you can see, this toss is straight up, no parabola. Action felt smooth, not forcing anything. Most importantly, it's balanced and repeatable for me specifically because I'm not coiling and I'm not using a "J toss" aka parabolic toss. That is too many moving parts for me, especially when tired during play.

I think this would be acceptable as a pro second serve, and is plenty for a first serve outside of those upper echelons. I simply don't think I have the flexibility or strength to serve much harder, even if I had a "perfect" motion. I'm 5-11" 145 lbs so quite thin, but more importantly I have bad flexibility.

 

[Raul_SJ]

DelPo's toss isn't really straight up and down. It travels forward quite a bit... 2 feet or more? His toss is pretty high. He doesn't get a whole lot of coil on his service motion but he makes up for it, in part, with his size and his considerable leg drive.

I have to wonder if the lack of coil on his serve contributes to all his time on the DL. Wrist and other injuries have Captain off quite a bit

Don't see the "straight up and down" with Djok. Leaving aside the right to left dimension, there is a parabolic back to front arc into the court.
At the point of release it is near the baseline and at point of contact it is ~ 2 feet into the court. So at least a 2 foot parabolic arc into the court.

 

[Raul_SJ]

I made a video of what I mean. Serve is probably mid 90s with top and side spin. The ball hit the fence a foot up on 1 bounce which is basically the best I ever did when I was trying to do a coil and a parabolic toss into the court and back toward me, me as is coached. As you can see, this toss is straight up, no parabola. Action felt smooth, not forcing anything. Most importantly, it's balanced and repeatable for me specifically because I'm not coiling and I'm not using a "J toss" aka parabolic toss. That is too many moving parts for me, especially when tired during play.

I think this would be acceptable as a pro second serve, and is plenty for a first serve outside of those upper echelons. I simply don't think I have the flexibility or strength to serve much harder, even if I had a "perfect" motion. I'm 5-11" 145 lbs so quite thin, but more importantly I have bad flexibility.

Have not examined your video closely, but in general, I do not like the idea of not coiling. One is leaving a lot of power on the table by not coiling and I don't see why one would not coil.
You can also definitely be tossing more into the court and moving contact point at least 1/2 foot forward. Again, leaving power on the table.

The challenge is combining the coil with toss into the court.


There are three tossing arm techniques:

1. Tossing arm roughly parallel to baseline (Federer, Karolovic, Sampras). This motion promotes great shoulder coil. Tough to toss inside the court.
2. Tossing arm roughly diagonal to baseline (Serena, Raonic). Promotes good shoulder coil. Easier to toss inside the court.
3. Tossing arm roughly perpendicular to baseline (Lisicki, Pliskova). Does NOT promote shoulder coil. Easiest to toss inside the court.

The challenge with the first two methods is how to toss the ball inside the court.

Nick from Intuitive Tennis video discusses the Parallel and Diagonal style. He mentions this caveat with Parallel style:

"I do not recommend this Parallel technique to recreational players as it's difficult to throw the ball inside the court. What needs to happen at the very end stage of the toss is that the arm has to go slightly forward to be able to bring the ball inside the court. The recreational players will have a hard time getting the ball to go forward.​

​

​

 

[AlexR]

Maybe I’m conflating not doing an arcing/parabolic/or “j” toss, with not coiling. It’s probably possible to ise a straight up and down toss, as well as coil. But, I think I am not coiling much anyway, which is on purpose. I used to try and add more coil. But now I just start with my hips and shoulders more or less 90 degrees perpendicular to the baseline and come around to the ball from there. I simply wasn’t seeing enough benefit from turning further away from the baseline. It was less balanced and less repeatable for me. Maybe I’ll work on adding more coil back in slowly. But I think I’m going to stick with the simpler straight up/down toss.

 

[movdqa]

I learned straight up and down and that's what I use. But I've been around long enough to know that different strokes for different folks.

 

[SystemicAnomaly]

To clarify... "J toss" is NOT the same as a parabolic path for the ball toss. J toss refers to the downward and then upward "hook" motion of the (ball tossing) hand as it prepares to lift the tossing arm and ball. Coined by Vic Braden in the 70s, I believe. Andy Murray has one of the closest implementations of this tossing style.

 

[SystemicAnomaly]

Maybe I’m conflating not doing an arcing/parabolic/or “j” toss, with not coiling. It’s probably possible to ise a straight up and down toss, as well as coil. But, I think I am not coiling much anyway, which is on purpose. I used to try and add more coil. But now I just start with my hips and shoulders more or less 90 degrees perpendicular to the baseline and come around to the ball from there. I simply wasn’t seeing enough benefit from turning further away from the baseline. It was less balanced and less repeatable for me. Maybe I’ll work on adding more coil back in slowly. But I think I’m going to stick with the simpler straight up/down toss.

Coiling is not really so much a product of the parabolic toss. It primarily is influenced by the direction the tossing arm rises. Coil is promoted by lifting the arm more or less parallel to the base line. Small angles, 45 degrees or less, will generally allow a decent coil. However, lifting the arm towards the net often significantly limits the amount coil unless the toss is very very high

 

[Raul_SJ]

Coiling is not really so much a product of the parabolic toss. It primarily is influenced by the direction the tossing arm rises. Coil is promoted by lifting the arm more or less parallel to the base line. Small angles, 45 degrees or less, will generally allow a decent coil. However, lifting the arm towards the net often significantly limits the amount coil unless the toss is very very high

Does Federer pretty much finish most of his shoulder coil before the toss is released? Maybe 70% of the shoulder coil is before release?

I think I am seeing little difference in the shoulder coil between frame #1 and frame #4 in the first pic. Looks like he adds a little shoulder coil after release but the main shoulder movement from frame 1 to 4, is increased shoulder tilt of the tossing shoulder to set up the "cartwheel" action. Any other major differences in terms of shoulders, hips, etc in those first 4 frames?

 

[Kevo]

I made a video of what I mean. Serve is probably mid 90s with top and side spin. The ball hit the fence a foot up on 1 bounce which is basically the best I ever did when I was trying to do a coil and a parabolic toss into the court and back toward me, me as is coached. As you can see, this toss is straight up, no parabola. Action felt smooth, not forcing anything. Most importantly, it's balanced and repeatable for me specifically because I'm not coiling and I'm not using a "J toss" aka parabolic toss. That is too many moving parts for me, especially when tired during play.

I think this would be acceptable as a pro second serve, and is plenty for a first serve outside of those upper echelons. I simply don't think I have the flexibility or strength to serve much harder, even if I had a "perfect" motion. I'm 5-11" 145 lbs so quite thin, but more importantly I have bad flexibility.

I don't see any problem with you using that toss motion. It looks simple and repeatable, and if you are consistent with it, keep using it.

I also would say don't worry to much about the speed for now or how high it is on the back fence. That's a pretty bad way to judge speed. I can pretty reliably hit the back fence on one bounce with speeds in the mid 50mph on my SpeedTrac radar. So there is a wide range of speed that your serve could be. What I would work on is placement and variety. If you want more pop you can certainly experiment with some of the tips that have been mentioned, and they will probably help, but I'd recommend getting comfortable and dialed in with your toss change first. Changing too many things too close together can cause you to backslide in your technique.

 

[heninfan99]

Maybe I’m conflating not doing an arcing/parabolic/or “j” toss, with not coiling. It’s probably possible to ise a straight up and down toss, as well as coil. But, I think I am not coiling much anyway, which is on purpose. I used to try and add more coil. But now I just start with my hips and shoulders more or less 90 degrees perpendicular to the baseline and come around to the ball from there. I simply wasn’t seeing enough benefit from turning further away from the baseline. It was less balanced and less repeatable for me. Maybe I’ll work on adding more coil back in slowly. But I think I’m going to stick with the simpler straight up/down toss.

All the guys you listed get shoulder turn(coil). Extreme shoulder turn is excellent for disguise.

 

[golden chicken]

I used to have a parabolic toss. I used a straight arm, starting around my right hip and lifting towards my left shoulder. My arm was roughly parallel to the baseline.

Since I play far less now than I did, I find using the straighter toss with my arm more perpendicular to the baseline gives me more toss consistency more quickly.