The Upside-Down Bow, and Why It’s Valuable to Play in Ways You Never Would on Stage

One of the most memorable exercises I can recall from my Suzuki days was the one where I was asked to turn my bow the other way, holding it at the tip instead of at the frog.

I remember it being more challenging than it looked (and it sort of freaked me out that I might break the bow), but this memory popped into my thoughts the other day and I started to wonder. Was this just a goofy creative thing my teacher did to keep me engaged, or is there a research basis for changing things up like this?

Serving velocity and accuracy

A team of Spanish researchers recruited 30 youngsters with about 2-3 years of tennis experience to practice improving the accuracy and velocity of their serves.

Each player started out by completing a baseline test of their serving abilities. Standing on the left (“ad”) side of the court, they were given 20 serves, and the instructions “Serve at the highest speed and with the highest accuracy possible aiming at the target (which was a 50cm-diameter area in the service box).”

Then, the players were randomly assigned to one of two groups, and completed 12 practice sessions (60 serves per session) over the course of 4-6 weeks.

The “consistency group” practiced the serve while standing in the exact same spot, using the same movements, in an effort to produce as consistently accurate and speedy a serve as they could.

The “variable group” did the same amount of practice, but introduced all sorts of variation into their practice from one serve to the next. They practiced serving on one leg, changing the width of their stance, and standing on a thick mat to increase their height. They also served from a position several feet in front of or behind the baseline. Or near the center of the court or way off to the left. With the ball tossed in front of them, in back, to the left, to the right, really high, or really low. They even served while facing 45° or 90° away from the net.

What’s the best approach?

After the final training session, everyone completed a performance test. As in the baseline test they took before the study, they hit 20 serves from the left side of the court, aiming at a 50cm-diameter target in the service box.

Both groups were now hitting the ball significantly harder than they had in the baseline test. The consistency group was hitting serves that traveled 4.8% faster than at baseline. And the variable group increased their velocity by 7.68% (faster on average than the consistency group, but not a statistically significant difference).

However, it was a different story when it came to accuracy.

As it turns out, practicing with consistency did not lead to any improvements in accuracy.

Conversely, practicing with all sorts of variations from one serve to the next led to a significant improvement in accuracy, reducing the participants’ error from 2.67m before training to 2.21m after (i.e. the participants were hitting the ball about half a yard, or 18 inches closer to the target than they did before the training).

How could practicing a serve one-legged, behind the baseline, and with the ball toss all over the place – “incorrectly,” in other words – actually result in better performance? Isn’t that totally backwards of how it should be?

Why?

“Schema Theory” suggests that in learning a motor skill, our brain develops a Generalized Motor Program (GMP), or a basic template for a particular movement pattern.

So instead of having one motor program for hitting a tennis ball which comes towards us at waist height with topspin and a second motor program for a ball which comes towards us at knee height with backspin, we have a single program which we can adjust and tweak in the moment and apply to both types of balls.

The unique variables of the ball (height, speed, spin, etc.) are the “parameters” which determine what sort of adjustments we have to make to our Generalized Motor Program. So when we practice, not only do we need to develop a generalized program, but we also have to ensure that this GMP is adaptable and can be used successfully across a range of parameters. A process which some have called “parameterization.”

In the case of the tennis players above, both groups already possessed a basic GMP for a serve. And while I imagine there’s always going to be some benefit from refining one’s GMP, it seems like the variable group’s practice emphasis on parameterization gave them an advantage over the consistency group.

So somewhat paradoxically, it seems that we may be able to increase consistency in performance, not by practicing for consistency, but by working to develop a more flexible skillset. So that you can make a phrase sound exactly the way you want whether a hall has the acoustics of a bathroom or a sound-proofed panic room.

Julian Martin on flexibility

I recently had a chance to ask pianist Julian Martin about how to adapt more quickly to unfamiliar pianos, and his take was that we need to practice producing the effect we want for a note, measure, or phrase in more than one way. So that even if the piano we are performing on feels nothing like our piano at home, our technique is flexible enough to adapt to the piano in front of us. Much like this tennis study suggests.

But how might one develop such flexibility?

He noted that Claudio Arrau’s teacher Martin Krause strongly believed that his students ought to be able to play a phrase with a huge range of dynamics and at a wide range of tempos – even if this means playing it (at least during practice) in a way that is totally inappropriate for the context of the piece. Because this way, we at least have cultivated the option of playing it in whatever manner we feel is most fitting to the context of the music, regardless of how dead or live our performance piano may be. Or how subpar our reeds are that day. Or how fast our heart might be beating in the moment. (And if this sort of artist insight/research overlap is totally your cup of tea, I’m co-developing an online performance psych course with Juilliard, which includes much more of this conversation with Julian Martin and other artist-faculty.)

Take action

Can you play tricky or important passages and shifts with more musical intention? Faster and slower? Louder? Super soft?

The key is to introduce relevant variability and practice parameterization, but without going too far past a point where it starts distorting the movement pattern beyond recognition. For instance, I think you can safely skip practicing arpeggios while hopping on one leg. But playing on different pianos, using your backup bow, even trying to play so as to be accused of playing “too musically” could all be good practice for that day when you have to still play beautifully when you’re exhausted from a long week of rehearsals and other gigs, the hall sounds like the inside of a middle school library, and your quartet-mates seem to have decided to pursue a very different tempo than you thought you had all agreed on…

Additional reading

Still wondering what this upside-down bow thing is all about? Here’s an entertaining account of a Suzuki alum’s experience of this exercise:

Upside Down Bow – Balanced Bow: Notes from my studies with Dr. Shinichi Suzuki @Stay Tuned